pacemaker  2.1.5-b7adf64e51
Scalable High-Availability cluster resource manager
unpack.c
Go to the documentation of this file.
1 /*
2  * Copyright 2004-2022 the Pacemaker project contributors
3  *
4  * The version control history for this file may have further details.
5  *
6  * This source code is licensed under the GNU Lesser General Public License
7  * version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY.
8  */
9 
10 #include <crm_internal.h>
11 
12 #include <stdio.h>
13 #include <string.h>
14 #include <glib.h>
15 #include <time.h>
16 
17 #include <crm/crm.h>
18 #include <crm/services.h>
19 #include <crm/msg_xml.h>
20 #include <crm/common/xml.h>
21 #include <crm/common/xml_internal.h>
22 
23 #include <crm/common/util.h>
24 #include <crm/pengine/rules.h>
25 #include <crm/pengine/internal.h>
26 #include <pe_status_private.h>
27 
28 CRM_TRACE_INIT_DATA(pe_status);
29 
30 /* This uses pcmk__set_flags_as()/pcmk__clear_flags_as() directly rather than
31  * use pe__set_working_set_flags()/pe__clear_working_set_flags() so that the
32  * flag is stringified more readably in log messages.
33  */
34 #define set_config_flag(data_set, option, flag) do { \
35  const char *scf_value = pe_pref((data_set)->config_hash, (option)); \
36  if (scf_value != NULL) { \
37  if (crm_is_true(scf_value)) { \
38  (data_set)->flags = pcmk__set_flags_as(__func__, __LINE__, \
39  LOG_TRACE, "Working set", \
40  crm_system_name, (data_set)->flags, \
41  (flag), #flag); \
42  } else { \
43  (data_set)->flags = pcmk__clear_flags_as(__func__, __LINE__,\
44  LOG_TRACE, "Working set", \
45  crm_system_name, (data_set)->flags, \
46  (flag), #flag); \
47  } \
48  } \
49  } while(0)
50 
51 static void unpack_rsc_op(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op,
52  xmlNode **last_failure,
53  enum action_fail_response *failed,
54  pe_working_set_t *data_set);
55 static void determine_remote_online_status(pe_working_set_t *data_set,
56  pe_node_t *this_node);
57 static void add_node_attrs(xmlNode *attrs, pe_node_t *node, bool overwrite,
58  pe_working_set_t *data_set);
59 static void determine_online_status(xmlNode *node_state, pe_node_t *this_node,
60  pe_working_set_t *data_set);
61 
62 static void unpack_node_lrm(pe_node_t *node, xmlNode *xml,
63  pe_working_set_t *data_set);
64 
65 
66 // Bitmask for warnings we only want to print once
67 uint32_t pe_wo = 0;
68 
69 static gboolean
70 is_dangling_guest_node(pe_node_t *node)
71 {
72  /* we are looking for a remote-node that was supposed to be mapped to a
73  * container resource, but all traces of that container have disappeared
74  * from both the config and the status section. */
75  if (pe__is_guest_or_remote_node(node) &&
76  node->details->remote_rsc &&
77  node->details->remote_rsc->container == NULL &&
78  pcmk_is_set(node->details->remote_rsc->flags,
79  pe_rsc_orphan_container_filler)) {
80  return TRUE;
81  }
82 
83  return FALSE;
84 }
85 
94 void
95 pe_fence_node(pe_working_set_t * data_set, pe_node_t * node,
96  const char *reason, bool priority_delay)
97 {
98  CRM_CHECK(node, return);
99 
100  /* A guest node is fenced by marking its container as failed */
101  if (pe__is_guest_node(node)) {
102  pe_resource_t *rsc = node->details->remote_rsc->container;
103 
104  if (!pcmk_is_set(rsc->flags, pe_rsc_failed)) {
105  if (!pcmk_is_set(rsc->flags, pe_rsc_managed)) {
106  crm_notice("Not fencing guest node %s "
107  "(otherwise would because %s): "
108  "its guest resource %s is unmanaged",
109  pe__node_name(node), reason, rsc->id);
110  } else {
111  crm_warn("Guest node %s will be fenced "
112  "(by recovering its guest resource %s): %s",
113  pe__node_name(node), rsc->id, reason);
114 
115  /* We don't mark the node as unclean because that would prevent the
116  * node from running resources. We want to allow it to run resources
117  * in this transition if the recovery succeeds.
118  */
119  node->details->remote_requires_reset = TRUE;
120  pe__set_resource_flags(rsc, pe_rsc_failed|pe_rsc_stop);
121  }
122  }
123 
124  } else if (is_dangling_guest_node(node)) {
125  crm_info("Cleaning up dangling connection for guest node %s: "
126  "fencing was already done because %s, "
127  "and guest resource no longer exists",
128  pe__node_name(node), reason);
129  pe__set_resource_flags(node->details->remote_rsc,
130  pe_rsc_failed|pe_rsc_stop);
131 
132  } else if (pe__is_remote_node(node)) {
133  pe_resource_t *rsc = node->details->remote_rsc;
134 
135  if ((rsc != NULL) && !pcmk_is_set(rsc->flags, pe_rsc_managed)) {
136  crm_notice("Not fencing remote node %s "
137  "(otherwise would because %s): connection is unmanaged",
138  pe__node_name(node), reason);
139  } else if(node->details->remote_requires_reset == FALSE) {
140  node->details->remote_requires_reset = TRUE;
141  crm_warn("Remote node %s %s: %s",
142  pe__node_name(node),
143  pe_can_fence(data_set, node)? "will be fenced" : "is unclean",
144  reason);
145  }
146  node->details->unclean = TRUE;
147  // No need to apply `priority-fencing-delay` for remote nodes
148  pe_fence_op(node, NULL, TRUE, reason, FALSE, data_set);
149 
150  } else if (node->details->unclean) {
151  crm_trace("Cluster node %s %s because %s",
152  pe__node_name(node),
153  pe_can_fence(data_set, node)? "would also be fenced" : "also is unclean",
154  reason);
155 
156  } else {
157  crm_warn("Cluster node %s %s: %s",
158  pe__node_name(node),
159  pe_can_fence(data_set, node)? "will be fenced" : "is unclean",
160  reason);
161  node->details->unclean = TRUE;
162  pe_fence_op(node, NULL, TRUE, reason, priority_delay, data_set);
163  }
164 }
165 
166 // @TODO xpaths can't handle templates, rules, or id-refs
167 
168 // nvpair with provides or requires set to unfencing
169 #define XPATH_UNFENCING_NVPAIR XML_CIB_TAG_NVPAIR \
170  "[(@" XML_NVPAIR_ATTR_NAME "='" PCMK_STONITH_PROVIDES "'" \
171  "or @" XML_NVPAIR_ATTR_NAME "='" XML_RSC_ATTR_REQUIRES "') " \
172  "and @" XML_NVPAIR_ATTR_VALUE "='" PCMK__VALUE_UNFENCING "']"
173 
174 // unfencing in rsc_defaults or any resource
175 #define XPATH_ENABLE_UNFENCING \
176  "/" XML_TAG_CIB "/" XML_CIB_TAG_CONFIGURATION "/" XML_CIB_TAG_RESOURCES \
177  "//" XML_TAG_META_SETS "/" XPATH_UNFENCING_NVPAIR \
178  "|/" XML_TAG_CIB "/" XML_CIB_TAG_CONFIGURATION "/" XML_CIB_TAG_RSCCONFIG \
179  "/" XML_TAG_META_SETS "/" XPATH_UNFENCING_NVPAIR
180 
181 static void
182 set_if_xpath(uint64_t flag, const char *xpath, pe_working_set_t *data_set)
183 {
184  xmlXPathObjectPtr result = NULL;
185 
186  if (!pcmk_is_set(data_set->flags, flag)) {
187  result = xpath_search(data_set->input, xpath);
188  if (result && (numXpathResults(result) > 0)) {
189  pe__set_working_set_flags(data_set, flag);
190  }
191  freeXpathObject(result);
192  }
193 }
194 
195 gboolean
196 unpack_config(xmlNode * config, pe_working_set_t * data_set)
197 {
198  const char *value = NULL;
199  GHashTable *config_hash = pcmk__strkey_table(free, free);
200 
201  pe_rule_eval_data_t rule_data = {
202  .node_hash = NULL,
203  .role = RSC_ROLE_UNKNOWN,
204  .now = data_set->now,
205  .match_data = NULL,
206  .rsc_data = NULL,
207  .op_data = NULL
208  };
209 
210  data_set->config_hash = config_hash;
211 
212  pe__unpack_dataset_nvpairs(config, XML_CIB_TAG_PROPSET, &rule_data, config_hash,
213  CIB_OPTIONS_FIRST, FALSE, data_set);
214 
215  verify_pe_options(data_set->config_hash);
216 
217  set_config_flag(data_set, "enable-startup-probes", pe_flag_startup_probes);
218  if (!pcmk_is_set(data_set->flags, pe_flag_startup_probes)) {
219  crm_info("Startup probes: disabled (dangerous)");
220  }
221 
222  value = pe_pref(data_set->config_hash, XML_ATTR_HAVE_WATCHDOG);
223  if (value && crm_is_true(value)) {
224  crm_info("Watchdog-based self-fencing will be performed via SBD if "
225  "fencing is required and stonith-watchdog-timeout is nonzero");
226  pe__set_working_set_flags(data_set, pe_flag_have_stonith_resource);
227  }
228 
229  /* Set certain flags via xpath here, so they can be used before the relevant
230  * configuration sections are unpacked.
231  */
232  set_if_xpath(pe_flag_enable_unfencing, XPATH_ENABLE_UNFENCING, data_set);
233 
234  value = pe_pref(data_set->config_hash, "stonith-timeout");
235  data_set->stonith_timeout = (int) crm_parse_interval_spec(value);
236  crm_debug("STONITH timeout: %d", data_set->stonith_timeout);
237 
238  set_config_flag(data_set, "stonith-enabled", pe_flag_stonith_enabled);
239  crm_debug("STONITH of failed nodes is %s",
240  pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)? "enabled" : "disabled");
241 
242  data_set->stonith_action = pe_pref(data_set->config_hash, "stonith-action");
243  if (!strcmp(data_set->stonith_action, "poweroff")) {
244  pe_warn_once(pe_wo_poweroff,
245  "Support for stonith-action of 'poweroff' is deprecated "
246  "and will be removed in a future release (use 'off' instead)");
247  data_set->stonith_action = "off";
248  }
249  crm_trace("STONITH will %s nodes", data_set->stonith_action);
250 
251  set_config_flag(data_set, "concurrent-fencing", pe_flag_concurrent_fencing);
252  crm_debug("Concurrent fencing is %s",
253  pcmk_is_set(data_set->flags, pe_flag_concurrent_fencing)? "enabled" : "disabled");
254 
255  value = pe_pref(data_set->config_hash,
256  XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY);
257  if (value) {
258  data_set->priority_fencing_delay = crm_parse_interval_spec(value) / 1000;
259  crm_trace("Priority fencing delay is %ds", data_set->priority_fencing_delay);
260  }
261 
262  set_config_flag(data_set, "stop-all-resources", pe_flag_stop_everything);
263  crm_debug("Stop all active resources: %s",
264  pcmk__btoa(pcmk_is_set(data_set->flags, pe_flag_stop_everything)));
265 
266  set_config_flag(data_set, "symmetric-cluster", pe_flag_symmetric_cluster);
267  if (pcmk_is_set(data_set->flags, pe_flag_symmetric_cluster)) {
268  crm_debug("Cluster is symmetric" " - resources can run anywhere by default");
269  }
270 
271  value = pe_pref(data_set->config_hash, "no-quorum-policy");
272 
273  if (pcmk__str_eq(value, "ignore", pcmk__str_casei)) {
274  data_set->no_quorum_policy = no_quorum_ignore;
275 
276  } else if (pcmk__str_eq(value, "freeze", pcmk__str_casei)) {
277  data_set->no_quorum_policy = no_quorum_freeze;
278 
279  } else if (pcmk__str_eq(value, "demote", pcmk__str_casei)) {
280  data_set->no_quorum_policy = no_quorum_demote;
281 
282  } else if (pcmk__str_eq(value, "suicide", pcmk__str_casei)) {
283  if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
284  int do_panic = 0;
285 
286  crm_element_value_int(data_set->input, XML_ATTR_QUORUM_PANIC,
287  &do_panic);
288  if (do_panic || pcmk_is_set(data_set->flags, pe_flag_have_quorum)) {
289  data_set->no_quorum_policy = no_quorum_suicide;
290  } else {
291  crm_notice("Resetting no-quorum-policy to 'stop': cluster has never had quorum");
292  data_set->no_quorum_policy = no_quorum_stop;
293  }
294  } else {
295  pcmk__config_err("Resetting no-quorum-policy to 'stop' because "
296  "fencing is disabled");
297  data_set->no_quorum_policy = no_quorum_stop;
298  }
299 
300  } else {
301  data_set->no_quorum_policy = no_quorum_stop;
302  }
303 
304  switch (data_set->no_quorum_policy) {
305  case no_quorum_freeze:
306  crm_debug("On loss of quorum: Freeze resources");
307  break;
308  case no_quorum_stop:
309  crm_debug("On loss of quorum: Stop ALL resources");
310  break;
311  case no_quorum_demote:
312  crm_debug("On loss of quorum: "
313  "Demote promotable resources and stop other resources");
314  break;
315  case no_quorum_suicide:
316  crm_notice("On loss of quorum: Fence all remaining nodes");
317  break;
318  case no_quorum_ignore:
319  crm_notice("On loss of quorum: Ignore");
320  break;
321  }
322 
323  set_config_flag(data_set, "stop-orphan-resources", pe_flag_stop_rsc_orphans);
324  crm_trace("Orphan resources are %s",
325  pcmk_is_set(data_set->flags, pe_flag_stop_rsc_orphans)? "stopped" : "ignored");
326 
327  set_config_flag(data_set, "stop-orphan-actions", pe_flag_stop_action_orphans);
328  crm_trace("Orphan resource actions are %s",
329  pcmk_is_set(data_set->flags, pe_flag_stop_action_orphans)? "stopped" : "ignored");
330 
331  value = pe_pref(data_set->config_hash, "remove-after-stop");
332  if (value != NULL) {
333  if (crm_is_true(value)) {
334  pe__set_working_set_flags(data_set, pe_flag_remove_after_stop);
335 #ifndef PCMK__COMPAT_2_0
336  pe_warn_once(pe_wo_remove_after,
337  "Support for the remove-after-stop cluster property is"
338  " deprecated and will be removed in a future release");
339 #endif
340  } else {
341  pe__clear_working_set_flags(data_set, pe_flag_remove_after_stop);
342  }
343  }
344 
345  set_config_flag(data_set, "maintenance-mode", pe_flag_maintenance_mode);
346  crm_trace("Maintenance mode: %s",
347  pcmk__btoa(pcmk_is_set(data_set->flags, pe_flag_maintenance_mode)));
348 
349  set_config_flag(data_set, "start-failure-is-fatal", pe_flag_start_failure_fatal);
350  crm_trace("Start failures are %s",
351  pcmk_is_set(data_set->flags, pe_flag_start_failure_fatal)? "always fatal" : "handled by failcount");
352 
353  if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
354  set_config_flag(data_set, "startup-fencing", pe_flag_startup_fencing);
355  }
356  if (pcmk_is_set(data_set->flags, pe_flag_startup_fencing)) {
357  crm_trace("Unseen nodes will be fenced");
358  } else {
359  pe_warn_once(pe_wo_blind, "Blind faith: not fencing unseen nodes");
360  }
361 
362  pe__unpack_node_health_scores(data_set);
363 
364  data_set->placement_strategy = pe_pref(data_set->config_hash, "placement-strategy");
365  crm_trace("Placement strategy: %s", data_set->placement_strategy);
366 
367  set_config_flag(data_set, "shutdown-lock", pe_flag_shutdown_lock);
368  crm_trace("Resources will%s be locked to cleanly shut down nodes",
369  (pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)? "" : " not"));
370  if (pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)) {
371  value = pe_pref(data_set->config_hash,
372  XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT);
373  data_set->shutdown_lock = crm_parse_interval_spec(value) / 1000;
374  crm_trace("Shutdown locks expire after %us", data_set->shutdown_lock);
375  }
376 
377  return TRUE;
378 }
379 
380 pe_node_t *
381 pe_create_node(const char *id, const char *uname, const char *type,
382  const char *score, pe_working_set_t * data_set)
383 {
384  pe_node_t *new_node = NULL;
385 
386  if (pe_find_node(data_set->nodes, uname) != NULL) {
387  pcmk__config_warn("More than one node entry has name '%s'", uname);
388  }
389 
390  new_node = calloc(1, sizeof(pe_node_t));
391  if (new_node == NULL) {
392  return NULL;
393  }
394 
395  new_node->weight = char2score(score);
396  new_node->fixed = FALSE;
397  new_node->details = calloc(1, sizeof(struct pe_node_shared_s));
398 
399  if (new_node->details == NULL) {
400  free(new_node);
401  return NULL;
402  }
403 
404  crm_trace("Creating node for entry %s/%s", uname, id);
405  new_node->details->id = id;
406  new_node->details->uname = uname;
407  new_node->details->online = FALSE;
408  new_node->details->shutdown = FALSE;
409  new_node->details->rsc_discovery_enabled = TRUE;
410  new_node->details->running_rsc = NULL;
411  new_node->details->data_set = data_set;
412 
413  if (pcmk__str_eq(type, "member", pcmk__str_null_matches | pcmk__str_casei)) {
414  new_node->details->type = node_member;
415 
416  } else if (pcmk__str_eq(type, "remote", pcmk__str_casei)) {
417  new_node->details->type = node_remote;
418  pe__set_working_set_flags(data_set, pe_flag_have_remote_nodes);
419 
420  } else {
421  /* @COMPAT 'ping' is the default for backward compatibility, but it
422  * should be changed to 'member' at a compatibility break
423  */
424  if (!pcmk__str_eq(type, "ping", pcmk__str_casei)) {
425  pcmk__config_warn("Node %s has unrecognized type '%s', "
426  "assuming 'ping'", pcmk__s(uname, "without name"),
427  type);
428  }
429  pe_warn_once(pe_wo_ping_node,
430  "Support for nodes of type 'ping' (such as %s) is "
431  "deprecated and will be removed in a future release",
432  pcmk__s(uname, "unnamed node"));
433  new_node->details->type = node_ping;
434  }
435 
436  new_node->details->attrs = pcmk__strkey_table(free, free);
437 
438  if (pe__is_guest_or_remote_node(new_node)) {
439  g_hash_table_insert(new_node->details->attrs, strdup(CRM_ATTR_KIND),
440  strdup("remote"));
441  } else {
442  g_hash_table_insert(new_node->details->attrs, strdup(CRM_ATTR_KIND),
443  strdup("cluster"));
444  }
445 
446  new_node->details->utilization = pcmk__strkey_table(free, free);
447  new_node->details->digest_cache = pcmk__strkey_table(free,
448  pe__free_digests);
449 
450  data_set->nodes = g_list_insert_sorted(data_set->nodes, new_node,
451  pe__cmp_node_name);
452  return new_node;
453 }
454 
455 static const char *
456 expand_remote_rsc_meta(xmlNode *xml_obj, xmlNode *parent, pe_working_set_t *data)
457 {
458  xmlNode *attr_set = NULL;
459  xmlNode *attr = NULL;
460 
461  const char *container_id = ID(xml_obj);
462  const char *remote_name = NULL;
463  const char *remote_server = NULL;
464  const char *remote_port = NULL;
465  const char *connect_timeout = "60s";
466  const char *remote_allow_migrate=NULL;
467  const char *is_managed = NULL;
468 
469  for (attr_set = pcmk__xe_first_child(xml_obj); attr_set != NULL;
470  attr_set = pcmk__xe_next(attr_set)) {
471 
472  if (!pcmk__str_eq((const char *)attr_set->name, XML_TAG_META_SETS,
473  pcmk__str_casei)) {
474  continue;
475  }
476 
477  for (attr = pcmk__xe_first_child(attr_set); attr != NULL;
478  attr = pcmk__xe_next(attr)) {
479  const char *value = crm_element_value(attr, XML_NVPAIR_ATTR_VALUE);
480  const char *name = crm_element_value(attr, XML_NVPAIR_ATTR_NAME);
481 
482  if (pcmk__str_eq(name, XML_RSC_ATTR_REMOTE_NODE, pcmk__str_casei)) {
483  remote_name = value;
484  } else if (pcmk__str_eq(name, "remote-addr", pcmk__str_casei)) {
485  remote_server = value;
486  } else if (pcmk__str_eq(name, "remote-port", pcmk__str_casei)) {
487  remote_port = value;
488  } else if (pcmk__str_eq(name, "remote-connect-timeout", pcmk__str_casei)) {
489  connect_timeout = value;
490  } else if (pcmk__str_eq(name, "remote-allow-migrate", pcmk__str_casei)) {
491  remote_allow_migrate=value;
492  } else if (pcmk__str_eq(name, XML_RSC_ATTR_MANAGED, pcmk__str_casei)) {
493  is_managed = value;
494  }
495  }
496  }
497 
498  if (remote_name == NULL) {
499  return NULL;
500  }
501 
502  if (pe_find_resource(data->resources, remote_name) != NULL) {
503  return NULL;
504  }
505 
506  pe_create_remote_xml(parent, remote_name, container_id,
507  remote_allow_migrate, is_managed,
508  connect_timeout, remote_server, remote_port);
509  return remote_name;
510 }
511 
512 static void
513 handle_startup_fencing(pe_working_set_t *data_set, pe_node_t *new_node)
514 {
515  if ((new_node->details->type == node_remote) && (new_node->details->remote_rsc == NULL)) {
516  /* Ignore fencing for remote nodes that don't have a connection resource
517  * associated with them. This happens when remote node entries get left
518  * in the nodes section after the connection resource is removed.
519  */
520  return;
521  }
522 
523  if (pcmk_is_set(data_set->flags, pe_flag_startup_fencing)) {
524  // All nodes are unclean until we've seen their status entry
525  new_node->details->unclean = TRUE;
526 
527  } else {
528  // Blind faith ...
529  new_node->details->unclean = FALSE;
530  }
531 
532  /* We need to be able to determine if a node's status section
533  * exists or not separate from whether the node is unclean. */
534  new_node->details->unseen = TRUE;
535 }
536 
537 gboolean
538 unpack_nodes(xmlNode * xml_nodes, pe_working_set_t * data_set)
539 {
540  xmlNode *xml_obj = NULL;
541  pe_node_t *new_node = NULL;
542  const char *id = NULL;
543  const char *uname = NULL;
544  const char *type = NULL;
545  const char *score = NULL;
546 
547  pe_rule_eval_data_t rule_data = {
548  .node_hash = NULL,
549  .role = RSC_ROLE_UNKNOWN,
550  .now = data_set->now,
551  .match_data = NULL,
552  .rsc_data = NULL,
553  .op_data = NULL
554  };
555 
556  for (xml_obj = pcmk__xe_first_child(xml_nodes); xml_obj != NULL;
557  xml_obj = pcmk__xe_next(xml_obj)) {
558 
559  if (pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_NODE, pcmk__str_none)) {
560  new_node = NULL;
561 
562  id = crm_element_value(xml_obj, XML_ATTR_ID);
563  uname = crm_element_value(xml_obj, XML_ATTR_UNAME);
564  type = crm_element_value(xml_obj, XML_ATTR_TYPE);
565  score = crm_element_value(xml_obj, XML_RULE_ATTR_SCORE);
566  crm_trace("Processing node %s/%s", uname, id);
567 
568  if (id == NULL) {
569  pcmk__config_err("Ignoring <" XML_CIB_TAG_NODE
570  "> entry in configuration without id");
571  continue;
572  }
573  new_node = pe_create_node(id, uname, type, score, data_set);
574 
575  if (new_node == NULL) {
576  return FALSE;
577  }
578 
579  handle_startup_fencing(data_set, new_node);
580 
581  add_node_attrs(xml_obj, new_node, FALSE, data_set);
582  pe__unpack_dataset_nvpairs(xml_obj, XML_TAG_UTILIZATION, &rule_data,
583  new_node->details->utilization, NULL,
584  FALSE, data_set);
585 
586  crm_trace("Done with node %s", crm_element_value(xml_obj, XML_ATTR_UNAME));
587  }
588  }
589 
590  if (data_set->localhost && pe_find_node(data_set->nodes, data_set->localhost) == NULL) {
591  crm_info("Creating a fake local node");
592  pe_create_node(data_set->localhost, data_set->localhost, NULL, 0,
593  data_set);
594  }
595 
596  return TRUE;
597 }
598 
599 static void
600 setup_container(pe_resource_t * rsc, pe_working_set_t * data_set)
601 {
602  const char *container_id = NULL;
603 
604  if (rsc->children) {
605  g_list_foreach(rsc->children, (GFunc) setup_container, data_set);
606  return;
607  }
608 
609  container_id = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_CONTAINER);
610  if (container_id && !pcmk__str_eq(container_id, rsc->id, pcmk__str_casei)) {
611  pe_resource_t *container = pe_find_resource(data_set->resources, container_id);
612 
613  if (container) {
614  rsc->container = container;
615  pe__set_resource_flags(container, pe_rsc_is_container);
616  container->fillers = g_list_append(container->fillers, rsc);
617  pe_rsc_trace(rsc, "Resource %s's container is %s", rsc->id, container_id);
618  } else {
619  pe_err("Resource %s: Unknown resource container (%s)", rsc->id, container_id);
620  }
621  }
622 }
623 
624 gboolean
625 unpack_remote_nodes(xmlNode * xml_resources, pe_working_set_t * data_set)
626 {
627  xmlNode *xml_obj = NULL;
628 
629  /* Create remote nodes and guest nodes from the resource configuration
630  * before unpacking resources.
631  */
632  for (xml_obj = pcmk__xe_first_child(xml_resources); xml_obj != NULL;
633  xml_obj = pcmk__xe_next(xml_obj)) {
634 
635  const char *new_node_id = NULL;
636 
637  /* Check for remote nodes, which are defined by ocf:pacemaker:remote
638  * primitives.
639  */
640  if (xml_contains_remote_node(xml_obj)) {
641  new_node_id = ID(xml_obj);
642  /* The "pe_find_node" check is here to make sure we don't iterate over
643  * an expanded node that has already been added to the node list. */
644  if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) {
645  crm_trace("Found remote node %s defined by resource %s",
646  new_node_id, ID(xml_obj));
647  pe_create_node(new_node_id, new_node_id, "remote", NULL,
648  data_set);
649  }
650  continue;
651  }
652 
653  /* Check for guest nodes, which are defined by special meta-attributes
654  * of a primitive of any type (for example, VirtualDomain or Xen).
655  */
656  if (pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_RESOURCE, pcmk__str_none)) {
657  /* This will add an ocf:pacemaker:remote primitive to the
658  * configuration for the guest node's connection, to be unpacked
659  * later.
660  */
661  new_node_id = expand_remote_rsc_meta(xml_obj, xml_resources, data_set);
662  if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) {
663  crm_trace("Found guest node %s in resource %s",
664  new_node_id, ID(xml_obj));
665  pe_create_node(new_node_id, new_node_id, "remote", NULL,
666  data_set);
667  }
668  continue;
669  }
670 
671  /* Check for guest nodes inside a group. Clones are currently not
672  * supported as guest nodes.
673  */
674  if (pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_GROUP, pcmk__str_none)) {
675  xmlNode *xml_obj2 = NULL;
676  for (xml_obj2 = pcmk__xe_first_child(xml_obj); xml_obj2 != NULL;
677  xml_obj2 = pcmk__xe_next(xml_obj2)) {
678 
679  new_node_id = expand_remote_rsc_meta(xml_obj2, xml_resources, data_set);
680 
681  if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) {
682  crm_trace("Found guest node %s in resource %s inside group %s",
683  new_node_id, ID(xml_obj2), ID(xml_obj));
684  pe_create_node(new_node_id, new_node_id, "remote", NULL,
685  data_set);
686  }
687  }
688  }
689  }
690  return TRUE;
691 }
692 
693 /* Call this after all the nodes and resources have been
694  * unpacked, but before the status section is read.
695  *
696  * A remote node's online status is reflected by the state
697  * of the remote node's connection resource. We need to link
698  * the remote node to this connection resource so we can have
699  * easy access to the connection resource during the scheduler calculations.
700  */
701 static void
702 link_rsc2remotenode(pe_working_set_t *data_set, pe_resource_t *new_rsc)
703 {
704  pe_node_t *remote_node = NULL;
705 
706  if (new_rsc->is_remote_node == FALSE) {
707  return;
708  }
709 
710  if (pcmk_is_set(data_set->flags, pe_flag_quick_location)) {
711  /* remote_nodes and remote_resources are not linked in quick location calculations */
712  return;
713  }
714 
715  remote_node = pe_find_node(data_set->nodes, new_rsc->id);
716  CRM_CHECK(remote_node != NULL, return);
717 
718  pe_rsc_trace(new_rsc, "Linking remote connection resource %s to %s",
719  new_rsc->id, pe__node_name(remote_node));
720  remote_node->details->remote_rsc = new_rsc;
721 
722  if (new_rsc->container == NULL) {
723  /* Handle start-up fencing for remote nodes (as opposed to guest nodes)
724  * the same as is done for cluster nodes.
725  */
726  handle_startup_fencing(data_set, remote_node);
727 
728  } else {
729  /* pe_create_node() marks the new node as "remote" or "cluster"; now
730  * that we know the node is a guest node, update it correctly.
731  */
732  g_hash_table_replace(remote_node->details->attrs, strdup(CRM_ATTR_KIND),
733  strdup("container"));
734  }
735 }
736 
737 static void
738 destroy_tag(gpointer data)
739 {
740  pe_tag_t *tag = data;
741 
742  if (tag) {
743  free(tag->id);
744  g_list_free_full(tag->refs, free);
745  free(tag);
746  }
747 }
748 
761 gboolean
762 unpack_resources(xmlNode * xml_resources, pe_working_set_t * data_set)
763 {
764  xmlNode *xml_obj = NULL;
765  GList *gIter = NULL;
766 
767  data_set->template_rsc_sets = pcmk__strkey_table(free, destroy_tag);
768 
769  for (xml_obj = pcmk__xe_first_child(xml_resources); xml_obj != NULL;
770  xml_obj = pcmk__xe_next(xml_obj)) {
771 
772  pe_resource_t *new_rsc = NULL;
773  const char *id = ID(xml_obj);
774 
775  if (pcmk__str_empty(id)) {
776  pcmk__config_err("Ignoring <%s> resource without ID",
777  crm_element_name(xml_obj));
778  continue;
779  }
780 
781  if (pcmk__str_eq((const char *) xml_obj->name, XML_CIB_TAG_RSC_TEMPLATE,
782  pcmk__str_none)) {
783  if (g_hash_table_lookup_extended(data_set->template_rsc_sets, id,
784  NULL, NULL) == FALSE) {
785  /* Record the template's ID for the knowledge of its existence anyway. */
786  g_hash_table_insert(data_set->template_rsc_sets, strdup(id), NULL);
787  }
788  continue;
789  }
790 
791  crm_trace("Unpacking <%s id='%s'>", crm_element_name(xml_obj), id);
792  if (pe__unpack_resource(xml_obj, &new_rsc, NULL,
793  data_set) == pcmk_rc_ok) {
794  data_set->resources = g_list_append(data_set->resources, new_rsc);
795  pe_rsc_trace(new_rsc, "Added resource %s", new_rsc->id);
796 
797  } else {
798  pcmk__config_err("Ignoring <%s> resource '%s' "
799  "because configuration is invalid",
800  crm_element_name(xml_obj), id);
801  }
802  }
803 
804  for (gIter = data_set->resources; gIter != NULL; gIter = gIter->next) {
805  pe_resource_t *rsc = (pe_resource_t *) gIter->data;
806 
807  setup_container(rsc, data_set);
808  link_rsc2remotenode(data_set, rsc);
809  }
810 
811  data_set->resources = g_list_sort(data_set->resources,
812  pe__cmp_rsc_priority);
813  if (pcmk_is_set(data_set->flags, pe_flag_quick_location)) {
814  /* Ignore */
815 
816  } else if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)
817  && !pcmk_is_set(data_set->flags, pe_flag_have_stonith_resource)) {
818 
819  pcmk__config_err("Resource start-up disabled since no STONITH resources have been defined");
820  pcmk__config_err("Either configure some or disable STONITH with the stonith-enabled option");
821  pcmk__config_err("NOTE: Clusters with shared data need STONITH to ensure data integrity");
822  }
823 
824  return TRUE;
825 }
826 
827 gboolean
828 unpack_tags(xmlNode * xml_tags, pe_working_set_t * data_set)
829 {
830  xmlNode *xml_tag = NULL;
831 
832  data_set->tags = pcmk__strkey_table(free, destroy_tag);
833 
834  for (xml_tag = pcmk__xe_first_child(xml_tags); xml_tag != NULL;
835  xml_tag = pcmk__xe_next(xml_tag)) {
836 
837  xmlNode *xml_obj_ref = NULL;
838  const char *tag_id = ID(xml_tag);
839 
840  if (!pcmk__str_eq((const char *)xml_tag->name, XML_CIB_TAG_TAG, pcmk__str_none)) {
841  continue;
842  }
843 
844  if (tag_id == NULL) {
845  pcmk__config_err("Ignoring <%s> without " XML_ATTR_ID,
846  crm_element_name(xml_tag));
847  continue;
848  }
849 
850  for (xml_obj_ref = pcmk__xe_first_child(xml_tag); xml_obj_ref != NULL;
851  xml_obj_ref = pcmk__xe_next(xml_obj_ref)) {
852 
853  const char *obj_ref = ID(xml_obj_ref);
854 
855  if (!pcmk__str_eq((const char *)xml_obj_ref->name, XML_CIB_TAG_OBJ_REF, pcmk__str_none)) {
856  continue;
857  }
858 
859  if (obj_ref == NULL) {
860  pcmk__config_err("Ignoring <%s> for tag '%s' without " XML_ATTR_ID,
861  crm_element_name(xml_obj_ref), tag_id);
862  continue;
863  }
864 
865  if (add_tag_ref(data_set->tags, tag_id, obj_ref) == FALSE) {
866  return FALSE;
867  }
868  }
869  }
870 
871  return TRUE;
872 }
873 
874 /* The ticket state section:
875  * "/cib/status/tickets/ticket_state" */
876 static gboolean
877 unpack_ticket_state(xmlNode * xml_ticket, pe_working_set_t * data_set)
878 {
879  const char *ticket_id = NULL;
880  const char *granted = NULL;
881  const char *last_granted = NULL;
882  const char *standby = NULL;
883  xmlAttrPtr xIter = NULL;
884 
885  pe_ticket_t *ticket = NULL;
886 
887  ticket_id = ID(xml_ticket);
888  if (pcmk__str_empty(ticket_id)) {
889  return FALSE;
890  }
891 
892  crm_trace("Processing ticket state for %s", ticket_id);
893 
894  ticket = g_hash_table_lookup(data_set->tickets, ticket_id);
895  if (ticket == NULL) {
896  ticket = ticket_new(ticket_id, data_set);
897  if (ticket == NULL) {
898  return FALSE;
899  }
900  }
901 
902  for (xIter = xml_ticket->properties; xIter; xIter = xIter->next) {
903  const char *prop_name = (const char *)xIter->name;
904  const char *prop_value = crm_element_value(xml_ticket, prop_name);
905 
906  if (pcmk__str_eq(prop_name, XML_ATTR_ID, pcmk__str_none)) {
907  continue;
908  }
909  g_hash_table_replace(ticket->state, strdup(prop_name), strdup(prop_value));
910  }
911 
912  granted = g_hash_table_lookup(ticket->state, "granted");
913  if (granted && crm_is_true(granted)) {
914  ticket->granted = TRUE;
915  crm_info("We have ticket '%s'", ticket->id);
916  } else {
917  ticket->granted = FALSE;
918  crm_info("We do not have ticket '%s'", ticket->id);
919  }
920 
921  last_granted = g_hash_table_lookup(ticket->state, "last-granted");
922  if (last_granted) {
923  long long last_granted_ll;
924 
925  pcmk__scan_ll(last_granted, &last_granted_ll, 0LL);
926  ticket->last_granted = (time_t) last_granted_ll;
927  }
928 
929  standby = g_hash_table_lookup(ticket->state, "standby");
930  if (standby && crm_is_true(standby)) {
931  ticket->standby = TRUE;
932  if (ticket->granted) {
933  crm_info("Granted ticket '%s' is in standby-mode", ticket->id);
934  }
935  } else {
936  ticket->standby = FALSE;
937  }
938 
939  crm_trace("Done with ticket state for %s", ticket_id);
940 
941  return TRUE;
942 }
943 
944 static gboolean
945 unpack_tickets_state(xmlNode * xml_tickets, pe_working_set_t * data_set)
946 {
947  xmlNode *xml_obj = NULL;
948 
949  for (xml_obj = pcmk__xe_first_child(xml_tickets); xml_obj != NULL;
950  xml_obj = pcmk__xe_next(xml_obj)) {
951 
952  if (!pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_TICKET_STATE, pcmk__str_none)) {
953  continue;
954  }
955  unpack_ticket_state(xml_obj, data_set);
956  }
957 
958  return TRUE;
959 }
960 
961 static void
962 unpack_handle_remote_attrs(pe_node_t *this_node, xmlNode *state, pe_working_set_t * data_set)
963 {
964  const char *resource_discovery_enabled = NULL;
965  xmlNode *attrs = NULL;
966  pe_resource_t *rsc = NULL;
967 
968  if (!pcmk__str_eq((const char *)state->name, XML_CIB_TAG_STATE, pcmk__str_none)) {
969  return;
970  }
971 
972  if ((this_node == NULL) || !pe__is_guest_or_remote_node(this_node)) {
973  return;
974  }
975  crm_trace("Processing Pacemaker Remote node %s", pe__node_name(this_node));
976 
977  pcmk__scan_min_int(crm_element_value(state, XML_NODE_IS_MAINTENANCE),
978  &(this_node->details->remote_maintenance), 0);
979 
980  rsc = this_node->details->remote_rsc;
981  if (this_node->details->remote_requires_reset == FALSE) {
982  this_node->details->unclean = FALSE;
983  this_node->details->unseen = FALSE;
984  }
985  attrs = find_xml_node(state, XML_TAG_TRANSIENT_NODEATTRS, FALSE);
986  add_node_attrs(attrs, this_node, TRUE, data_set);
987 
988  if (pe__shutdown_requested(this_node)) {
989  crm_info("%s is shutting down", pe__node_name(this_node));
990  this_node->details->shutdown = TRUE;
991  }
992 
993  if (crm_is_true(pe_node_attribute_raw(this_node, "standby"))) {
994  crm_info("%s is in standby mode", pe__node_name(this_node));
995  this_node->details->standby = TRUE;
996  }
997 
998  if (crm_is_true(pe_node_attribute_raw(this_node, "maintenance")) ||
999  ((rsc != NULL) && !pcmk_is_set(rsc->flags, pe_rsc_managed))) {
1000  crm_info("%s is in maintenance mode", pe__node_name(this_node));
1001  this_node->details->maintenance = TRUE;
1002  }
1003 
1004  resource_discovery_enabled = pe_node_attribute_raw(this_node, XML_NODE_ATTR_RSC_DISCOVERY);
1005  if (resource_discovery_enabled && !crm_is_true(resource_discovery_enabled)) {
1006  if (pe__is_remote_node(this_node)
1007  && !pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
1008  crm_warn("Ignoring " XML_NODE_ATTR_RSC_DISCOVERY
1009  " attribute on Pacemaker Remote node %s"
1010  " because fencing is disabled",
1011  pe__node_name(this_node));
1012  } else {
1013  /* This is either a remote node with fencing enabled, or a guest
1014  * node. We don't care whether fencing is enabled when fencing guest
1015  * nodes, because they are "fenced" by recovering their containing
1016  * resource.
1017  */
1018  crm_info("%s has resource discovery disabled",
1019  pe__node_name(this_node));
1020  this_node->details->rsc_discovery_enabled = FALSE;
1021  }
1022  }
1023 }
1024 
1033 static void
1034 unpack_transient_attributes(xmlNode *state, pe_node_t *node,
1035  pe_working_set_t *data_set)
1036 {
1037  const char *discovery = NULL;
1038  xmlNode *attrs = find_xml_node(state, XML_TAG_TRANSIENT_NODEATTRS, FALSE);
1039 
1040  add_node_attrs(attrs, node, TRUE, data_set);
1041 
1042  if (crm_is_true(pe_node_attribute_raw(node, "standby"))) {
1043  crm_info("%s is in standby mode", pe__node_name(node));
1044  node->details->standby = TRUE;
1045  }
1046 
1047  if (crm_is_true(pe_node_attribute_raw(node, "maintenance"))) {
1048  crm_info("%s is in maintenance mode", pe__node_name(node));
1049  node->details->maintenance = TRUE;
1050  }
1051 
1052  discovery = pe_node_attribute_raw(node, XML_NODE_ATTR_RSC_DISCOVERY);
1053  if ((discovery != NULL) && !crm_is_true(discovery)) {
1054  crm_warn("Ignoring " XML_NODE_ATTR_RSC_DISCOVERY
1055  " attribute for %s because disabling resource discovery "
1056  "is not allowed for cluster nodes", pe__node_name(node));
1057  }
1058 }
1059 
1072 static void
1073 unpack_node_state(xmlNode *state, pe_working_set_t *data_set)
1074 {
1075  const char *id = NULL;
1076  const char *uname = NULL;
1077  pe_node_t *this_node = NULL;
1078 
1079  id = crm_element_value(state, XML_ATTR_ID);
1080  if (id == NULL) {
1081  crm_warn("Ignoring malformed " XML_CIB_TAG_STATE " entry without "
1082  XML_ATTR_ID);
1083  return;
1084  }
1085 
1086  uname = crm_element_value(state, XML_ATTR_UNAME);
1087  if (uname == NULL) {
1088  crm_warn("Ignoring malformed " XML_CIB_TAG_STATE " entry without "
1089  XML_ATTR_UNAME);
1090  return;
1091  }
1092 
1093  this_node = pe_find_node_any(data_set->nodes, id, uname);
1094  if (this_node == NULL) {
1095  pcmk__config_warn("Ignoring recorded node state for '%s' because "
1096  "it is no longer in the configuration", uname);
1097  return;
1098  }
1099 
1100  if (pe__is_guest_or_remote_node(this_node)) {
1101  /* We can't determine the online status of Pacemaker Remote nodes until
1102  * after all resource history has been unpacked. In this first pass, we
1103  * do need to mark whether the node has been fenced, as this plays a
1104  * role during unpacking cluster node resource state.
1105  */
1106  pcmk__scan_min_int(crm_element_value(state, XML_NODE_IS_FENCED),
1107  &(this_node->details->remote_was_fenced), 0);
1108  return;
1109  }
1110 
1111  unpack_transient_attributes(state, this_node, data_set);
1112 
1113  /* Provisionally mark this cluster node as clean. We have at least seen it
1114  * in the current cluster's lifetime.
1115  */
1116  this_node->details->unclean = FALSE;
1117  this_node->details->unseen = FALSE;
1118 
1119  crm_trace("Determining online status of cluster node %s (id %s)",
1120  pe__node_name(this_node), id);
1121  determine_online_status(state, this_node, data_set);
1122 
1123  if (!pcmk_is_set(data_set->flags, pe_flag_have_quorum)
1124  && this_node->details->online
1125  && (data_set->no_quorum_policy == no_quorum_suicide)) {
1126  /* Everything else should flow from this automatically
1127  * (at least until the scheduler becomes able to migrate off
1128  * healthy resources)
1129  */
1130  pe_fence_node(data_set, this_node, "cluster does not have quorum",
1131  FALSE);
1132  }
1133 }
1134 
1152 static int
1153 unpack_node_history(xmlNode *status, bool fence, pe_working_set_t *data_set)
1154 {
1155  int rc = pcmk_rc_ok;
1156 
1157  // Loop through all node_state entries in CIB status
1158  for (xmlNode *state = first_named_child(status, XML_CIB_TAG_STATE);
1159  state != NULL; state = crm_next_same_xml(state)) {
1160 
1161  const char *id = ID(state);
1162  const char *uname = crm_element_value(state, XML_ATTR_UNAME);
1163  pe_node_t *this_node = NULL;
1164 
1165  if ((id == NULL) || (uname == NULL)) {
1166  // Warning already logged in first pass through status section
1167  crm_trace("Not unpacking resource history from malformed "
1168  XML_CIB_TAG_STATE " without id and/or uname");
1169  continue;
1170  }
1171 
1172  this_node = pe_find_node_any(data_set->nodes, id, uname);
1173  if (this_node == NULL) {
1174  // Warning already logged in first pass through status section
1175  crm_trace("Not unpacking resource history for node %s because "
1176  "no longer in configuration", id);
1177  continue;
1178  }
1179 
1180  if (this_node->details->unpacked) {
1181  crm_trace("Not unpacking resource history for node %s because "
1182  "already unpacked", id);
1183  continue;
1184  }
1185 
1186  if (fence) {
1187  // We're processing all remaining nodes
1188 
1189  } else if (pe__is_guest_node(this_node)) {
1190  /* We can unpack a guest node's history only after we've unpacked
1191  * other resource history to the point that we know that the node's
1192  * connection and containing resource are both up.
1193  */
1194  pe_resource_t *rsc = this_node->details->remote_rsc;
1195 
1196  if ((rsc == NULL) || (rsc->role != RSC_ROLE_STARTED)
1197  || (rsc->container->role != RSC_ROLE_STARTED)) {
1198  crm_trace("Not unpacking resource history for guest node %s "
1199  "because container and connection are not known to "
1200  "be up", id);
1201  continue;
1202  }
1203 
1204  } else if (pe__is_remote_node(this_node)) {
1205  /* We can unpack a remote node's history only after we've unpacked
1206  * other resource history to the point that we know that the node's
1207  * connection is up, with the exception of when shutdown locks are
1208  * in use.
1209  */
1210  pe_resource_t *rsc = this_node->details->remote_rsc;
1211 
1212  if ((rsc == NULL)
1213  || (!pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)
1214  && (rsc->role != RSC_ROLE_STARTED))) {
1215  crm_trace("Not unpacking resource history for remote node %s "
1216  "because connection is not known to be up", id);
1217  continue;
1218  }
1219 
1220  /* If fencing and shutdown locks are disabled and we're not processing
1221  * unseen nodes, then we don't want to unpack offline nodes until online
1222  * nodes have been unpacked. This allows us to number active clone
1223  * instances first.
1224  */
1225  } else if (!pcmk_any_flags_set(data_set->flags, pe_flag_stonith_enabled
1226  |pe_flag_shutdown_lock)
1227  && !this_node->details->online) {
1228  crm_trace("Not unpacking resource history for offline "
1229  "cluster node %s", id);
1230  continue;
1231  }
1232 
1233  if (pe__is_guest_or_remote_node(this_node)) {
1234  determine_remote_online_status(data_set, this_node);
1235  unpack_handle_remote_attrs(this_node, state, data_set);
1236  }
1237 
1238  crm_trace("Unpacking resource history for %snode %s",
1239  (fence? "unseen " : ""), id);
1240 
1241  this_node->details->unpacked = TRUE;
1242  unpack_node_lrm(this_node, state, data_set);
1243 
1244  rc = EAGAIN; // Other node histories might depend on this one
1245  }
1246  return rc;
1247 }
1248 
1249 /* remove nodes that are down, stopping */
1250 /* create positive rsc_to_node constraints between resources and the nodes they are running on */
1251 /* anything else? */
1252 gboolean
1253 unpack_status(xmlNode * status, pe_working_set_t * data_set)
1254 {
1255  xmlNode *state = NULL;
1256 
1257  crm_trace("Beginning unpack");
1258 
1259  if (data_set->tickets == NULL) {
1260  data_set->tickets = pcmk__strkey_table(free, destroy_ticket);
1261  }
1262 
1263  for (state = pcmk__xe_first_child(status); state != NULL;
1264  state = pcmk__xe_next(state)) {
1265 
1266  if (pcmk__str_eq((const char *)state->name, XML_CIB_TAG_TICKETS, pcmk__str_none)) {
1267  unpack_tickets_state((xmlNode *) state, data_set);
1268 
1269  } else if (pcmk__str_eq((const char *)state->name, XML_CIB_TAG_STATE, pcmk__str_none)) {
1270  unpack_node_state(state, data_set);
1271  }
1272  }
1273 
1274  while (unpack_node_history(status, FALSE, data_set) == EAGAIN) {
1275  crm_trace("Another pass through node resource histories is needed");
1276  }
1277 
1278  // Now catch any nodes we didn't see
1279  unpack_node_history(status,
1280  pcmk_is_set(data_set->flags, pe_flag_stonith_enabled),
1281  data_set);
1282 
1283  /* Now that we know where resources are, we can schedule stops of containers
1284  * with failed bundle connections
1285  */
1286  if (data_set->stop_needed != NULL) {
1287  for (GList *item = data_set->stop_needed; item; item = item->next) {
1288  pe_resource_t *container = item->data;
1289  pe_node_t *node = pe__current_node(container);
1290 
1291  if (node) {
1292  stop_action(container, node, FALSE);
1293  }
1294  }
1295  g_list_free(data_set->stop_needed);
1296  data_set->stop_needed = NULL;
1297  }
1298 
1299  /* Now that we know status of all Pacemaker Remote connections and nodes,
1300  * we can stop connections for node shutdowns, and check the online status
1301  * of remote/guest nodes that didn't have any node history to unpack.
1302  */
1303  for (GList *gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) {
1304  pe_node_t *this_node = gIter->data;
1305 
1306  if (!pe__is_guest_or_remote_node(this_node)) {
1307  continue;
1308  }
1309  if (this_node->details->shutdown
1310  && (this_node->details->remote_rsc != NULL)) {
1311  pe__set_next_role(this_node->details->remote_rsc, RSC_ROLE_STOPPED,
1312  "remote shutdown");
1313  }
1314  if (!this_node->details->unpacked) {
1315  determine_remote_online_status(data_set, this_node);
1316  }
1317  }
1318 
1319  return TRUE;
1320 }
1321 
1322 static gboolean
1323 determine_online_status_no_fencing(pe_working_set_t * data_set, xmlNode * node_state,
1324  pe_node_t * this_node)
1325 {
1326  gboolean online = FALSE;
1327  const char *join = crm_element_value(node_state, XML_NODE_JOIN_STATE);
1328  const char *is_peer = crm_element_value(node_state, XML_NODE_IS_PEER);
1329  const char *in_cluster = crm_element_value(node_state, XML_NODE_IN_CLUSTER);
1330  const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED);
1331 
1332  if (!crm_is_true(in_cluster)) {
1333  crm_trace("Node is down: in_cluster=%s",
1334  pcmk__s(in_cluster, "<null>"));
1335 
1336  } else if (pcmk__str_eq(is_peer, ONLINESTATUS, pcmk__str_casei)) {
1337  if (pcmk__str_eq(join, CRMD_JOINSTATE_MEMBER, pcmk__str_casei)) {
1338  online = TRUE;
1339  } else {
1340  crm_debug("Node is not ready to run resources: %s", join);
1341  }
1342 
1343  } else if (this_node->details->expected_up == FALSE) {
1344  crm_trace("Controller is down: "
1345  "in_cluster=%s is_peer=%s join=%s expected=%s",
1346  pcmk__s(in_cluster, "<null>"), pcmk__s(is_peer, "<null>"),
1347  pcmk__s(join, "<null>"), pcmk__s(exp_state, "<null>"));
1348 
1349  } else {
1350  /* mark it unclean */
1351  pe_fence_node(data_set, this_node, "peer is unexpectedly down", FALSE);
1352  crm_info("in_cluster=%s is_peer=%s join=%s expected=%s",
1353  pcmk__s(in_cluster, "<null>"), pcmk__s(is_peer, "<null>"),
1354  pcmk__s(join, "<null>"), pcmk__s(exp_state, "<null>"));
1355  }
1356  return online;
1357 }
1358 
1359 static gboolean
1360 determine_online_status_fencing(pe_working_set_t * data_set, xmlNode * node_state,
1361  pe_node_t * this_node)
1362 {
1363  gboolean online = FALSE;
1364  gboolean do_terminate = FALSE;
1365  bool crmd_online = FALSE;
1366  const char *join = crm_element_value(node_state, XML_NODE_JOIN_STATE);
1367  const char *is_peer = crm_element_value(node_state, XML_NODE_IS_PEER);
1368  const char *in_cluster = crm_element_value(node_state, XML_NODE_IN_CLUSTER);
1369  const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED);
1370  const char *terminate = pe_node_attribute_raw(this_node, "terminate");
1371 
1372 /*
1373  - XML_NODE_IN_CLUSTER ::= true|false
1374  - XML_NODE_IS_PEER ::= online|offline
1375  - XML_NODE_JOIN_STATE ::= member|down|pending|banned
1376  - XML_NODE_EXPECTED ::= member|down
1377 */
1378 
1379  if (crm_is_true(terminate)) {
1380  do_terminate = TRUE;
1381 
1382  } else if (terminate != NULL && strlen(terminate) > 0) {
1383  /* could be a time() value */
1384  char t = terminate[0];
1385 
1386  if (t != '0' && isdigit(t)) {
1387  do_terminate = TRUE;
1388  }
1389  }
1390 
1391  crm_trace("%s: in_cluster=%s is_peer=%s join=%s expected=%s term=%d",
1392  pe__node_name(this_node), pcmk__s(in_cluster, "<null>"),
1393  pcmk__s(is_peer, "<null>"), pcmk__s(join, "<null>"),
1394  pcmk__s(exp_state, "<null>"), do_terminate);
1395 
1396  online = crm_is_true(in_cluster);
1397  crmd_online = pcmk__str_eq(is_peer, ONLINESTATUS, pcmk__str_casei);
1398  if (exp_state == NULL) {
1399  exp_state = CRMD_JOINSTATE_DOWN;
1400  }
1401 
1402  if (this_node->details->shutdown) {
1403  crm_debug("%s is shutting down", pe__node_name(this_node));
1404 
1405  /* Slightly different criteria since we can't shut down a dead peer */
1406  online = crmd_online;
1407 
1408  } else if (in_cluster == NULL) {
1409  pe_fence_node(data_set, this_node, "peer has not been seen by the cluster", FALSE);
1410 
1411  } else if (pcmk__str_eq(join, CRMD_JOINSTATE_NACK, pcmk__str_casei)) {
1412  pe_fence_node(data_set, this_node,
1413  "peer failed Pacemaker membership criteria", FALSE);
1414 
1415  } else if (do_terminate == FALSE && pcmk__str_eq(exp_state, CRMD_JOINSTATE_DOWN, pcmk__str_casei)) {
1416 
1417  if (crm_is_true(in_cluster) || crmd_online) {
1418  crm_info("- %s is not ready to run resources",
1419  pe__node_name(this_node));
1420  this_node->details->standby = TRUE;
1421  this_node->details->pending = TRUE;
1422 
1423  } else {
1424  crm_trace("%s is down or still coming up",
1425  pe__node_name(this_node));
1426  }
1427 
1428  } else if (do_terminate && pcmk__str_eq(join, CRMD_JOINSTATE_DOWN, pcmk__str_casei)
1429  && crm_is_true(in_cluster) == FALSE && !crmd_online) {
1430  crm_info("%s was just shot", pe__node_name(this_node));
1431  online = FALSE;
1432 
1433  } else if (crm_is_true(in_cluster) == FALSE) {
1434  // Consider `priority-fencing-delay` for lost nodes
1435  pe_fence_node(data_set, this_node, "peer is no longer part of the cluster", TRUE);
1436 
1437  } else if (!crmd_online) {
1438  pe_fence_node(data_set, this_node, "peer process is no longer available", FALSE);
1439 
1440  /* Everything is running at this point, now check join state */
1441  } else if (do_terminate) {
1442  pe_fence_node(data_set, this_node, "termination was requested", FALSE);
1443 
1444  } else if (pcmk__str_eq(join, CRMD_JOINSTATE_MEMBER, pcmk__str_casei)) {
1445  crm_info("%s is active", pe__node_name(this_node));
1446 
1447  } else if (pcmk__strcase_any_of(join, CRMD_JOINSTATE_PENDING, CRMD_JOINSTATE_DOWN, NULL)) {
1448  crm_info("%s is not ready to run resources", pe__node_name(this_node));
1449  this_node->details->standby = TRUE;
1450  this_node->details->pending = TRUE;
1451 
1452  } else {
1453  pe_fence_node(data_set, this_node, "peer was in an unknown state", FALSE);
1454  crm_warn("%s: in-cluster=%s is-peer=%s join=%s expected=%s term=%d shutdown=%d",
1455  pe__node_name(this_node), pcmk__s(in_cluster, "<null>"),
1456  pcmk__s(is_peer, "<null>"), pcmk__s(join, "<null>"),
1457  pcmk__s(exp_state, "<null>"), do_terminate,
1458  this_node->details->shutdown);
1459  }
1460 
1461  return online;
1462 }
1463 
1464 static void
1465 determine_remote_online_status(pe_working_set_t * data_set, pe_node_t * this_node)
1466 {
1467  pe_resource_t *rsc = this_node->details->remote_rsc;
1468  pe_resource_t *container = NULL;
1469  pe_node_t *host = NULL;
1470 
1471  /* If there is a node state entry for a (former) Pacemaker Remote node
1472  * but no resource creating that node, the node's connection resource will
1473  * be NULL. Consider it an offline remote node in that case.
1474  */
1475  if (rsc == NULL) {
1476  this_node->details->online = FALSE;
1477  goto remote_online_done;
1478  }
1479 
1480  container = rsc->container;
1481 
1482  if (container && pcmk__list_of_1(rsc->running_on)) {
1483  host = rsc->running_on->data;
1484  }
1485 
1486  /* If the resource is currently started, mark it online. */
1487  if (rsc->role == RSC_ROLE_STARTED) {
1488  crm_trace("%s node %s presumed ONLINE because connection resource is started",
1489  (container? "Guest" : "Remote"), this_node->details->id);
1490  this_node->details->online = TRUE;
1491  }
1492 
1493  /* consider this node shutting down if transitioning start->stop */
1494  if (rsc->role == RSC_ROLE_STARTED && rsc->next_role == RSC_ROLE_STOPPED) {
1495  crm_trace("%s node %s shutting down because connection resource is stopping",
1496  (container? "Guest" : "Remote"), this_node->details->id);
1497  this_node->details->shutdown = TRUE;
1498  }
1499 
1500  /* Now check all the failure conditions. */
1501  if(container && pcmk_is_set(container->flags, pe_rsc_failed)) {
1502  crm_trace("Guest node %s UNCLEAN because guest resource failed",
1503  this_node->details->id);
1504  this_node->details->online = FALSE;
1505  this_node->details->remote_requires_reset = TRUE;
1506 
1507  } else if (pcmk_is_set(rsc->flags, pe_rsc_failed)) {
1508  crm_trace("%s node %s OFFLINE because connection resource failed",
1509  (container? "Guest" : "Remote"), this_node->details->id);
1510  this_node->details->online = FALSE;
1511 
1512  } else if (rsc->role == RSC_ROLE_STOPPED
1513  || (container && container->role == RSC_ROLE_STOPPED)) {
1514 
1515  crm_trace("%s node %s OFFLINE because its resource is stopped",
1516  (container? "Guest" : "Remote"), this_node->details->id);
1517  this_node->details->online = FALSE;
1518  this_node->details->remote_requires_reset = FALSE;
1519 
1520  } else if (host && (host->details->online == FALSE)
1521  && host->details->unclean) {
1522  crm_trace("Guest node %s UNCLEAN because host is unclean",
1523  this_node->details->id);
1524  this_node->details->online = FALSE;
1525  this_node->details->remote_requires_reset = TRUE;
1526  }
1527 
1528 remote_online_done:
1529  crm_trace("Remote node %s online=%s",
1530  this_node->details->id, this_node->details->online ? "TRUE" : "FALSE");
1531 }
1532 
1533 static void
1534 determine_online_status(xmlNode * node_state, pe_node_t * this_node, pe_working_set_t * data_set)
1535 {
1536  gboolean online = FALSE;
1537  const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED);
1538 
1539  CRM_CHECK(this_node != NULL, return);
1540 
1541  this_node->details->shutdown = FALSE;
1542  this_node->details->expected_up = FALSE;
1543 
1544  if (pe__shutdown_requested(this_node)) {
1545  this_node->details->shutdown = TRUE;
1546 
1547  } else if (pcmk__str_eq(exp_state, CRMD_JOINSTATE_MEMBER, pcmk__str_casei)) {
1548  this_node->details->expected_up = TRUE;
1549  }
1550 
1551  if (this_node->details->type == node_ping) {
1552  this_node->details->unclean = FALSE;
1553  online = FALSE; /* As far as resource management is concerned,
1554  * the node is safely offline.
1555  * Anyone caught abusing this logic will be shot
1556  */
1557 
1558  } else if (!pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
1559  online = determine_online_status_no_fencing(data_set, node_state, this_node);
1560 
1561  } else {
1562  online = determine_online_status_fencing(data_set, node_state, this_node);
1563  }
1564 
1565  if (online) {
1566  this_node->details->online = TRUE;
1567 
1568  } else {
1569  /* remove node from contention */
1570  this_node->fixed = TRUE;
1571  this_node->weight = -INFINITY;
1572  }
1573 
1574  if (online && this_node->details->shutdown) {
1575  /* don't run resources here */
1576  this_node->fixed = TRUE;
1577  this_node->weight = -INFINITY;
1578  }
1579 
1580  if (this_node->details->type == node_ping) {
1581  crm_info("%s is not a Pacemaker node", pe__node_name(this_node));
1582 
1583  } else if (this_node->details->unclean) {
1584  pe_proc_warn("%s is unclean", pe__node_name(this_node));
1585 
1586  } else if (this_node->details->online) {
1587  crm_info("%s is %s", pe__node_name(this_node),
1588  this_node->details->shutdown ? "shutting down" :
1589  this_node->details->pending ? "pending" :
1590  this_node->details->standby ? "standby" :
1591  this_node->details->maintenance ? "maintenance" : "online");
1592 
1593  } else {
1594  crm_trace("%s is offline", pe__node_name(this_node));
1595  }
1596 }
1597 
1606 const char *
1607 pe_base_name_end(const char *id)
1608 {
1609  if (!pcmk__str_empty(id)) {
1610  const char *end = id + strlen(id) - 1;
1611 
1612  for (const char *s = end; s > id; --s) {
1613  switch (*s) {
1614  case '0':
1615  case '1':
1616  case '2':
1617  case '3':
1618  case '4':
1619  case '5':
1620  case '6':
1621  case '7':
1622  case '8':
1623  case '9':
1624  break;
1625  case ':':
1626  return (s == end)? s : (s - 1);
1627  default:
1628  return end;
1629  }
1630  }
1631  return end;
1632  }
1633  return NULL;
1634 }
1635 
1646 char *
1647 clone_strip(const char *last_rsc_id)
1648 {
1649  const char *end = pe_base_name_end(last_rsc_id);
1650  char *basename = NULL;
1651 
1652  CRM_ASSERT(end);
1653  basename = strndup(last_rsc_id, end - last_rsc_id + 1);
1654  CRM_ASSERT(basename);
1655  return basename;
1656 }
1657 
1668 char *
1669 clone_zero(const char *last_rsc_id)
1670 {
1671  const char *end = pe_base_name_end(last_rsc_id);
1672  size_t base_name_len = end - last_rsc_id + 1;
1673  char *zero = NULL;
1674 
1675  CRM_ASSERT(end);
1676  zero = calloc(base_name_len + 3, sizeof(char));
1677  CRM_ASSERT(zero);
1678  memcpy(zero, last_rsc_id, base_name_len);
1679  zero[base_name_len] = ':';
1680  zero[base_name_len + 1] = '0';
1681  return zero;
1682 }
1683 
1684 static pe_resource_t *
1685 create_fake_resource(const char *rsc_id, xmlNode * rsc_entry, pe_working_set_t * data_set)
1686 {
1687  pe_resource_t *rsc = NULL;
1688  xmlNode *xml_rsc = create_xml_node(NULL, XML_CIB_TAG_RESOURCE);
1689 
1690  copy_in_properties(xml_rsc, rsc_entry);
1691  crm_xml_add(xml_rsc, XML_ATTR_ID, rsc_id);
1692  crm_log_xml_debug(xml_rsc, "Orphan resource");
1693 
1694  if (pe__unpack_resource(xml_rsc, &rsc, NULL, data_set) != pcmk_rc_ok) {
1695  return NULL;
1696  }
1697 
1698  if (xml_contains_remote_node(xml_rsc)) {
1699  pe_node_t *node;
1700 
1701  crm_debug("Detected orphaned remote node %s", rsc_id);
1702  node = pe_find_node(data_set->nodes, rsc_id);
1703  if (node == NULL) {
1704  node = pe_create_node(rsc_id, rsc_id, "remote", NULL, data_set);
1705  }
1706  link_rsc2remotenode(data_set, rsc);
1707 
1708  if (node) {
1709  crm_trace("Setting node %s as shutting down due to orphaned connection resource", rsc_id);
1710  node->details->shutdown = TRUE;
1711  }
1712  }
1713 
1714  if (crm_element_value(rsc_entry, XML_RSC_ATTR_CONTAINER)) {
1715  /* This orphaned rsc needs to be mapped to a container. */
1716  crm_trace("Detected orphaned container filler %s", rsc_id);
1717  pe__set_resource_flags(rsc, pe_rsc_orphan_container_filler);
1718  }
1719  pe__set_resource_flags(rsc, pe_rsc_orphan);
1720  data_set->resources = g_list_append(data_set->resources, rsc);
1721  return rsc;
1722 }
1723 
1728 static pe_resource_t *
1729 create_anonymous_orphan(pe_resource_t *parent, const char *rsc_id,
1730  pe_node_t *node, pe_working_set_t *data_set)
1731 {
1732  pe_resource_t *top = pe__create_clone_child(parent, data_set);
1733 
1734  // find_rsc() because we might be a cloned group
1735  pe_resource_t *orphan = top->fns->find_rsc(top, rsc_id, NULL, pe_find_clone);
1736 
1737  pe_rsc_debug(parent, "Created orphan %s for %s: %s on %s",
1738  top->id, parent->id, rsc_id, pe__node_name(node));
1739  return orphan;
1740 }
1741 
1756 static pe_resource_t *
1757 find_anonymous_clone(pe_working_set_t * data_set, pe_node_t * node, pe_resource_t * parent,
1758  const char *rsc_id)
1759 {
1760  GList *rIter = NULL;
1761  pe_resource_t *rsc = NULL;
1762  pe_resource_t *inactive_instance = NULL;
1763  gboolean skip_inactive = FALSE;
1764 
1765  CRM_ASSERT(parent != NULL);
1766  CRM_ASSERT(pe_rsc_is_clone(parent));
1767  CRM_ASSERT(!pcmk_is_set(parent->flags, pe_rsc_unique));
1768 
1769  // Check for active (or partially active, for cloned groups) instance
1770  pe_rsc_trace(parent, "Looking for %s on %s in %s",
1771  rsc_id, pe__node_name(node), parent->id);
1772  for (rIter = parent->children; rsc == NULL && rIter; rIter = rIter->next) {
1773  GList *locations = NULL;
1774  pe_resource_t *child = rIter->data;
1775 
1776  /* Check whether this instance is already known to be active or pending
1777  * anywhere, at this stage of unpacking. Because this function is called
1778  * for a resource before the resource's individual operation history
1779  * entries are unpacked, locations will generally not contain the
1780  * desired node.
1781  *
1782  * However, there are three exceptions:
1783  * (1) when child is a cloned group and we have already unpacked the
1784  * history of another member of the group on the same node;
1785  * (2) when we've already unpacked the history of another numbered
1786  * instance on the same node (which can happen if globally-unique
1787  * was flipped from true to false); and
1788  * (3) when we re-run calculations on the same data set as part of a
1789  * simulation.
1790  */
1791  child->fns->location(child, &locations, 2);
1792  if (locations) {
1793  /* We should never associate the same numbered anonymous clone
1794  * instance with multiple nodes, and clone instances can't migrate,
1795  * so there must be only one location, regardless of history.
1796  */
1797  CRM_LOG_ASSERT(locations->next == NULL);
1798 
1799  if (((pe_node_t *)locations->data)->details == node->details) {
1800  /* This child instance is active on the requested node, so check
1801  * for a corresponding configured resource. We use find_rsc()
1802  * instead of child because child may be a cloned group, and we
1803  * need the particular member corresponding to rsc_id.
1804  *
1805  * If the history entry is orphaned, rsc will be NULL.
1806  */
1807  rsc = parent->fns->find_rsc(child, rsc_id, NULL, pe_find_clone);
1808  if (rsc) {
1809  /* If there are multiple instance history entries for an
1810  * anonymous clone in a single node's history (which can
1811  * happen if globally-unique is switched from true to
1812  * false), we want to consider the instances beyond the
1813  * first as orphans, even if there are inactive instance
1814  * numbers available.
1815  */
1816  if (rsc->running_on) {
1817  crm_notice("Active (now-)anonymous clone %s has "
1818  "multiple (orphan) instance histories on %s",
1819  parent->id, pe__node_name(node));
1820  skip_inactive = TRUE;
1821  rsc = NULL;
1822  } else {
1823  pe_rsc_trace(parent, "Resource %s, active", rsc->id);
1824  }
1825  }
1826  }
1827  g_list_free(locations);
1828 
1829  } else {
1830  pe_rsc_trace(parent, "Resource %s, skip inactive", child->id);
1831  if (!skip_inactive && !inactive_instance
1832  && !pcmk_is_set(child->flags, pe_rsc_block)) {
1833  // Remember one inactive instance in case we don't find active
1834  inactive_instance = parent->fns->find_rsc(child, rsc_id, NULL,
1835  pe_find_clone);
1836 
1837  /* ... but don't use it if it was already associated with a
1838  * pending action on another node
1839  */
1840  if (inactive_instance && inactive_instance->pending_node
1841  && (inactive_instance->pending_node->details != node->details)) {
1842  inactive_instance = NULL;
1843  }
1844  }
1845  }
1846  }
1847 
1848  if ((rsc == NULL) && !skip_inactive && (inactive_instance != NULL)) {
1849  pe_rsc_trace(parent, "Resource %s, empty slot", inactive_instance->id);
1850  rsc = inactive_instance;
1851  }
1852 
1853  /* If the resource has "requires" set to "quorum" or "nothing", and we don't
1854  * have a clone instance for every node, we don't want to consume a valid
1855  * instance number for unclean nodes. Such instances may appear to be active
1856  * according to the history, but should be considered inactive, so we can
1857  * start an instance elsewhere. Treat such instances as orphans.
1858  *
1859  * An exception is instances running on guest nodes -- since guest node
1860  * "fencing" is actually just a resource stop, requires shouldn't apply.
1861  *
1862  * @TODO Ideally, we'd use an inactive instance number if it is not needed
1863  * for any clean instances. However, we don't know that at this point.
1864  */
1865  if ((rsc != NULL) && !pcmk_is_set(rsc->flags, pe_rsc_needs_fencing)
1866  && (!node->details->online || node->details->unclean)
1867  && !pe__is_guest_node(node)
1868  && !pe__is_universal_clone(parent, data_set)) {
1869 
1870  rsc = NULL;
1871  }
1872 
1873  if (rsc == NULL) {
1874  rsc = create_anonymous_orphan(parent, rsc_id, node, data_set);
1875  pe_rsc_trace(parent, "Resource %s, orphan", rsc->id);
1876  }
1877  return rsc;
1878 }
1879 
1880 static pe_resource_t *
1881 unpack_find_resource(pe_working_set_t * data_set, pe_node_t * node, const char *rsc_id,
1882  xmlNode * rsc_entry)
1883 {
1884  pe_resource_t *rsc = NULL;
1885  pe_resource_t *parent = NULL;
1886 
1887  crm_trace("looking for %s", rsc_id);
1888  rsc = pe_find_resource(data_set->resources, rsc_id);
1889 
1890  if (rsc == NULL) {
1891  /* If we didn't find the resource by its name in the operation history,
1892  * check it again as a clone instance. Even when clone-max=0, we create
1893  * a single :0 orphan to match against here.
1894  */
1895  char *clone0_id = clone_zero(rsc_id);
1896  pe_resource_t *clone0 = pe_find_resource(data_set->resources, clone0_id);
1897 
1898  if (clone0 && !pcmk_is_set(clone0->flags, pe_rsc_unique)) {
1899  rsc = clone0;
1900  parent = uber_parent(clone0);
1901  crm_trace("%s found as %s (%s)", rsc_id, clone0_id, parent->id);
1902  } else {
1903  crm_trace("%s is not known as %s either (orphan)",
1904  rsc_id, clone0_id);
1905  }
1906  free(clone0_id);
1907 
1908  } else if (rsc->variant > pe_native) {
1909  crm_trace("Resource history for %s is orphaned because it is no longer primitive",
1910  rsc_id);
1911  return NULL;
1912 
1913  } else {
1914  parent = uber_parent(rsc);
1915  }
1916 
1917  if (pe_rsc_is_anon_clone(parent)) {
1918 
1919  if (pe_rsc_is_bundled(parent)) {
1920  rsc = pe__find_bundle_replica(parent->parent, node);
1921  } else {
1922  char *base = clone_strip(rsc_id);
1923 
1924  rsc = find_anonymous_clone(data_set, node, parent, base);
1925  free(base);
1926  CRM_ASSERT(rsc != NULL);
1927  }
1928  }
1929 
1930  if (rsc && !pcmk__str_eq(rsc_id, rsc->id, pcmk__str_casei)
1931  && !pcmk__str_eq(rsc_id, rsc->clone_name, pcmk__str_casei)) {
1932 
1933  pcmk__str_update(&rsc->clone_name, rsc_id);
1934  pe_rsc_debug(rsc, "Internally renamed %s on %s to %s%s",
1935  rsc_id, pe__node_name(node), rsc->id,
1936  (pcmk_is_set(rsc->flags, pe_rsc_orphan)? " (ORPHAN)" : ""));
1937  }
1938  return rsc;
1939 }
1940 
1941 static pe_resource_t *
1942 process_orphan_resource(xmlNode * rsc_entry, pe_node_t * node, pe_working_set_t * data_set)
1943 {
1944  pe_resource_t *rsc = NULL;
1945  const char *rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID);
1946 
1947  crm_debug("Detected orphan resource %s on %s", rsc_id, pe__node_name(node));
1948  rsc = create_fake_resource(rsc_id, rsc_entry, data_set);
1949  if (rsc == NULL) {
1950  return NULL;
1951  }
1952 
1953  if (!pcmk_is_set(data_set->flags, pe_flag_stop_rsc_orphans)) {
1954  pe__clear_resource_flags(rsc, pe_rsc_managed);
1955 
1956  } else {
1957  CRM_CHECK(rsc != NULL, return NULL);
1958  pe_rsc_trace(rsc, "Added orphan %s", rsc->id);
1959  resource_location(rsc, NULL, -INFINITY, "__orphan_do_not_run__", data_set);
1960  }
1961  return rsc;
1962 }
1963 
1964 static void
1965 process_rsc_state(pe_resource_t * rsc, pe_node_t * node,
1966  enum action_fail_response on_fail,
1967  xmlNode * migrate_op, pe_working_set_t * data_set)
1968 {
1969  pe_node_t *tmpnode = NULL;
1970  char *reason = NULL;
1971  enum action_fail_response save_on_fail = action_fail_ignore;
1972 
1973  CRM_ASSERT(rsc);
1974  pe_rsc_trace(rsc, "Resource %s is %s on %s: on_fail=%s",
1975  rsc->id, role2text(rsc->role), pe__node_name(node),
1976  fail2text(on_fail));
1977 
1978  /* process current state */
1979  if (rsc->role != RSC_ROLE_UNKNOWN) {
1980  pe_resource_t *iter = rsc;
1981 
1982  while (iter) {
1983  if (g_hash_table_lookup(iter->known_on, node->details->id) == NULL) {
1984  pe_node_t *n = pe__copy_node(node);
1985 
1986  pe_rsc_trace(rsc, "%s%s%s known on %s",
1987  rsc->id,
1988  ((rsc->clone_name == NULL)? "" : " also known as "),
1989  ((rsc->clone_name == NULL)? "" : rsc->clone_name),
1990  pe__node_name(n));
1991  g_hash_table_insert(iter->known_on, (gpointer) n->details->id, n);
1992  }
1993  if (pcmk_is_set(iter->flags, pe_rsc_unique)) {
1994  break;
1995  }
1996  iter = iter->parent;
1997  }
1998  }
1999 
2000  /* If a managed resource is believed to be running, but node is down ... */
2001  if (rsc->role > RSC_ROLE_STOPPED
2002  && node->details->online == FALSE
2003  && node->details->maintenance == FALSE
2004  && pcmk_is_set(rsc->flags, pe_rsc_managed)) {
2005 
2006  gboolean should_fence = FALSE;
2007 
2008  /* If this is a guest node, fence it (regardless of whether fencing is
2009  * enabled, because guest node fencing is done by recovery of the
2010  * container resource rather than by the fencer). Mark the resource
2011  * we're processing as failed. When the guest comes back up, its
2012  * operation history in the CIB will be cleared, freeing the affected
2013  * resource to run again once we are sure we know its state.
2014  */
2015  if (pe__is_guest_node(node)) {
2016  pe__set_resource_flags(rsc, pe_rsc_failed|pe_rsc_stop);
2017  should_fence = TRUE;
2018 
2019  } else if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
2020  if (pe__is_remote_node(node) && node->details->remote_rsc
2021  && !pcmk_is_set(node->details->remote_rsc->flags, pe_rsc_failed)) {
2022 
2023  /* Setting unseen means that fencing of the remote node will
2024  * occur only if the connection resource is not going to start
2025  * somewhere. This allows connection resources on a failed
2026  * cluster node to move to another node without requiring the
2027  * remote nodes to be fenced as well.
2028  */
2029  node->details->unseen = TRUE;
2030  reason = crm_strdup_printf("%s is active there (fencing will be"
2031  " revoked if remote connection can "
2032  "be re-established elsewhere)",
2033  rsc->id);
2034  }
2035  should_fence = TRUE;
2036  }
2037 
2038  if (should_fence) {
2039  if (reason == NULL) {
2040  reason = crm_strdup_printf("%s is thought to be active there", rsc->id);
2041  }
2042  pe_fence_node(data_set, node, reason, FALSE);
2043  }
2044  free(reason);
2045  }
2046 
2047  /* In order to calculate priority_fencing_delay correctly, save the failure information and pass it to native_add_running(). */
2048  save_on_fail = on_fail;
2049 
2050  if (node->details->unclean) {
2051  /* No extra processing needed
2052  * Also allows resources to be started again after a node is shot
2053  */
2054  on_fail = action_fail_ignore;
2055  }
2056 
2057  switch (on_fail) {
2058  case action_fail_ignore:
2059  /* nothing to do */
2060  break;
2061 
2062  case action_fail_demote:
2063  pe__set_resource_flags(rsc, pe_rsc_failed);
2064  demote_action(rsc, node, FALSE);
2065  break;
2066 
2067  case action_fail_fence:
2068  /* treat it as if it is still running
2069  * but also mark the node as unclean
2070  */
2071  reason = crm_strdup_printf("%s failed there", rsc->id);
2072  pe_fence_node(data_set, node, reason, FALSE);
2073  free(reason);
2074  break;
2075 
2076  case action_fail_standby:
2077  node->details->standby = TRUE;
2078  node->details->standby_onfail = TRUE;
2079  break;
2080 
2081  case action_fail_block:
2082  /* is_managed == FALSE will prevent any
2083  * actions being sent for the resource
2084  */
2085  pe__clear_resource_flags(rsc, pe_rsc_managed);
2086  pe__set_resource_flags(rsc, pe_rsc_block);
2087  break;
2088 
2089  case action_fail_migrate:
2090  /* make sure it comes up somewhere else
2091  * or not at all
2092  */
2093  resource_location(rsc, node, -INFINITY, "__action_migration_auto__", data_set);
2094  break;
2095 
2096  case action_fail_stop:
2097  pe__set_next_role(rsc, RSC_ROLE_STOPPED, "on-fail=stop");
2098  break;
2099 
2100  case action_fail_recover:
2101  if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) {
2102  pe__set_resource_flags(rsc, pe_rsc_failed|pe_rsc_stop);
2103  stop_action(rsc, node, FALSE);
2104  }
2105  break;
2106 
2107  case action_fail_restart_container:
2108  pe__set_resource_flags(rsc, pe_rsc_failed|pe_rsc_stop);
2109  if (rsc->container && pe_rsc_is_bundled(rsc)) {
2110  /* A bundle's remote connection can run on a different node than
2111  * the bundle's container. We don't necessarily know where the
2112  * container is running yet, so remember it and add a stop
2113  * action for it later.
2114  */
2115  data_set->stop_needed = g_list_prepend(data_set->stop_needed,
2116  rsc->container);
2117  } else if (rsc->container) {
2118  stop_action(rsc->container, node, FALSE);
2119  } else if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) {
2120  stop_action(rsc, node, FALSE);
2121  }
2122  break;
2123 
2124  case action_fail_reset_remote:
2125  pe__set_resource_flags(rsc, pe_rsc_failed|pe_rsc_stop);
2126  if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
2127  tmpnode = NULL;
2128  if (rsc->is_remote_node) {
2129  tmpnode = pe_find_node(data_set->nodes, rsc->id);
2130  }
2131  if (tmpnode &&
2132  pe__is_remote_node(tmpnode) &&
2133  tmpnode->details->remote_was_fenced == 0) {
2134 
2135  /* The remote connection resource failed in a way that
2136  * should result in fencing the remote node.
2137  */
2138  pe_fence_node(data_set, tmpnode,
2139  "remote connection is unrecoverable", FALSE);
2140  }
2141  }
2142 
2143  /* require the stop action regardless if fencing is occurring or not. */
2144  if (rsc->role > RSC_ROLE_STOPPED) {
2145  stop_action(rsc, node, FALSE);
2146  }
2147 
2148  /* if reconnect delay is in use, prevent the connection from exiting the
2149  * "STOPPED" role until the failure is cleared by the delay timeout. */
2150  if (rsc->remote_reconnect_ms) {
2151  pe__set_next_role(rsc, RSC_ROLE_STOPPED, "remote reset");
2152  }
2153  break;
2154  }
2155 
2156  /* ensure a remote-node connection failure forces an unclean remote-node
2157  * to be fenced. By setting unseen = FALSE, the remote-node failure will
2158  * result in a fencing operation regardless if we're going to attempt to
2159  * reconnect to the remote-node in this transition or not. */
2160  if (pcmk_is_set(rsc->flags, pe_rsc_failed) && rsc->is_remote_node) {
2161  tmpnode = pe_find_node(data_set->nodes, rsc->id);
2162  if (tmpnode && tmpnode->details->unclean) {
2163  tmpnode->details->unseen = FALSE;
2164  }
2165  }
2166 
2167  if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) {
2168  if (pcmk_is_set(rsc->flags, pe_rsc_orphan)) {
2169  if (pcmk_is_set(rsc->flags, pe_rsc_managed)) {
2170  pcmk__config_warn("Detected active orphan %s running on %s",
2171  rsc->id, pe__node_name(node));
2172  } else {
2173  pcmk__config_warn("Resource '%s' must be stopped manually on "
2174  "%s because cluster is configured not to "
2175  "stop active orphans",
2176  rsc->id, pe__node_name(node));
2177  }
2178  }
2179 
2180  native_add_running(rsc, node, data_set, (save_on_fail != action_fail_ignore));
2181  switch (on_fail) {
2182  case action_fail_ignore:
2183  break;
2184  case action_fail_demote:
2185  case action_fail_block:
2186  pe__set_resource_flags(rsc, pe_rsc_failed);
2187  break;
2188  default:
2189  pe__set_resource_flags(rsc, pe_rsc_failed|pe_rsc_stop);
2190  break;
2191  }
2192 
2193  } else if (rsc->clone_name && strchr(rsc->clone_name, ':') != NULL) {
2194  /* Only do this for older status sections that included instance numbers
2195  * Otherwise stopped instances will appear as orphans
2196  */
2197  pe_rsc_trace(rsc, "Resetting clone_name %s for %s (stopped)", rsc->clone_name, rsc->id);
2198  free(rsc->clone_name);
2199  rsc->clone_name = NULL;
2200 
2201  } else {
2202  GList *possible_matches = pe__resource_actions(rsc, node, RSC_STOP,
2203  FALSE);
2204  GList *gIter = possible_matches;
2205 
2206  for (; gIter != NULL; gIter = gIter->next) {
2207  pe_action_t *stop = (pe_action_t *) gIter->data;
2208 
2209  pe__set_action_flags(stop, pe_action_optional);
2210  }
2211 
2212  g_list_free(possible_matches);
2213  }
2214 
2215  /* A successful stop after migrate_to on the migration source doesn't make
2216  * the partially migrated resource stopped on the migration target.
2217  */
2218  if (rsc->role == RSC_ROLE_STOPPED
2219  && rsc->partial_migration_source
2220  && rsc->partial_migration_source->details == node->details
2221  && rsc->partial_migration_target
2222  && rsc->running_on) {
2223 
2224  rsc->role = RSC_ROLE_STARTED;
2225  }
2226 }
2227 
2228 /* create active recurring operations as optional */
2229 static void
2230 process_recurring(pe_node_t * node, pe_resource_t * rsc,
2231  int start_index, int stop_index,
2232  GList *sorted_op_list, pe_working_set_t * data_set)
2233 {
2234  int counter = -1;
2235  const char *task = NULL;
2236  const char *status = NULL;
2237  GList *gIter = sorted_op_list;
2238 
2239  CRM_ASSERT(rsc);
2240  pe_rsc_trace(rsc, "%s: Start index %d, stop index = %d", rsc->id, start_index, stop_index);
2241 
2242  for (; gIter != NULL; gIter = gIter->next) {
2243  xmlNode *rsc_op = (xmlNode *) gIter->data;
2244 
2245  guint interval_ms = 0;
2246  char *key = NULL;
2247  const char *id = ID(rsc_op);
2248 
2249  counter++;
2250 
2251  if (node->details->online == FALSE) {
2252  pe_rsc_trace(rsc, "Skipping %s on %s: node is offline",
2253  rsc->id, pe__node_name(node));
2254  break;
2255 
2256  /* Need to check if there's a monitor for role="Stopped" */
2257  } else if (start_index < stop_index && counter <= stop_index) {
2258  pe_rsc_trace(rsc, "Skipping %s on %s: resource is not active",
2259  id, pe__node_name(node));
2260  continue;
2261 
2262  } else if (counter < start_index) {
2263  pe_rsc_trace(rsc, "Skipping %s on %s: old %d",
2264  id, pe__node_name(node), counter);
2265  continue;
2266  }
2267 
2268  crm_element_value_ms(rsc_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms);
2269  if (interval_ms == 0) {
2270  pe_rsc_trace(rsc, "Skipping %s on %s: non-recurring",
2271  id, pe__node_name(node));
2272  continue;
2273  }
2274 
2275  status = crm_element_value(rsc_op, XML_LRM_ATTR_OPSTATUS);
2276  if (pcmk__str_eq(status, "-1", pcmk__str_casei)) {
2277  pe_rsc_trace(rsc, "Skipping %s on %s: status",
2278  id, pe__node_name(node));
2279  continue;
2280  }
2281  task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK);
2282  /* create the action */
2283  key = pcmk__op_key(rsc->id, task, interval_ms);
2284  pe_rsc_trace(rsc, "Creating %s on %s", key, pe__node_name(node));
2285  custom_action(rsc, key, task, node, TRUE, TRUE, data_set);
2286  }
2287 }
2288 
2289 void
2290 calculate_active_ops(GList *sorted_op_list, int *start_index, int *stop_index)
2291 {
2292  int counter = -1;
2293  int implied_monitor_start = -1;
2294  int implied_clone_start = -1;
2295  const char *task = NULL;
2296  const char *status = NULL;
2297  GList *gIter = sorted_op_list;
2298 
2299  *stop_index = -1;
2300  *start_index = -1;
2301 
2302  for (; gIter != NULL; gIter = gIter->next) {
2303  xmlNode *rsc_op = (xmlNode *) gIter->data;
2304 
2305  counter++;
2306 
2307  task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK);
2308  status = crm_element_value(rsc_op, XML_LRM_ATTR_OPSTATUS);
2309 
2310  if (pcmk__str_eq(task, CRMD_ACTION_STOP, pcmk__str_casei)
2311  && pcmk__str_eq(status, "0", pcmk__str_casei)) {
2312  *stop_index = counter;
2313 
2314  } else if (pcmk__strcase_any_of(task, CRMD_ACTION_START, CRMD_ACTION_MIGRATED, NULL)) {
2315  *start_index = counter;
2316 
2317  } else if ((implied_monitor_start <= *stop_index) && pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_casei)) {
2318  const char *rc = crm_element_value(rsc_op, XML_LRM_ATTR_RC);
2319 
2320  if (pcmk__strcase_any_of(rc, "0", "8", NULL)) {
2321  implied_monitor_start = counter;
2322  }
2323  } else if (pcmk__strcase_any_of(task, CRMD_ACTION_PROMOTE, CRMD_ACTION_DEMOTE, NULL)) {
2324  implied_clone_start = counter;
2325  }
2326  }
2327 
2328  if (*start_index == -1) {
2329  if (implied_clone_start != -1) {
2330  *start_index = implied_clone_start;
2331  } else if (implied_monitor_start != -1) {
2332  *start_index = implied_monitor_start;
2333  }
2334  }
2335 }
2336 
2337 // If resource history entry has shutdown lock, remember lock node and time
2338 static void
2339 unpack_shutdown_lock(xmlNode *rsc_entry, pe_resource_t *rsc, pe_node_t *node,
2340  pe_working_set_t *data_set)
2341 {
2342  time_t lock_time = 0; // When lock started (i.e. node shutdown time)
2343 
2344  if ((crm_element_value_epoch(rsc_entry, XML_CONFIG_ATTR_SHUTDOWN_LOCK,
2345  &lock_time) == pcmk_ok) && (lock_time != 0)) {
2346 
2347  if ((data_set->shutdown_lock > 0)
2348  && (get_effective_time(data_set)
2349  > (lock_time + data_set->shutdown_lock))) {
2350  pe_rsc_info(rsc, "Shutdown lock for %s on %s expired",
2351  rsc->id, pe__node_name(node));
2352  pe__clear_resource_history(rsc, node, data_set);
2353  } else {
2354  rsc->lock_node = node;
2355  rsc->lock_time = lock_time;
2356  }
2357  }
2358 }
2359 
2370 static pe_resource_t *
2371 unpack_lrm_resource(pe_node_t *node, xmlNode *lrm_resource,
2372  pe_working_set_t *data_set)
2373 {
2374  GList *gIter = NULL;
2375  int stop_index = -1;
2376  int start_index = -1;
2377  enum rsc_role_e req_role = RSC_ROLE_UNKNOWN;
2378 
2379  const char *task = NULL;
2380  const char *rsc_id = ID(lrm_resource);
2381 
2382  pe_resource_t *rsc = NULL;
2383  GList *op_list = NULL;
2384  GList *sorted_op_list = NULL;
2385 
2386  xmlNode *migrate_op = NULL;
2387  xmlNode *rsc_op = NULL;
2388  xmlNode *last_failure = NULL;
2389 
2390  enum action_fail_response on_fail = action_fail_ignore;
2391  enum rsc_role_e saved_role = RSC_ROLE_UNKNOWN;
2392 
2393  if (rsc_id == NULL) {
2394  crm_warn("Ignoring malformed " XML_LRM_TAG_RESOURCE
2395  " entry without id");
2396  return NULL;
2397  }
2398  crm_trace("Unpacking " XML_LRM_TAG_RESOURCE " for %s on %s",
2399  rsc_id, pe__node_name(node));
2400 
2401  // Build a list of individual lrm_rsc_op entries, so we can sort them
2402  for (rsc_op = first_named_child(lrm_resource, XML_LRM_TAG_RSC_OP);
2403  rsc_op != NULL; rsc_op = crm_next_same_xml(rsc_op)) {
2404 
2405  op_list = g_list_prepend(op_list, rsc_op);
2406  }
2407 
2408  if (!pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)) {
2409  if (op_list == NULL) {
2410  // If there are no operations, there is nothing to do
2411  return NULL;
2412  }
2413  }
2414 
2415  /* find the resource */
2416  rsc = unpack_find_resource(data_set, node, rsc_id, lrm_resource);
2417  if (rsc == NULL) {
2418  if (op_list == NULL) {
2419  // If there are no operations, there is nothing to do
2420  return NULL;
2421  } else {
2422  rsc = process_orphan_resource(lrm_resource, node, data_set);
2423  }
2424  }
2425  CRM_ASSERT(rsc != NULL);
2426 
2427  // Check whether the resource is "shutdown-locked" to this node
2428  if (pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)) {
2429  unpack_shutdown_lock(lrm_resource, rsc, node, data_set);
2430  }
2431 
2432  /* process operations */
2433  saved_role = rsc->role;
2434  rsc->role = RSC_ROLE_UNKNOWN;
2435  sorted_op_list = g_list_sort(op_list, sort_op_by_callid);
2436 
2437  for (gIter = sorted_op_list; gIter != NULL; gIter = gIter->next) {
2438  xmlNode *rsc_op = (xmlNode *) gIter->data;
2439 
2440  task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK);
2441  if (pcmk__str_eq(task, CRMD_ACTION_MIGRATED, pcmk__str_casei)) {
2442  migrate_op = rsc_op;
2443  }
2444 
2445  unpack_rsc_op(rsc, node, rsc_op, &last_failure, &on_fail, data_set);
2446  }
2447 
2448  /* create active recurring operations as optional */
2449  calculate_active_ops(sorted_op_list, &start_index, &stop_index);
2450  process_recurring(node, rsc, start_index, stop_index, sorted_op_list, data_set);
2451 
2452  /* no need to free the contents */
2453  g_list_free(sorted_op_list);
2454 
2455  process_rsc_state(rsc, node, on_fail, migrate_op, data_set);
2456 
2457  if (get_target_role(rsc, &req_role)) {
2458  if (rsc->next_role == RSC_ROLE_UNKNOWN || req_role < rsc->next_role) {
2459  pe__set_next_role(rsc, req_role, XML_RSC_ATTR_TARGET_ROLE);
2460 
2461  } else if (req_role > rsc->next_role) {
2462  pe_rsc_info(rsc, "%s: Not overwriting calculated next role %s"
2463  " with requested next role %s",
2464  rsc->id, role2text(rsc->next_role), role2text(req_role));
2465  }
2466  }
2467 
2468  if (saved_role > rsc->role) {
2469  rsc->role = saved_role;
2470  }
2471 
2472  return rsc;
2473 }
2474 
2475 static void
2476 handle_orphaned_container_fillers(xmlNode * lrm_rsc_list, pe_working_set_t * data_set)
2477 {
2478  xmlNode *rsc_entry = NULL;
2479  for (rsc_entry = pcmk__xe_first_child(lrm_rsc_list); rsc_entry != NULL;
2480  rsc_entry = pcmk__xe_next(rsc_entry)) {
2481 
2482  pe_resource_t *rsc;
2483  pe_resource_t *container;
2484  const char *rsc_id;
2485  const char *container_id;
2486 
2487  if (!pcmk__str_eq((const char *)rsc_entry->name, XML_LRM_TAG_RESOURCE, pcmk__str_casei)) {
2488  continue;
2489  }
2490 
2491  container_id = crm_element_value(rsc_entry, XML_RSC_ATTR_CONTAINER);
2492  rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID);
2493  if (container_id == NULL || rsc_id == NULL) {
2494  continue;
2495  }
2496 
2497  container = pe_find_resource(data_set->resources, container_id);
2498  if (container == NULL) {
2499  continue;
2500  }
2501 
2502  rsc = pe_find_resource(data_set->resources, rsc_id);
2503  if (rsc == NULL ||
2504  !pcmk_is_set(rsc->flags, pe_rsc_orphan_container_filler) ||
2505  rsc->container != NULL) {
2506  continue;
2507  }
2508 
2509  pe_rsc_trace(rsc, "Mapped container of orphaned resource %s to %s",
2510  rsc->id, container_id);
2511  rsc->container = container;
2512  container->fillers = g_list_append(container->fillers, rsc);
2513  }
2514 }
2515 
2524 static void
2525 unpack_node_lrm(pe_node_t *node, xmlNode *xml, pe_working_set_t *data_set)
2526 {
2527  bool found_orphaned_container_filler = false;
2528 
2529  // Drill down to lrm_resources section
2530  xml = find_xml_node(xml, XML_CIB_TAG_LRM, FALSE);
2531  if (xml == NULL) {
2532  return;
2533  }
2534  xml = find_xml_node(xml, XML_LRM_TAG_RESOURCES, FALSE);
2535  if (xml == NULL) {
2536  return;
2537  }
2538 
2539  // Unpack each lrm_resource entry
2540  for (xmlNode *rsc_entry = first_named_child(xml, XML_LRM_TAG_RESOURCE);
2541  rsc_entry != NULL; rsc_entry = crm_next_same_xml(rsc_entry)) {
2542 
2543  pe_resource_t *rsc = unpack_lrm_resource(node, rsc_entry, data_set);
2544 
2545  if ((rsc != NULL)
2546  && pcmk_is_set(rsc->flags, pe_rsc_orphan_container_filler)) {
2547  found_orphaned_container_filler = true;
2548  }
2549  }
2550 
2551  /* Now that all resource state has been unpacked for this node, map any
2552  * orphaned container fillers to their container resource.
2553  */
2554  if (found_orphaned_container_filler) {
2555  handle_orphaned_container_fillers(xml, data_set);
2556  }
2557 }
2558 
2559 static void
2560 set_active(pe_resource_t * rsc)
2561 {
2562  pe_resource_t *top = uber_parent(rsc);
2563 
2564  if (top && pcmk_is_set(top->flags, pe_rsc_promotable)) {
2565  rsc->role = RSC_ROLE_UNPROMOTED;
2566  } else {
2567  rsc->role = RSC_ROLE_STARTED;
2568  }
2569 }
2570 
2571 static void
2572 set_node_score(gpointer key, gpointer value, gpointer user_data)
2573 {
2574  pe_node_t *node = value;
2575  int *score = user_data;
2576 
2577  node->weight = *score;
2578 }
2579 
2580 static xmlNode *
2581 find_lrm_op(const char *resource, const char *op, const char *node, const char *source,
2582  int target_rc, pe_working_set_t *data_set)
2583 {
2584  GString *xpath = NULL;
2585  xmlNode *xml = NULL;
2586 
2587  CRM_CHECK((resource != NULL) && (op != NULL) && (node != NULL),
2588  return NULL);
2589 
2590  xpath = g_string_sized_new(256);
2591  pcmk__g_strcat(xpath,
2592  "//" XML_CIB_TAG_STATE "[@" XML_ATTR_UNAME "='", node, "']"
2593  "//" XML_LRM_TAG_RESOURCE
2594  "[@" XML_ATTR_ID "='", resource, "']"
2595  "/" XML_LRM_TAG_RSC_OP "[@" XML_LRM_ATTR_TASK "='", op, "'",
2596  NULL);
2597 
2598  /* Need to check against transition_magic too? */
2599  if ((source != NULL) && (strcmp(op, CRMD_ACTION_MIGRATE) == 0)) {
2600  pcmk__g_strcat(xpath,
2601  " and @" XML_LRM_ATTR_MIGRATE_TARGET "='", source, "']",
2602  NULL);
2603 
2604  } else if ((source != NULL) && (strcmp(op, CRMD_ACTION_MIGRATED) == 0)) {
2605  pcmk__g_strcat(xpath,
2606  " and @" XML_LRM_ATTR_MIGRATE_SOURCE "='", source, "']",
2607  NULL);
2608  } else {
2609  g_string_append_c(xpath, ']');
2610  }
2611 
2612  xml = get_xpath_object((const char *) xpath->str, data_set->input,
2613  LOG_DEBUG);
2614  g_string_free(xpath, TRUE);
2615 
2616  if (xml && target_rc >= 0) {
2617  int rc = PCMK_OCF_UNKNOWN_ERROR;
2618  int status = PCMK_EXEC_ERROR;
2619 
2620  crm_element_value_int(xml, XML_LRM_ATTR_RC, &rc);
2621  crm_element_value_int(xml, XML_LRM_ATTR_OPSTATUS, &status);
2622  if ((rc != target_rc) || (status != PCMK_EXEC_DONE)) {
2623  return NULL;
2624  }
2625  }
2626  return xml;
2627 }
2628 
2629 static xmlNode *
2630 find_lrm_resource(const char *rsc_id, const char *node_name,
2631  pe_working_set_t *data_set)
2632 {
2633  GString *xpath = NULL;
2634  xmlNode *xml = NULL;
2635 
2636  CRM_CHECK((rsc_id != NULL) && (node_name != NULL), return NULL);
2637 
2638  xpath = g_string_sized_new(256);
2639  pcmk__g_strcat(xpath,
2640  "//" XML_CIB_TAG_STATE
2641  "[@" XML_ATTR_UNAME "='", node_name, "']"
2642  "//" XML_LRM_TAG_RESOURCE
2643  "[@" XML_ATTR_ID "='", rsc_id, "']",
2644  NULL);
2645 
2646  xml = get_xpath_object((const char *) xpath->str, data_set->input,
2647  LOG_DEBUG);
2648 
2649  g_string_free(xpath, TRUE);
2650  return xml;
2651 }
2652 
2653 static bool
2654 unknown_on_node(const char *rsc_id, const char *node_name,
2655  pe_working_set_t *data_set)
2656 {
2657  xmlNode *lrm_resource = NULL;
2658 
2659  lrm_resource = find_lrm_resource(rsc_id, node_name, data_set);
2660 
2661  /* If the resource has no lrm_rsc_op history on the node, that means its
2662  * state is unknown there.
2663  */
2664  return (lrm_resource == NULL
2665  || first_named_child(lrm_resource, XML_LRM_TAG_RSC_OP) == NULL);
2666 }
2667 
2679 static bool
2680 monitor_not_running_after(const char *rsc_id, const char *node_name,
2681  xmlNode *xml_op, bool same_node,
2682  pe_working_set_t *data_set)
2683 {
2684  /* Any probe/monitor operation on the node indicating it was not running
2685  * there
2686  */
2687  xmlNode *monitor = find_lrm_op(rsc_id, CRMD_ACTION_STATUS, node_name,
2688  NULL, PCMK_OCF_NOT_RUNNING, data_set);
2689 
2690  return (monitor && pe__is_newer_op(monitor, xml_op, same_node) > 0);
2691 }
2692 
2705 static bool
2706 non_monitor_after(const char *rsc_id, const char *node_name, xmlNode *xml_op,
2707  bool same_node, pe_working_set_t *data_set)
2708 {
2709  xmlNode *lrm_resource = NULL;
2710 
2711  lrm_resource = find_lrm_resource(rsc_id, node_name, data_set);
2712  if (lrm_resource == NULL) {
2713  return false;
2714  }
2715 
2716  for (xmlNode *op = first_named_child(lrm_resource, XML_LRM_TAG_RSC_OP);
2717  op != NULL; op = crm_next_same_xml(op)) {
2718  const char * task = NULL;
2719 
2720  if (op == xml_op) {
2721  continue;
2722  }
2723 
2724  task = crm_element_value(op, XML_LRM_ATTR_TASK);
2725 
2726  if (pcmk__str_any_of(task, CRMD_ACTION_START, CRMD_ACTION_STOP,
2727  CRMD_ACTION_MIGRATE, CRMD_ACTION_MIGRATED, NULL)
2728  && pe__is_newer_op(op, xml_op, same_node) > 0) {
2729  return true;
2730  }
2731  }
2732 
2733  return false;
2734 }
2735 
2748 static bool
2749 newer_state_after_migrate(const char *rsc_id, const char *node_name,
2750  xmlNode *migrate_to, xmlNode *migrate_from,
2751  pe_working_set_t *data_set)
2752 {
2753  xmlNode *xml_op = migrate_to;
2754  const char *source = NULL;
2755  const char *target = NULL;
2756  bool same_node = false;
2757 
2758  if (migrate_from) {
2759  xml_op = migrate_from;
2760  }
2761 
2762  source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE);
2763  target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
2764 
2765  /* It's preferred to compare to the migrate event on the same node if
2766  * existing, since call ids are more reliable.
2767  */
2768  if (pcmk__str_eq(node_name, target, pcmk__str_casei)) {
2769  if (migrate_from) {
2770  xml_op = migrate_from;
2771  same_node = true;
2772 
2773  } else {
2774  xml_op = migrate_to;
2775  }
2776 
2777  } else if (pcmk__str_eq(node_name, source, pcmk__str_casei)) {
2778  if (migrate_to) {
2779  xml_op = migrate_to;
2780  same_node = true;
2781 
2782  } else {
2783  xml_op = migrate_from;
2784  }
2785  }
2786 
2787  /* If there's any newer non-monitor operation on the node, or any newer
2788  * probe/monitor operation on the node indicating it was not running there,
2789  * the migration events potentially no longer matter for the node.
2790  */
2791  return non_monitor_after(rsc_id, node_name, xml_op, same_node, data_set)
2792  || monitor_not_running_after(rsc_id, node_name, xml_op, same_node,
2793  data_set);
2794 }
2795 
2796 static void
2797 unpack_migrate_to_success(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op,
2798  pe_working_set_t *data_set)
2799 {
2800  /* A successful migration sequence is:
2801  * migrate_to on source node
2802  * migrate_from on target node
2803  * stop on source node
2804  *
2805  * But there could be scenarios like (It's easier to produce with cluster
2806  * property batch-limit=1):
2807  *
2808  * - rscA is live-migrating from node1 to node2.
2809  *
2810  * - Before migrate_to on node1 returns, put node2 into standby.
2811  *
2812  * - Transition aborts upon return of successful migrate_to on node1. New
2813  * transition is going to stop the rscA on both nodes and start it on
2814  * node1.
2815  *
2816  * - While it is stopping on node1, run something that is going to make
2817  * the transition abort again like:
2818  * crm_resource --resource rscA --ban --node node2
2819  *
2820  * - Transition aborts upon return of stop on node1.
2821  *
2822  * Now although there's a stop on node1, it's still a partial migration and
2823  * rscA is still potentially active on node2.
2824  *
2825  * So even if a migrate_to is followed by a stop, we still need to check
2826  * whether there's a corresponding migrate_from or any newer operation on
2827  * the target.
2828  *
2829  * If no migrate_from has happened, the migration is considered to be
2830  * "partial". If the migrate_from failed, make sure the resource gets
2831  * stopped on both source and target (if up).
2832  *
2833  * If the migrate_to and migrate_from both succeeded (which also implies the
2834  * resource is no longer running on the source), but there is no stop, the
2835  * migration is considered to be "dangling". Schedule a stop on the source
2836  * in this case.
2837  */
2838  int from_rc = 0;
2839  int from_status = 0;
2840  pe_node_t *target_node = NULL;
2841  pe_node_t *source_node = NULL;
2842  xmlNode *migrate_from = NULL;
2843  const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE);
2844  const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
2845  bool source_newer_op = false;
2846  bool target_newer_state = false;
2847 
2848  // Sanity check
2849  CRM_CHECK(source && target && !strcmp(source, node->details->uname), return);
2850 
2851  /* If there's any newer non-monitor operation on the source, this migrate_to
2852  * potentially no longer matters for the source.
2853  */
2854  source_newer_op = non_monitor_after(rsc->id, source, xml_op, true,
2855  data_set);
2856 
2857  // Check whether there was a migrate_from action on the target
2858  migrate_from = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATED, target,
2859  source, -1, data_set);
2860 
2861  /* Even if there's a newer non-monitor operation on the source, we still
2862  * need to check how this migrate_to might matter for the target.
2863  */
2864  if (source_newer_op && migrate_from) {
2865  return;
2866  }
2867 
2868  /* If the resource has newer state on the target after the migration
2869  * events, this migrate_to no longer matters for the target.
2870  */
2871  target_newer_state = newer_state_after_migrate(rsc->id, target, xml_op,
2872  migrate_from, data_set);
2873 
2874  if (source_newer_op && target_newer_state) {
2875  return;
2876  }
2877 
2878  // Clones are not allowed to migrate, so role can't be promoted
2879  rsc->role = RSC_ROLE_STARTED;
2880 
2881  target_node = pe_find_node(data_set->nodes, target);
2882  source_node = pe_find_node(data_set->nodes, source);
2883 
2884  if (migrate_from) {
2885  crm_element_value_int(migrate_from, XML_LRM_ATTR_RC, &from_rc);
2886  crm_element_value_int(migrate_from, XML_LRM_ATTR_OPSTATUS, &from_status);
2887  pe_rsc_trace(rsc, "%s op on %s exited with status=%d, rc=%d",
2888  ID(migrate_from), target, from_status, from_rc);
2889  }
2890 
2891  if (migrate_from && from_rc == PCMK_OCF_OK
2892  && (from_status == PCMK_EXEC_DONE)) {
2893  /* The migrate_to and migrate_from both succeeded, so mark the migration
2894  * as "dangling". This will be used to schedule a stop action on the
2895  * source without affecting the target.
2896  */
2897  pe_rsc_trace(rsc, "Detected dangling migration op: %s on %s", ID(xml_op),
2898  source);
2899  rsc->role = RSC_ROLE_STOPPED;
2900  rsc->dangling_migrations = g_list_prepend(rsc->dangling_migrations, node);
2901 
2902  } else if (migrate_from && (from_status != PCMK_EXEC_PENDING)) { // Failed
2903  /* If the resource has newer state on the target, this migrate_to no
2904  * longer matters for the target.
2905  */
2906  if (!target_newer_state
2907  && target_node && target_node->details->online) {
2908  pe_rsc_trace(rsc, "Marking active on %s %p %d", target, target_node,
2909  target_node->details->online);
2910  native_add_running(rsc, target_node, data_set, TRUE);
2911 
2912  } else {
2913  /* With the earlier bail logic, migrate_from != NULL here implies
2914  * source_newer_op is false, meaning this migrate_to still matters
2915  * for the source.
2916  * Consider it failed here - forces a restart, prevents migration
2917  */
2918  pe__set_resource_flags(rsc, pe_rsc_failed|pe_rsc_stop);
2919  pe__clear_resource_flags(rsc, pe_rsc_allow_migrate);
2920  }
2921 
2922  } else { // Pending, or complete but erased
2923  /* If the resource has newer state on the target, this migrate_to no
2924  * longer matters for the target.
2925  */
2926  if (!target_newer_state
2927  && target_node && target_node->details->online) {
2928  pe_rsc_trace(rsc, "Marking active on %s %p %d", target, target_node,
2929  target_node->details->online);
2930 
2931  native_add_running(rsc, target_node, data_set, FALSE);
2932  if (source_node && source_node->details->online) {
2933  /* This is a partial migration: the migrate_to completed
2934  * successfully on the source, but the migrate_from has not
2935  * completed. Remember the source and target; if the newly
2936  * chosen target remains the same when we schedule actions
2937  * later, we may continue with the migration.
2938  */
2939  rsc->partial_migration_target = target_node;
2940  rsc->partial_migration_source = source_node;
2941  }
2942  } else if (!source_newer_op) {
2943  /* This migrate_to matters for the source only if it's the last
2944  * non-monitor operation here.
2945  * Consider it failed here - forces a restart, prevents migration
2946  */
2947  pe__set_resource_flags(rsc, pe_rsc_failed|pe_rsc_stop);
2948  pe__clear_resource_flags(rsc, pe_rsc_allow_migrate);
2949  }
2950  }
2951 }
2952 
2953 static void
2954 unpack_migrate_to_failure(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op,
2955  pe_working_set_t *data_set)
2956 {
2957  xmlNode *target_migrate_from = NULL;
2958  const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE);
2959  const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
2960 
2961  // Sanity check
2962  CRM_CHECK(source && target && !strcmp(source, node->details->uname), return);
2963 
2964  /* If a migration failed, we have to assume the resource is active. Clones
2965  * are not allowed to migrate, so role can't be promoted.
2966  */
2967  rsc->role = RSC_ROLE_STARTED;
2968 
2969  // Check for migrate_from on the target
2970  target_migrate_from = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATED, target,
2971  source, PCMK_OCF_OK, data_set);
2972 
2973  if (/* If the resource state is unknown on the target, it will likely be
2974  * probed there.
2975  * Don't just consider it running there. We will get back here anyway in
2976  * case the probe detects it's running there.
2977  */
2978  !unknown_on_node(rsc->id, target, data_set)
2979  /* If the resource has newer state on the target after the migration
2980  * events, this migrate_to no longer matters for the target.
2981  */
2982  && !newer_state_after_migrate(rsc->id, target, xml_op, target_migrate_from,
2983  data_set)) {
2984  /* The resource has no newer state on the target, so assume it's still
2985  * active there.
2986  * (if it is up).
2987  */
2988  pe_node_t *target_node = pe_find_node(data_set->nodes, target);
2989 
2990  if (target_node && target_node->details->online) {
2991  native_add_running(rsc, target_node, data_set, FALSE);
2992  }
2993 
2994  } else if (!non_monitor_after(rsc->id, source, xml_op, true, data_set)) {
2995  /* We know the resource has newer state on the target, but this
2996  * migrate_to still matters for the source as long as there's no newer
2997  * non-monitor operation there.
2998  */
2999 
3000  // Mark node as having dangling migration so we can force a stop later
3001  rsc->dangling_migrations = g_list_prepend(rsc->dangling_migrations, node);
3002  }
3003 }
3004 
3005 static void
3006 unpack_migrate_from_failure(pe_resource_t *rsc, pe_node_t *node,
3007  xmlNode *xml_op, pe_working_set_t *data_set)
3008 {
3009  xmlNode *source_migrate_to = NULL;
3010  const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE);
3011  const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
3012 
3013  // Sanity check
3014  CRM_CHECK(source && target && !strcmp(target, node->details->uname), return);
3015 
3016  /* If a migration failed, we have to assume the resource is active. Clones
3017  * are not allowed to migrate, so role can't be promoted.
3018  */
3019  rsc->role = RSC_ROLE_STARTED;
3020 
3021  // Check for a migrate_to on the source
3022  source_migrate_to = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATE,
3023  source, target, PCMK_OCF_OK, data_set);
3024 
3025  if (/* If the resource state is unknown on the source, it will likely be
3026  * probed there.
3027  * Don't just consider it running there. We will get back here anyway in
3028  * case the probe detects it's running there.
3029  */
3030  !unknown_on_node(rsc->id, source, data_set)
3031  /* If the resource has newer state on the source after the migration
3032  * events, this migrate_from no longer matters for the source.
3033  */
3034  && !newer_state_after_migrate(rsc->id, source, source_migrate_to, xml_op,
3035  data_set)) {
3036  /* The resource has no newer state on the source, so assume it's still
3037  * active there (if it is up).
3038  */
3039  pe_node_t *source_node = pe_find_node(data_set->nodes, source);
3040 
3041  if (source_node && source_node->details->online) {
3042  native_add_running(rsc, source_node, data_set, TRUE);
3043  }
3044  }
3045 }
3046 
3047 static void
3048 record_failed_op(xmlNode *op, const pe_node_t *node,
3049  const pe_resource_t *rsc, pe_working_set_t *data_set)
3050 {
3051  xmlNode *xIter = NULL;
3052  const char *op_key = crm_element_value(op, XML_LRM_ATTR_TASK_KEY);
3053 
3054  if (node->details->online == FALSE) {
3055  return;
3056  }
3057 
3058  for (xIter = data_set->failed->children; xIter; xIter = xIter->next) {
3059  const char *key = crm_element_value(xIter, XML_LRM_ATTR_TASK_KEY);
3060  const char *uname = crm_element_value(xIter, XML_ATTR_UNAME);
3061 
3062  if(pcmk__str_eq(op_key, key, pcmk__str_casei) && pcmk__str_eq(uname, node->details->uname, pcmk__str_casei)) {
3063  crm_trace("Skipping duplicate entry %s on %s",
3064  op_key, pe__node_name(node));
3065  return;
3066  }
3067  }
3068 
3069  crm_trace("Adding entry %s on %s", op_key, pe__node_name(node));
3070  crm_xml_add(op, XML_ATTR_UNAME, node->details->uname);
3071  crm_xml_add(op, XML_LRM_ATTR_RSCID, rsc->id);
3072  add_node_copy(data_set->failed, op);
3073 }
3074 
3075 static const char *get_op_key(xmlNode *xml_op)
3076 {
3077  const char *key = crm_element_value(xml_op, XML_LRM_ATTR_TASK_KEY);
3078  if(key == NULL) {
3079  key = ID(xml_op);
3080  }
3081  return key;
3082 }
3083 
3084 static const char *
3085 last_change_str(const xmlNode *xml_op)
3086 {
3087  time_t when;
3088  const char *when_s = NULL;
3089 
3090  if (crm_element_value_epoch(xml_op, XML_RSC_OP_LAST_CHANGE,
3091  &when) == pcmk_ok) {
3092  when_s = pcmk__epoch2str(&when);
3093  if (when_s) {
3094  // Skip day of week to make message shorter
3095  when_s = strchr(when_s, ' ');
3096  if (when_s) {
3097  ++when_s;
3098  }
3099  }
3100  }
3101  return ((when_s && *when_s)? when_s : "unknown time");
3102 }
3103 
3116 static int
3117 cmp_on_fail(enum action_fail_response first, enum action_fail_response second)
3118 {
3119  switch (first) {
3120  case action_fail_demote:
3121  switch (second) {
3122  case action_fail_ignore:
3123  return 1;
3124  case action_fail_demote:
3125  return 0;
3126  default:
3127  return -1;
3128  }
3129  break;
3130 
3131  case action_fail_reset_remote:
3132  switch (second) {
3133  case action_fail_ignore:
3134  case action_fail_demote:
3135  case action_fail_recover:
3136  return 1;
3137  case action_fail_reset_remote:
3138  return 0;
3139  default:
3140  return -1;
3141  }
3142  break;
3143 
3144  case action_fail_restart_container:
3145  switch (second) {
3146  case action_fail_ignore:
3147  case action_fail_demote:
3148  case action_fail_recover:
3149  case action_fail_reset_remote:
3150  return 1;
3151  case action_fail_restart_container:
3152  return 0;
3153  default:
3154  return -1;
3155  }
3156  break;
3157 
3158  default:
3159  break;
3160  }
3161  switch (second) {
3162  case action_fail_demote:
3163  return (first == action_fail_ignore)? -1 : 1;
3164 
3165  case action_fail_reset_remote:
3166  switch (first) {
3167  case action_fail_ignore:
3168  case action_fail_demote:
3169  case action_fail_recover:
3170  return -1;
3171  default:
3172  return 1;
3173  }
3174  break;
3175 
3176  case action_fail_restart_container:
3177  switch (first) {
3178  case action_fail_ignore:
3179  case action_fail_demote:
3180  case action_fail_recover:
3181  case action_fail_reset_remote:
3182  return -1;
3183  default:
3184  return 1;
3185  }
3186  break;
3187 
3188  default:
3189  break;
3190  }
3191  return first - second;
3192 }
3193 
3194 static void
3195 unpack_rsc_op_failure(pe_resource_t * rsc, pe_node_t * node, int rc, xmlNode * xml_op, xmlNode ** last_failure,
3196  enum action_fail_response * on_fail, pe_working_set_t * data_set)
3197 {
3198  bool is_probe = false;
3199  pe_action_t *action = NULL;
3200 
3201  const char *key = get_op_key(xml_op);
3202  const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK);
3203  const char *exit_reason = crm_element_value(xml_op,
3204  XML_LRM_ATTR_EXIT_REASON);
3205 
3206  CRM_ASSERT(rsc);
3207  CRM_CHECK(task != NULL, return);
3208 
3209  *last_failure = xml_op;
3210 
3211  is_probe = pcmk_xe_is_probe(xml_op);
3212 
3213  if (exit_reason == NULL) {
3214  exit_reason = "";
3215  }
3216 
3217  if (!pcmk_is_set(data_set->flags, pe_flag_symmetric_cluster)
3218  && (rc == PCMK_OCF_NOT_INSTALLED)) {
3219  crm_trace("Unexpected result (%s%s%s) was recorded for "
3220  "%s of %s on %s at %s " CRM_XS " rc=%d id=%s",
3221  services_ocf_exitcode_str(rc),
3222  (*exit_reason? ": " : ""), exit_reason,
3223  (is_probe? "probe" : task), rsc->id, pe__node_name(node),
3224  last_change_str(xml_op), rc, ID(xml_op));
3225  } else {
3226  crm_warn("Unexpected result (%s%s%s) was recorded for "
3227  "%s of %s on %s at %s " CRM_XS " rc=%d id=%s",
3228  services_ocf_exitcode_str(rc),
3229  (*exit_reason? ": " : ""), exit_reason,
3230  (is_probe? "probe" : task), rsc->id, pe__node_name(node),
3231  last_change_str(xml_op), rc, ID(xml_op));
3232 
3233  if (is_probe && (rc != PCMK_OCF_OK)
3234  && (rc != PCMK_OCF_NOT_RUNNING)
3235  && (rc != PCMK_OCF_RUNNING_PROMOTED)) {
3236 
3237  /* A failed (not just unexpected) probe result could mean the user
3238  * didn't know resources will be probed even where they can't run.
3239  */
3240  crm_notice("If it is not possible for %s to run on %s, see "
3241  "the resource-discovery option for location constraints",
3242  rsc->id, pe__node_name(node));
3243  }
3244 
3245  record_failed_op(xml_op, node, rsc, data_set);
3246  }
3247 
3248  action = custom_action(rsc, strdup(key), task, NULL, TRUE, FALSE, data_set);
3249  if (cmp_on_fail(*on_fail, action->on_fail) < 0) {
3250  pe_rsc_trace(rsc, "on-fail %s -> %s for %s (%s)", fail2text(*on_fail),
3251  fail2text(action->on_fail), action->uuid, key);
3252  *on_fail = action->on_fail;
3253  }
3254 
3255  if (!strcmp(task, CRMD_ACTION_STOP)) {
3256  resource_location(rsc, node, -INFINITY, "__stop_fail__", data_set);
3257 
3258  } else if (!strcmp(task, CRMD_ACTION_MIGRATE)) {
3259  unpack_migrate_to_failure(rsc, node, xml_op, data_set);
3260 
3261  } else if (!strcmp(task, CRMD_ACTION_MIGRATED)) {
3262  unpack_migrate_from_failure(rsc, node, xml_op, data_set);
3263 
3264  } else if (!strcmp(task, CRMD_ACTION_PROMOTE)) {
3265  rsc->role = RSC_ROLE_PROMOTED;
3266 
3267  } else if (!strcmp(task, CRMD_ACTION_DEMOTE)) {
3268  if (action->on_fail == action_fail_block) {
3269  rsc->role = RSC_ROLE_PROMOTED;
3270  pe__set_next_role(rsc, RSC_ROLE_STOPPED,
3271  "demote with on-fail=block");
3272 
3273  } else if(rc == PCMK_OCF_NOT_RUNNING) {
3274  rsc->role = RSC_ROLE_STOPPED;
3275 
3276  } else {
3277  /* Staying in the promoted role would put the scheduler and
3278  * controller into a loop. Setting the role to unpromoted is not
3279  * dangerous because the resource will be stopped as part of
3280  * recovery, and any promotion will be ordered after that stop.
3281  */
3282  rsc->role = RSC_ROLE_UNPROMOTED;
3283  }
3284  }
3285 
3286  if(is_probe && rc == PCMK_OCF_NOT_INSTALLED) {
3287  /* leave stopped */
3288  pe_rsc_trace(rsc, "Leaving %s stopped", rsc->id);
3289  rsc->role = RSC_ROLE_STOPPED;
3290 
3291  } else if (rsc->role < RSC_ROLE_STARTED) {
3292  pe_rsc_trace(rsc, "Setting %s active", rsc->id);
3293  set_active(rsc);
3294  }
3295 
3296  pe_rsc_trace(rsc, "Resource %s: role=%s, unclean=%s, on_fail=%s, fail_role=%s",
3297  rsc->id, role2text(rsc->role),
3298  pcmk__btoa(node->details->unclean),
3299  fail2text(action->on_fail), role2text(action->fail_role));
3300 
3301  if (action->fail_role != RSC_ROLE_STARTED && rsc->next_role < action->fail_role) {
3302  pe__set_next_role(rsc, action->fail_role, "failure");
3303  }
3304 
3305  if (action->fail_role == RSC_ROLE_STOPPED) {
3306  int score = -INFINITY;
3307 
3308  pe_resource_t *fail_rsc = rsc;
3309 
3310  if (fail_rsc->parent) {
3311  pe_resource_t *parent = uber_parent(fail_rsc);
3312 
3313  if (pe_rsc_is_clone(parent)
3314  && !pcmk_is_set(parent->flags, pe_rsc_unique)) {
3315  /* For clone resources, if a child fails on an operation
3316  * with on-fail = stop, all the resources fail. Do this by preventing
3317  * the parent from coming up again. */
3318  fail_rsc = parent;
3319  }
3320  }
3321  crm_notice("%s will not be started under current conditions",
3322  fail_rsc->id);
3323  /* make sure it doesn't come up again */
3324  if (fail_rsc->allowed_nodes != NULL) {
3325  g_hash_table_destroy(fail_rsc->allowed_nodes);
3326  }
3327  fail_rsc->allowed_nodes = pe__node_list2table(data_set->nodes);
3328  g_hash_table_foreach(fail_rsc->allowed_nodes, set_node_score, &score);
3329  }
3330 
3331  pe_free_action(action);
3332 }
3333 
3347 static void
3348 check_recoverable(pe_resource_t *rsc, pe_node_t *node, const char *task,
3349  int exit_status, const xmlNode *xml_op)
3350 {
3351  const char *exit_reason = NULL;
3352 
3353  if (strcmp(task, CRMD_ACTION_STOP) != 0) {
3354  return; // All actions besides stop are always recoverable
3355  }
3356  if (pe_can_fence(node->details->data_set, node)) {
3357  return; // Failed stops are recoverable via fencing
3358  }
3359 
3360  exit_reason = crm_element_value(xml_op, XML_LRM_ATTR_EXIT_REASON);
3361  pe_proc_err("No further recovery can be attempted for %s "
3362  "because %s on %s failed (%s%s%s) at %s "
3363  CRM_XS " rc=%d id=%s", rsc->id, task, pe__node_name(node),
3364  services_ocf_exitcode_str(exit_status),
3365  ((exit_reason == NULL)? "" : ": "), pcmk__s(exit_reason, ""),
3366  last_change_str(xml_op), exit_status, ID(xml_op));
3367 
3368  pe__clear_resource_flags(rsc, pe_rsc_managed);
3369  pe__set_resource_flags(rsc, pe_rsc_block);
3370 }
3371 
3399 static void
3400 remap_operation(xmlNode *xml_op, pe_resource_t *rsc, pe_node_t *node,
3401  pe_working_set_t *data_set, enum action_fail_response *on_fail,
3402  int target_rc, int *rc, int *status) {
3403  bool is_probe = false;
3404  const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK);
3405  const char *key = get_op_key(xml_op);
3406  const char *exit_reason = crm_element_value(xml_op,
3407  XML_LRM_ATTR_EXIT_REASON);
3408 
3409  if (pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_none)) {
3410  int remapped_rc = pcmk__effective_rc(*rc);
3411 
3412  if (*rc != remapped_rc) {
3413  crm_trace("Remapping monitor result %d to %d", *rc, remapped_rc);
3414  if (!node->details->shutdown || node->details->online) {
3415  record_failed_op(xml_op, node, rsc, data_set);
3416  }
3417 
3418  *rc = remapped_rc;
3419  }
3420  }
3421 
3422  if (!pe_rsc_is_bundled(rsc) && pcmk_xe_mask_probe_failure(xml_op)) {
3423  *status = PCMK_EXEC_DONE;
3424  *rc = PCMK_OCF_NOT_RUNNING;
3425  }
3426 
3427  /* If the executor reported an operation status of anything but done or
3428  * error, consider that final. But for done or error, we know better whether
3429  * it should be treated as a failure or not, because we know the expected
3430  * result.
3431  */
3432  if (*status != PCMK_EXEC_DONE && *status != PCMK_EXEC_ERROR) {
3433  return;
3434  }
3435 
3436  CRM_ASSERT(rsc);
3437  CRM_CHECK(task != NULL,
3438  *status = PCMK_EXEC_ERROR; return);
3439 
3440  *status = PCMK_EXEC_DONE;
3441 
3442  if (exit_reason == NULL) {
3443  exit_reason = "";
3444  }
3445 
3446  is_probe = pcmk_xe_is_probe(xml_op);
3447 
3448  if (is_probe) {
3449  task = "probe";
3450  }
3451 
3452  if (target_rc < 0) {
3453  /* Pre-1.0 Pacemaker versions, and Pacemaker 1.1.6 or earlier with
3454  * Heartbeat 2.0.7 or earlier as the cluster layer, did not include the
3455  * target_rc in the transition key, which (along with the similar case
3456  * of a corrupted transition key in the CIB) will be reported to this
3457  * function as -1. Pacemaker 2.0+ does not support rolling upgrades from
3458  * those versions or processing of saved CIB files from those versions,
3459  * so we do not need to care much about this case.
3460  */
3461  *status = PCMK_EXEC_ERROR;
3462  crm_warn("Expected result not found for %s on %s (corrupt or obsolete CIB?)",
3463  key, pe__node_name(node));
3464 
3465  } else if (target_rc != *rc) {
3466  *status = PCMK_EXEC_ERROR;
3467  pe_rsc_debug(rsc, "%s on %s: expected %d (%s), got %d (%s%s%s)",
3468  key, pe__node_name(node),
3469  target_rc, services_ocf_exitcode_str(target_rc),
3470  *rc, services_ocf_exitcode_str(*rc),
3471  (*exit_reason? ": " : ""), exit_reason);
3472  }
3473 
3474  switch (*rc) {
3475  case PCMK_OCF_OK:
3476  if (is_probe && (target_rc == PCMK_OCF_NOT_RUNNING)) {
3477  *status = PCMK_EXEC_DONE;
3478  pe_rsc_info(rsc, "Probe found %s active on %s at %s",
3479  rsc->id, pe__node_name(node),
3480  last_change_str(xml_op));
3481  }
3482  break;
3483 
3484  case PCMK_OCF_NOT_RUNNING:
3485  if (is_probe || (target_rc == *rc)
3486  || !pcmk_is_set(rsc->flags, pe_rsc_managed)) {
3487 
3488  *status = PCMK_EXEC_DONE;
3489  rsc->role = RSC_ROLE_STOPPED;
3490 
3491  /* clear any previous failure actions */
3492  *on_fail = action_fail_ignore;
3493  pe__set_next_role(rsc, RSC_ROLE_UNKNOWN, "not running");
3494  }
3495  break;
3496 
3497  case PCMK_OCF_RUNNING_PROMOTED:
3498  if (is_probe && (*rc != target_rc)) {
3499  *status = PCMK_EXEC_DONE;
3500  pe_rsc_info(rsc,
3501  "Probe found %s active and promoted on %s at %s",
3502  rsc->id, pe__node_name(node),
3503  last_change_str(xml_op));
3504  }
3505  rsc->role = RSC_ROLE_PROMOTED;
3506  break;
3507 
3508  case PCMK_OCF_DEGRADED_PROMOTED:
3509  case PCMK_OCF_FAILED_PROMOTED:
3510  rsc->role = RSC_ROLE_PROMOTED;
3511  *status = PCMK_EXEC_ERROR;
3512  break;
3513 
3514  case PCMK_OCF_NOT_CONFIGURED:
3515  *status = PCMK_EXEC_ERROR_FATAL;
3516  break;
3517 
3518  case PCMK_OCF_UNIMPLEMENT_FEATURE:
3519  {
3520  guint interval_ms = 0;
3521  crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS,
3522  &interval_ms);
3523 
3524  if (interval_ms == 0) {
3525  check_recoverable(rsc, node, task, *rc, xml_op);
3526  *status = PCMK_EXEC_ERROR_HARD;
3527  } else {
3528  *status = PCMK_EXEC_NOT_SUPPORTED;
3529  }
3530  }
3531  break;
3532 
3533  case PCMK_OCF_NOT_INSTALLED:
3534  case PCMK_OCF_INVALID_PARAM:
3535  case PCMK_OCF_INSUFFICIENT_PRIV:
3536  check_recoverable(rsc, node, task, *rc, xml_op);
3537  *status = PCMK_EXEC_ERROR_HARD;
3538  break;
3539 
3540  default:
3541  if (*status == PCMK_EXEC_DONE) {
3542  crm_info("Treating unknown exit status %d from %s of %s "
3543  "on %s at %s as failure",
3544  *rc, task, rsc->id, pe__node_name(node),
3545  last_change_str(xml_op));
3546  *status = PCMK_EXEC_ERROR;
3547  }
3548  break;
3549  }
3550 
3551  pe_rsc_trace(rsc, "Remapped %s status to '%s'",
3552  key, pcmk_exec_status_str(*status));
3553 }
3554 
3555 // return TRUE if start or monitor last failure but parameters changed
3556 static bool
3557 should_clear_for_param_change(xmlNode *xml_op, const char *task,
3558  pe_resource_t *rsc, pe_node_t *node,
3559  pe_working_set_t *data_set)
3560 {
3561  if (!strcmp(task, "start") || !strcmp(task, "monitor")) {
3562 
3563  if (pe__bundle_needs_remote_name(rsc, data_set)) {
3564  /* We haven't allocated resources yet, so we can't reliably
3565  * substitute addr parameters for the REMOTE_CONTAINER_HACK.
3566  * When that's needed, defer the check until later.
3567  */
3568  pe__add_param_check(xml_op, rsc, node, pe_check_last_failure,
3569  data_set);
3570 
3571  } else {
3572  op_digest_cache_t *digest_data = NULL;
3573 
3574  digest_data = rsc_action_digest_cmp(rsc, xml_op, node, data_set);
3575  switch (digest_data->rc) {
3576  case RSC_DIGEST_UNKNOWN:
3577  crm_trace("Resource %s history entry %s on %s"
3578  " has no digest to compare",
3579  rsc->id, get_op_key(xml_op), node->details->id);
3580  break;
3581  case RSC_DIGEST_MATCH:
3582  break;
3583  default:
3584  return TRUE;
3585  }
3586  }
3587  }
3588  return FALSE;
3589 }
3590 
3591 // Order action after fencing of remote node, given connection rsc
3592 static void
3593 order_after_remote_fencing(pe_action_t *action, pe_resource_t *remote_conn,
3594  pe_working_set_t *data_set)
3595 {
3596  pe_node_t *remote_node = pe_find_node(data_set->nodes, remote_conn->id);
3597 
3598  if (remote_node) {
3599  pe_action_t *fence = pe_fence_op(remote_node, NULL, TRUE, NULL,
3600  FALSE, data_set);
3601 
3602  order_actions(fence, action, pe_order_implies_then);
3603  }
3604 }
3605 
3606 static bool
3607 should_ignore_failure_timeout(pe_resource_t *rsc, xmlNode *xml_op,
3608  const char *task, guint interval_ms,
3609  bool is_last_failure, pe_working_set_t *data_set)
3610 {
3611  /* Clearing failures of recurring monitors has special concerns. The
3612  * executor reports only changes in the monitor result, so if the
3613  * monitor is still active and still getting the same failure result,
3614  * that will go undetected after the failure is cleared.
3615  *
3616  * Also, the operation history will have the time when the recurring
3617  * monitor result changed to the given code, not the time when the
3618  * result last happened.
3619  *
3620  * @TODO We probably should clear such failures only when the failure
3621  * timeout has passed since the last occurrence of the failed result.
3622  * However we don't record that information. We could maybe approximate
3623  * that by clearing only if there is a more recent successful monitor or
3624  * stop result, but we don't even have that information at this point
3625  * since we are still unpacking the resource's operation history.
3626  *
3627  * This is especially important for remote connection resources with a
3628  * reconnect interval, so in that case, we skip clearing failures
3629  * if the remote node hasn't been fenced.
3630  */
3631  if (rsc->remote_reconnect_ms
3632  && pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)
3633  && (interval_ms != 0) && pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_casei)) {
3634 
3635  pe_node_t *remote_node = pe_find_node(data_set->nodes, rsc->id);
3636 
3637  if (remote_node && !remote_node->details->remote_was_fenced) {
3638  if (is_last_failure) {
3639  crm_info("Waiting to clear monitor failure for remote node %s"
3640  " until fencing has occurred", rsc->id);
3641  }
3642  return TRUE;
3643  }
3644  }
3645  return FALSE;
3646 }
3647 
3670 static bool
3671 check_operation_expiry(pe_resource_t *rsc, pe_node_t *node, int rc,
3672  xmlNode *xml_op, pe_working_set_t *data_set)
3673 {
3674  bool expired = FALSE;
3675  bool is_last_failure = pcmk__ends_with(ID(xml_op), "_last_failure_0");
3676  time_t last_run = 0;
3677  guint interval_ms = 0;
3678  int unexpired_fail_count = 0;
3679  const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK);
3680  const char *clear_reason = NULL;
3681 
3682  crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms);
3683 
3684  if ((rsc->failure_timeout > 0)
3685  && (crm_element_value_epoch(xml_op, XML_RSC_OP_LAST_CHANGE,
3686  &last_run) == 0)) {
3687 
3688  // Resource has a failure-timeout, and history entry has a timestamp
3689 
3690  time_t now = get_effective_time(data_set);
3691  time_t last_failure = 0;
3692 
3693  // Is this particular operation history older than the failure timeout?
3694  if ((now >= (last_run + rsc->failure_timeout))
3695  && !should_ignore_failure_timeout(rsc, xml_op, task, interval_ms,
3696  is_last_failure, data_set)) {
3697  expired = TRUE;
3698  }
3699 
3700  // Does the resource as a whole have an unexpired fail count?
3701  unexpired_fail_count = pe_get_failcount(node, rsc, &last_failure,
3702  pe_fc_effective, xml_op,
3703  data_set);
3704 
3705  // Update scheduler recheck time according to *last* failure
3706  crm_trace("%s@%lld is %sexpired @%lld with unexpired_failures=%d timeout=%ds"
3707  " last-failure@%lld",
3708  ID(xml_op), (long long) last_run, (expired? "" : "not "),
3709  (long long) now, unexpired_fail_count, rsc->failure_timeout,
3710  (long long) last_failure);
3711  last_failure += rsc->failure_timeout + 1;
3712  if (unexpired_fail_count && (now < last_failure)) {
3713  pe__update_recheck_time(last_failure, data_set);
3714  }
3715  }
3716 
3717  if (expired) {
3718  if (pe_get_failcount(node, rsc, NULL, pe_fc_default, xml_op, data_set)) {
3719 
3720  // There is a fail count ignoring timeout
3721 
3722  if (unexpired_fail_count == 0) {
3723  // There is no fail count considering timeout
3724  clear_reason = "it expired";
3725 
3726  } else {
3727  /* This operation is old, but there is an unexpired fail count.
3728  * In a properly functioning cluster, this should only be
3729  * possible if this operation is not a failure (otherwise the
3730  * fail count should be expired too), so this is really just a
3731  * failsafe.
3732  */
3733  expired = FALSE;
3734  }
3735 
3736  } else if (is_last_failure && rsc->remote_reconnect_ms) {
3737  /* Clear any expired last failure when reconnect interval is set,
3738  * even if there is no fail count.
3739  */
3740  clear_reason = "reconnect interval is set";
3741  }
3742  }
3743 
3744  if (!expired && is_last_failure
3745  && should_clear_for_param_change(xml_op, task, rsc, node, data_set)) {
3746  clear_reason = "resource parameters have changed";
3747  }
3748 
3749  if (clear_reason != NULL) {
3750  // Schedule clearing of the fail count
3751  pe_action_t *clear_op = pe__clear_failcount(rsc, node, clear_reason,
3752  data_set);
3753 
3754  if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)
3755  && rsc->remote_reconnect_ms) {
3756  /* If we're clearing a remote connection due to a reconnect
3757  * interval, we want to wait until any scheduled fencing
3758  * completes.
3759  *
3760  * We could limit this to remote_node->details->unclean, but at
3761  * this point, that's always true (it won't be reliable until
3762  * after unpack_node_history() is done).
3763  */
3764  crm_info("Clearing %s failure will wait until any scheduled "
3765  "fencing of %s completes", task, rsc->id);
3766  order_after_remote_fencing(clear_op, rsc, data_set);
3767  }
3768  }
3769 
3770  if (expired && (interval_ms == 0) && pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_casei)) {
3771  switch(rc) {
3772  case PCMK_OCF_OK:
3773  case PCMK_OCF_NOT_RUNNING:
3774  case PCMK_OCF_RUNNING_PROMOTED:
3775  case PCMK_OCF_DEGRADED:
3776  case PCMK_OCF_DEGRADED_PROMOTED:
3777  // Don't expire probes that return these values
3778  expired = FALSE;
3779  break;
3780  }
3781  }
3782 
3783  return expired;
3784 }
3785 
3786 int pe__target_rc_from_xml(xmlNode *xml_op)
3787 {
3788  int target_rc = 0;
3789  const char *key = crm_element_value(xml_op, XML_ATTR_TRANSITION_KEY);
3790 
3791  if (key == NULL) {
3792  return -1;
3793  }
3794  decode_transition_key(key, NULL, NULL, NULL, &target_rc);
3795  return target_rc;
3796 }
3797 
3798 static enum action_fail_response
3799 get_action_on_fail(pe_resource_t *rsc, const char *key, const char *task, pe_working_set_t * data_set)
3800 {
3801  enum action_fail_response result = action_fail_recover;
3802  pe_action_t *action = custom_action(rsc, strdup(key), task, NULL, TRUE, FALSE, data_set);
3803 
3804  result = action->on_fail;
3805  pe_free_action(action);
3806 
3807  return result;
3808 }
3809 
3810 static void
3811 update_resource_state(pe_resource_t * rsc, pe_node_t * node, xmlNode * xml_op, const char * task, int rc,
3812  xmlNode * last_failure, enum action_fail_response * on_fail, pe_working_set_t * data_set)
3813 {
3814  gboolean clear_past_failure = FALSE;
3815 
3816  CRM_ASSERT(rsc);
3817  CRM_ASSERT(xml_op);
3818 
3819  if (rc == PCMK_OCF_NOT_INSTALLED || (!pe_rsc_is_bundled(rsc) && pcmk_xe_mask_probe_failure(xml_op))) {
3820  rsc->role = RSC_ROLE_STOPPED;
3821 
3822  } else if (rc == PCMK_OCF_NOT_RUNNING) {
3823  clear_past_failure = TRUE;
3824 
3825  } else if (pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_casei)) {
3826  if (last_failure) {
3827  const char *op_key = get_op_key(xml_op);
3828  const char *last_failure_key = get_op_key(last_failure);
3829 
3830  if (pcmk__str_eq(op_key, last_failure_key, pcmk__str_casei)) {
3831  clear_past_failure = TRUE;
3832  }
3833  }
3834 
3835  if (rsc->role < RSC_ROLE_STARTED) {
3836  set_active(rsc);
3837  }
3838 
3839  } else if (pcmk__str_eq(task, CRMD_ACTION_START, pcmk__str_casei)) {
3840  rsc->role = RSC_ROLE_STARTED;
3841  clear_past_failure = TRUE;
3842 
3843  } else if (pcmk__str_eq(task, CRMD_ACTION_STOP, pcmk__str_casei)) {
3844  rsc->role = RSC_ROLE_STOPPED;
3845  clear_past_failure = TRUE;
3846 
3847  } else if (pcmk__str_eq(task, CRMD_ACTION_PROMOTE, pcmk__str_casei)) {
3848  rsc->role = RSC_ROLE_PROMOTED;
3849  clear_past_failure = TRUE;
3850 
3851  } else if (pcmk__str_eq(task, CRMD_ACTION_DEMOTE, pcmk__str_casei)) {
3852 
3853  if (*on_fail == action_fail_demote) {
3854  // Demote clears an error only if on-fail=demote
3855  clear_past_failure = TRUE;
3856  }
3857  rsc->role = RSC_ROLE_UNPROMOTED;
3858 
3859  } else if (pcmk__str_eq(task, CRMD_ACTION_MIGRATED, pcmk__str_casei)) {
3860  rsc->role = RSC_ROLE_STARTED;
3861  clear_past_failure = TRUE;
3862 
3863  } else if (pcmk__str_eq(task, CRMD_ACTION_MIGRATE, pcmk__str_casei)) {
3864  unpack_migrate_to_success(rsc, node, xml_op, data_set);
3865 
3866  } else if (rsc->role < RSC_ROLE_STARTED) {
3867  pe_rsc_trace(rsc, "%s active on %s", rsc->id, pe__node_name(node));
3868  set_active(rsc);
3869  }
3870 
3871  /* clear any previous failure actions */
3872  if (clear_past_failure) {
3873  switch (*on_fail) {
3874  case action_fail_stop:
3875  case action_fail_fence:
3876  case action_fail_migrate:
3877  case action_fail_standby:
3878  pe_rsc_trace(rsc, "%s.%s is not cleared by a completed stop",
3879  rsc->id, fail2text(*on_fail));
3880  break;
3881 
3882  case action_fail_block:
3883  case action_fail_ignore:
3884  case action_fail_demote:
3885  case action_fail_recover:
3886  case action_fail_restart_container:
3887  *on_fail = action_fail_ignore;
3888  pe__set_next_role(rsc, RSC_ROLE_UNKNOWN, "clear past failures");
3889  break;
3890  case action_fail_reset_remote:
3891  if (rsc->remote_reconnect_ms == 0) {
3892  /* With no reconnect interval, the connection is allowed to
3893  * start again after the remote node is fenced and
3894  * completely stopped. (With a reconnect interval, we wait
3895  * for the failure to be cleared entirely before attempting
3896  * to reconnect.)
3897  */
3898  *on_fail = action_fail_ignore;
3899  pe__set_next_role(rsc, RSC_ROLE_UNKNOWN,
3900  "clear past failures and reset remote");
3901  }
3902  break;
3903  }
3904  }
3905 }
3906 
3907 static void
3908 unpack_rsc_op(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op,
3909  xmlNode **last_failure, enum action_fail_response *on_fail,
3910  pe_working_set_t *data_set)
3911 {
3912  int rc = 0;
3913  int old_rc = 0;
3914  int task_id = 0;
3915  int target_rc = 0;
3916  int old_target_rc = 0;
3917  int status = PCMK_EXEC_UNKNOWN;
3918  guint interval_ms = 0;
3919  const char *task = NULL;
3920  const char *task_key = NULL;
3921  const char *exit_reason = NULL;
3922  bool expired = false;
3923  pe_resource_t *parent = rsc;
3924  enum action_fail_response failure_strategy = action_fail_recover;
3925  bool maskable_probe_failure = false;
3926 
3927  CRM_CHECK(rsc && node && xml_op, return);
3928 
3929  target_rc = pe__target_rc_from_xml(xml_op);
3930  task_key = get_op_key(xml_op);
3931  task = crm_element_value(xml_op, XML_LRM_ATTR_TASK);
3932  exit_reason = crm_element_value(xml_op, XML_LRM_ATTR_EXIT_REASON);
3933  if (exit_reason == NULL) {
3934  exit_reason = "";
3935  }
3936 
3937  crm_element_value_int(xml_op, XML_LRM_ATTR_RC, &rc);
3938  crm_element_value_int(xml_op, XML_LRM_ATTR_CALLID, &task_id);
3939  crm_element_value_int(xml_op, XML_LRM_ATTR_OPSTATUS, &status);
3940  crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms);
3941 
3942  CRM_CHECK(task != NULL, return);
3943  CRM_CHECK((status >= PCMK_EXEC_PENDING) && (status <= PCMK_EXEC_MAX),
3944  return);
3945 
3946  if (!strcmp(task, CRMD_ACTION_NOTIFY) ||
3947  !strcmp(task, CRMD_ACTION_METADATA)) {
3948  /* safe to ignore these */
3949  return;
3950  }
3951 
3952  if (!pcmk_is_set(rsc->flags, pe_rsc_unique)) {
3953  parent = uber_parent(rsc);
3954  }
3955 
3956  pe_rsc_trace(rsc, "Unpacking task %s/%s (call_id=%d, status=%d, rc=%d) on %s (role=%s)",
3957  task_key, task, task_id, status, rc, pe__node_name(node),
3958  role2text(rsc->role));
3959 
3960  if (node->details->unclean) {
3961  pe_rsc_trace(rsc,
3962  "%s is running on %s, which is unclean (further action "
3963  "depends on value of stop's on-fail attribute)",
3964  rsc->id, pe__node_name(node));
3965  }
3966 
3967  /* It should be possible to call remap_operation() first then call
3968  * check_operation_expiry() only if rc != target_rc, because there should
3969  * never be a fail count without at least one unexpected result in the
3970  * resource history. That would be more efficient by avoiding having to call
3971  * check_operation_expiry() for expected results.
3972  *
3973  * However, we do have such configurations in the scheduler regression
3974  * tests, even if it shouldn't be possible with the current code. It's
3975  * probably a good idea anyway, but that would require updating the test
3976  * inputs to something currently possible.
3977  */
3978 
3979  if ((status != PCMK_EXEC_NOT_INSTALLED)
3980  && check_operation_expiry(rsc, node, rc, xml_op, data_set)) {
3981  expired = true;
3982  }
3983 
3984  old_rc = rc;
3985  old_target_rc = target_rc;
3986 
3987  remap_operation(xml_op, rsc, node, data_set, on_fail, target_rc,
3988  &rc, &status);
3989 
3990  maskable_probe_failure = !pe_rsc_is_bundled(rsc) && pcmk_xe_mask_probe_failure(xml_op);
3991 
3992  if (expired && maskable_probe_failure && old_rc != old_target_rc) {
3993  if (rsc->role <= RSC_ROLE_STOPPED) {
3994  rsc->role = RSC_ROLE_UNKNOWN;
3995  }
3996 
3997  goto done;
3998 
3999  } else if (expired && (rc != target_rc)) {
4000  const char *magic = crm_element_value(xml_op, XML_ATTR_TRANSITION_MAGIC);
4001 
4002  if (interval_ms == 0) {
4003  crm_notice("Ignoring expired %s failure on %s "
4004  CRM_XS " actual=%d expected=%d magic=%s",
4005  task_key, pe__node_name(node), rc, target_rc, magic);
4006  goto done;
4007 
4008  } else if(node->details->online && node->details->unclean == FALSE) {
4009  /* Reschedule the recurring monitor. schedule_cancel() won't work at
4010  * this stage, so as a hacky workaround, forcibly change the restart
4011  * digest so pcmk__check_action_config() does what we want later.
4012  *
4013  * @TODO We should skip this if there is a newer successful monitor.
4014  * Also, this causes rescheduling only if the history entry
4015  * has an op-digest (which the expire-non-blocked-failure
4016  * scheduler regression test doesn't, but that may not be a
4017  * realistic scenario in production).
4018  */
4019  crm_notice("Rescheduling %s after failure expired on %s "
4020  CRM_XS " actual=%d expected=%d magic=%s",
4021  task_key, pe__node_name(node), rc, target_rc, magic);
4022  crm_xml_add(xml_op, XML_LRM_ATTR_RESTART_DIGEST, "calculated-failure-timeout");
4023  goto done;
4024  }
4025  }
4026 
4027  if (maskable_probe_failure) {
4028  crm_notice("Treating probe result '%s' for %s on %s as 'not running'",
4029  services_ocf_exitcode_str(old_rc), rsc->id,
4030  pe__node_name(node));
4031  update_resource_state(rsc, node, xml_op, task, target_rc, *last_failure,
4032  on_fail, data_set);
4033  crm_xml_add(xml_op, XML_ATTR_UNAME, node->details->uname);
4034 
4035  record_failed_op(xml_op, node, rsc, data_set);
4036  resource_location(parent, node, -INFINITY, "masked-probe-failure", data_set);
4037  goto done;
4038  }
4039 
4040  switch (status) {
4041  case PCMK_EXEC_CANCELLED:
4042  // Should never happen
4043  pe_err("Resource history contains cancellation '%s' "
4044  "(%s of %s on %s at %s)",
4045  ID(xml_op), task, rsc->id, pe__node_name(node),
4046  last_change_str(xml_op));
4047  goto done;
4048 
4049  case PCMK_EXEC_PENDING:
4050  if (!strcmp(task, CRMD_ACTION_START)) {
4051  pe__set_resource_flags(rsc, pe_rsc_start_pending);
4052  set_active(rsc);
4053 
4054  } else if (!strcmp(task, CRMD_ACTION_PROMOTE)) {
4055  rsc->role = RSC_ROLE_PROMOTED;
4056 
4057  } else if (!strcmp(task, CRMD_ACTION_MIGRATE) && node->details->unclean) {
4058  /* If a pending migrate_to action is out on a unclean node,
4059  * we have to force the stop action on the target. */
4060  const char *migrate_target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
4061  pe_node_t *target = pe_find_node(data_set->nodes, migrate_target);
4062  if (target) {
4063  stop_action(rsc, target, FALSE);
4064  }
4065  }
4066 
4067  if (rsc->pending_task == NULL) {
4068  if ((interval_ms != 0) || strcmp(task, CRMD_ACTION_STATUS)) {
4069  rsc->pending_task = strdup(task);
4070  rsc->pending_node = node;
4071  } else {
4072  /* Pending probes are not printed, even if pending
4073  * operations are requested. If someone ever requests that
4074  * behavior, enable the below and the corresponding part of
4075  * native.c:native_pending_task().
4076  */
4077 #if 0
4078  rsc->pending_task = strdup("probe");
4079  rsc->pending_node = node;
4080 #endif
4081  }
4082  }
4083  goto done;
4084 
4085  case PCMK_EXEC_DONE:
4086  pe_rsc_trace(rsc, "%s of %s on %s completed at %s " CRM_XS " id=%s",
4087  task, rsc->id, pe__node_name(node),
4088  last_change_str(xml_op), ID(xml_op));
4089  update_resource_state(rsc, node, xml_op, task, rc, *last_failure, on_fail, data_set);
4090  goto done;
4091 
4092  case PCMK_EXEC_NOT_INSTALLED:
4093  failure_strategy = get_action_on_fail(rsc, task_key, task, data_set);
4094  if (failure_strategy == action_fail_ignore) {
4095  crm_warn("Cannot ignore failed %s of %s on %s: "
4096  "Resource agent doesn't exist "
4097  CRM_XS " status=%d rc=%d id=%s",
4098  task, rsc->id, pe__node_name(node), status, rc,
4099  ID(xml_op));
4100  /* Also for printing it as "FAILED" by marking it as pe_rsc_failed later */
4101  *on_fail = action_fail_migrate;
4102  }
4103  resource_location(parent, node, -INFINITY, "hard-error", data_set);
4104  unpack_rsc_op_failure(rsc, node, rc, xml_op, last_failure, on_fail, data_set);
4105  goto done;
4106 
4107  case PCMK_EXEC_NOT_CONNECTED:
4108  if (pe__is_guest_or_remote_node(node)
4109  && pcmk_is_set(node->details->remote_rsc->flags, pe_rsc_managed)) {
4110  /* We should never get into a situation where a managed remote
4111  * connection resource is considered OK but a resource action
4112  * behind the connection gets a "not connected" status. But as a
4113  * fail-safe in case a bug or unusual circumstances do lead to
4114  * that, ensure the remote connection is considered failed.
4115  */
4116  pe__set_resource_flags(node->details->remote_rsc,
4117  pe_rsc_failed|pe_rsc_stop);
4118  }
4119  break; // Not done, do error handling
4120 
4121  case PCMK_EXEC_ERROR:
4122  case PCMK_EXEC_ERROR_HARD:
4123  case PCMK_EXEC_ERROR_FATAL:
4124  case PCMK_EXEC_TIMEOUT:
4125  case PCMK_EXEC_NOT_SUPPORTED:
4126  case PCMK_EXEC_INVALID:
4127  break; // Not done, do error handling
4128 
4129  case PCMK_EXEC_NO_FENCE_DEVICE:
4130  case PCMK_EXEC_NO_SECRETS:
4131  status = PCMK_EXEC_ERROR_HARD;
4132  break; // Not done, do error handling
4133  }
4134 
4135  failure_strategy = get_action_on_fail(rsc, task_key, task, data_set);
4136  if ((failure_strategy == action_fail_ignore)
4137  || (failure_strategy == action_fail_restart_container
4138  && !strcmp(task, CRMD_ACTION_STOP))) {
4139 
4140  crm_warn("Pretending failed %s (%s%s%s) of %s on %s at %s "
4141  "succeeded " CRM_XS " rc=%d id=%s",
4142  task, services_ocf_exitcode_str(rc),
4143  (*exit_reason? ": " : ""), exit_reason, rsc->id,
4144  pe__node_name(node), last_change_str(xml_op), rc,
4145  ID(xml_op));
4146 
4147  update_resource_state(rsc, node, xml_op, task, target_rc, *last_failure,
4148  on_fail, data_set);
4149  crm_xml_add(xml_op, XML_ATTR_UNAME, node->details->uname);
4150  pe__set_resource_flags(rsc, pe_rsc_failure_ignored);
4151 
4152  record_failed_op(xml_op, node, rsc, data_set);
4153 
4154  if ((failure_strategy == action_fail_restart_container)
4155  && cmp_on_fail(*on_fail, action_fail_recover) <= 0) {
4156  *on_fail = failure_strategy;
4157  }
4158 
4159  } else {
4160  unpack_rsc_op_failure(rsc, node, rc, xml_op, last_failure, on_fail,
4161  data_set);
4162 
4163  if (status == PCMK_EXEC_ERROR_HARD) {
4164  do_crm_log(rc != PCMK_OCF_NOT_INSTALLED?LOG_ERR:LOG_NOTICE,
4165  "Preventing %s from restarting on %s because "
4166  "of hard failure (%s%s%s)" CRM_XS " rc=%d id=%s",
4167  parent->id, pe__node_name(node),
4168  services_ocf_exitcode_str(rc),
4169  (*exit_reason? ": " : ""), exit_reason,
4170  rc, ID(xml_op));
4171  resource_location(parent, node, -INFINITY, "hard-error", data_set);
4172 
4173  } else if (status == PCMK_EXEC_ERROR_FATAL) {
4174  crm_err("Preventing %s from restarting anywhere because "
4175  "of fatal failure (%s%s%s) " CRM_XS " rc=%d id=%s",
4176  parent->id, services_ocf_exitcode_str(rc),
4177  (*exit_reason? ": " : ""), exit_reason,
4178  rc, ID(xml_op));
4179  resource_location(parent, NULL, -INFINITY, "fatal-error", data_set);
4180  }
4181  }
4182 
4183 done:
4184  pe_rsc_trace(rsc, "Resource %s after %s: role=%s, next=%s",
4185  rsc->id, task, role2text(rsc->role),
4186  role2text(rsc->next_role));
4187 }
4188 
4189 static void
4190 add_node_attrs(xmlNode *xml_obj, pe_node_t *node, bool overwrite,
4191  pe_working_set_t *data_set)
4192 {
4193  const char *cluster_name = NULL;
4194 
4195  pe_rule_eval_data_t rule_data = {
4196  .node_hash = NULL,
4197  .role = RSC_ROLE_UNKNOWN,
4198  .now = data_set->now,
4199  .match_data = NULL,
4200  .rsc_data = NULL,
4201  .op_data = NULL
4202  };
4203 
4204  g_hash_table_insert(node->details->attrs,
4205  strdup(CRM_ATTR_UNAME), strdup(node->details->uname));
4206 
4207  g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_ID),
4208  strdup(node->details->id));
4209  if (pcmk__str_eq(node->details->id, data_set->dc_uuid, pcmk__str_casei)) {
4210  data_set->dc_node = node;
4211  node->details->is_dc = TRUE;
4212  g_hash_table_insert(node->details->attrs,
4213  strdup(CRM_ATTR_IS_DC), strdup(XML_BOOLEAN_TRUE));
4214  } else {
4215  g_hash_table_insert(node->details->attrs,
4216  strdup(CRM_ATTR_IS_DC), strdup(XML_BOOLEAN_FALSE));
4217  }
4218 
4219  cluster_name = g_hash_table_lookup(data_set->config_hash, "cluster-name");
4220  if (cluster_name) {
4221  g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_CLUSTER_NAME),
4222  strdup(cluster_name));
4223  }
4224 
4225  pe__unpack_dataset_nvpairs(xml_obj, XML_TAG_ATTR_SETS, &rule_data,
4226  node->details->attrs, NULL, overwrite, data_set);
4227 
4228  if (pe_node_attribute_raw(node, CRM_ATTR_SITE_NAME) == NULL) {
4229  const char *site_name = pe_node_attribute_raw(node, "site-name");
4230 
4231  if (site_name) {
4232  g_hash_table_insert(node->details->attrs,
4233  strdup(CRM_ATTR_SITE_NAME),
4234  strdup(site_name));
4235 
4236  } else if (cluster_name) {
4237  /* Default to cluster-name if unset */
4238  g_hash_table_insert(node->details->attrs,
4239  strdup(CRM_ATTR_SITE_NAME),
4240  strdup(cluster_name));
4241  }
4242  }
4243 }
4244 
4245 static GList *
4246 extract_operations(const char *node, const char *rsc, xmlNode * rsc_entry, gboolean active_filter)
4247 {
4248  int counter = -1;
4249  int stop_index = -1;
4250  int start_index = -1;
4251 
4252  xmlNode *rsc_op = NULL;
4253 
4254  GList *gIter = NULL;
4255  GList *op_list = NULL;
4256  GList *sorted_op_list = NULL;
4257 
4258  /* extract operations */
4259  op_list = NULL;
4260  sorted_op_list = NULL;
4261 
4262  for (rsc_op = pcmk__xe_first_child(rsc_entry);
4263  rsc_op != NULL; rsc_op = pcmk__xe_next(rsc_op)) {
4264 
4265  if (pcmk__str_eq((const char *)rsc_op->name, XML_LRM_TAG_RSC_OP,
4266  pcmk__str_none)) {
4267  crm_xml_add(rsc_op, "resource", rsc);
4268  crm_xml_add(rsc_op, XML_ATTR_UNAME, node);
4269  op_list = g_list_prepend(op_list, rsc_op);
4270  }
4271  }
4272 
4273  if (op_list == NULL) {
4274  /* if there are no operations, there is nothing to do */
4275  return NULL;
4276  }
4277 
4278  sorted_op_list = g_list_sort(op_list, sort_op_by_callid);
4279 
4280  /* create active recurring operations as optional */
4281  if (active_filter == FALSE) {
4282  return sorted_op_list;
4283  }
4284 
4285  op_list = NULL;
4286 
4287  calculate_active_ops(sorted_op_list, &start_index, &stop_index);
4288 
4289  for (gIter = sorted_op_list; gIter != NULL; gIter = gIter->next) {
4290  xmlNode *rsc_op = (xmlNode *) gIter->data;
4291 
4292  counter++;
4293 
4294  if (start_index < stop_index) {
4295  crm_trace("Skipping %s: not active", ID(rsc_entry));
4296  break;
4297 
4298  } else if (counter < start_index) {
4299  crm_trace("Skipping %s: old", ID(rsc_op));
4300  continue;
4301  }
4302  op_list = g_list_append(op_list, rsc_op);
4303  }
4304 
4305  g_list_free(sorted_op_list);
4306  return op_list;
4307 }
4308 
4309 GList *
4310 find_operations(const char *rsc, const char *node, gboolean active_filter,
4311  pe_working_set_t * data_set)
4312 {
4313  GList *output = NULL;
4314  GList *intermediate = NULL;
4315 
4316  xmlNode *tmp = NULL;
4317  xmlNode *status = find_xml_node(data_set->input, XML_CIB_TAG_STATUS, TRUE);
4318 
4319  pe_node_t *this_node = NULL;
4320 
4321  xmlNode *node_state = NULL;
4322 
4323  for (node_state = pcmk__xe_first_child(status); node_state != NULL;
4324  node_state = pcmk__xe_next(node_state)) {
4325 
4326  if (pcmk__str_eq((const char *)node_state->name, XML_CIB_TAG_STATE, pcmk__str_none)) {
4327  const char *uname = crm_element_value(node_state, XML_ATTR_UNAME);
4328 
4329  if (node != NULL && !pcmk__str_eq(uname, node, pcmk__str_casei)) {
4330  continue;
4331  }
4332 
4333  this_node = pe_find_node(data_set->nodes, uname);
4334  if(this_node == NULL) {
4335  CRM_LOG_ASSERT(this_node != NULL);
4336  continue;
4337 
4338  } else if (pe__is_guest_or_remote_node(this_node)) {
4339  determine_remote_online_status(data_set, this_node);
4340 
4341  } else {
4342  determine_online_status(node_state, this_node, data_set);
4343  }
4344 
4345  if (this_node->details->online
4346  || pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
4347  /* offline nodes run no resources...
4348  * unless stonith is enabled in which case we need to
4349  * make sure rsc start events happen after the stonith
4350  */
4351  xmlNode *lrm_rsc = NULL;
4352 
4353  tmp = find_xml_node(node_state, XML_CIB_TAG_LRM, FALSE);
4354  tmp = find_xml_node(tmp, XML_LRM_TAG_RESOURCES, FALSE);
4355 
4356  for (lrm_rsc = pcmk__xe_first_child(tmp); lrm_rsc != NULL;
4357  lrm_rsc = pcmk__xe_next(lrm_rsc)) {
4358 
4359  if (pcmk__str_eq((const char *)lrm_rsc->name,
4360  XML_LRM_TAG_RESOURCE, pcmk__str_none)) {
4361 
4362  const char *rsc_id = crm_element_value(lrm_rsc, XML_ATTR_ID);
4363 
4364  if (rsc != NULL && !pcmk__str_eq(rsc_id, rsc, pcmk__str_casei)) {
4365  continue;
4366  }
4367 
4368  intermediate = extract_operations(uname, rsc_id, lrm_rsc, active_filter);
4369  output = g_list_concat(output, intermediate);
4370  }
4371  }
4372  }
4373  }
4374  }
4375 
4376  return output;
4377 }
pe_working_set_s::tags
GHashTable * tags
Definition: pe_types.h:187
services.h
Services API.
host
pcmk__cpg_host_t host
Definition: cpg.c:49
CRM_CHECK
#define CRM_CHECK(expr, failure_action)
Definition: logging.h:227
XML_RSC_OP_LAST_CHANGE
#define XML_RSC_OP_LAST_CHANGE
Definition: msg_xml.h:320
verify_pe_options
void verify_pe_options(GHashTable *options)
Definition: common.c:301
pe_find_node
pe_node_t * pe_find_node(GList *node_list, const char *uname)
Definition: status.c:443
pe_working_set_s::no_quorum_policy
enum pe_quorum_policy no_quorum_policy
Definition: pe_types.h:156
RSC_STOP
#define RSC_STOP
Definition: crm.h:202
crm.h
A dumping ground.
pe_native
Definition: pe_types.h:38
pe_order_implies_then
Definition: pe_types.h:485
pcmk__str_none
Definition: strings_internal.h:25
pe_find_node_any
pe_node_t * pe_find_node_any(GList *node_list, const char *id, const char *uname)
Definition: status.c:415
PCMK_OCF_FAILED_PROMOTED
Service failed and possibly in promoted role.
Definition: results.h:173
pe__node_list2table
GHashTable * pe__node_list2table(GList *list)
Definition: utils.c:151
crm_notice
#define crm_notice(fmt, args...)
Definition: logging.h:361
pe_resource_s::known_on
GHashTable * known_on
Definition: pe_types.h:374
CRMD_ACTION_MIGRATED
#define CRMD_ACTION_MIGRATED
Definition: crm.h:172
pe_working_set_s::failed
xmlNode * failed
Definition: pe_types.h:172
PCMK_EXEC_NOT_CONNECTED
No connection to executor.
Definition: results.h:320
pe_flag_stop_action_orphans
#define pe_flag_stop_action_orphans
Definition: pe_types.h:105
pe__is_guest_or_remote_node
bool pe__is_guest_or_remote_node(const pe_node_t *node)
Definition: remote.c:41
pe_node_shared_s::attrs
GHashTable * attrs
Definition: pe_types.h:241
pe_rsc_debug
#define pe_rsc_debug(rsc, fmt, args...)
Definition: internal.h:46
pe_node_shared_s::unseen
gboolean unseen
Definition: pe_types.h:225
XML_CONFIG_ATTR_SHUTDOWN_LOCK
#define XML_CONFIG_ATTR_SHUTDOWN_LOCK
Definition: msg_xml.h:401
pe_node_s::fixed
gboolean fixed
Definition: pe_types.h:250
data
char data[0]
Definition: cpg.c:55
INFINITY
#define INFINITY
Definition: crm.h:99
PCMK_OCF_RUNNING_PROMOTED
Service active and promoted.
Definition: results.h:172
pe__set_action_flags
#define pe__set_action_flags(action, flags_to_set)
Definition: internal.h:86
CRM_ATTR_KIND
#define CRM_ATTR_KIND
Definition: crm.h:115
XML_NODE_IS_FENCED
#define XML_NODE_IS_FENCED
Definition: msg_xml.h:283
XML_ATTR_TRANSITION_MAGIC
#define XML_ATTR_TRANSITION_MAGIC
Definition: msg_xml.h:412
pe_ticket_s::state
GHashTable * state
Definition: pe_types.h:457
pcmk__scan_min_int
int pcmk__scan_min_int(const char *text, int *result, int minimum)
Definition: strings.c:127
CRM_ATTR_IS_DC
#define CRM_ATTR_IS_DC
Definition: crm.h:117
stop_action
#define stop_action(rsc, node, optional)
Definition: internal.h:415
pe__add_param_check
void pe__add_param_check(xmlNode *rsc_op, pe_resource_t *rsc, pe_node_t *node, enum pe_check_parameters, pe_working_set_t *data_set)
Definition: remote.c:220
pe_resource_s::container
pe_resource_t * container
Definition: pe_types.h:387
name
const char * name
Definition: cib.c:24
pcmk__strcase_any_of
bool pcmk__strcase_any_of(const char *s,...) G_GNUC_NULL_TERMINATED
Definition: strings.c:928
pe_resource_s::partial_migration_source
pe_node_t * partial_migration_source
Definition: pe_types.h:372
XML_ATTR_QUORUM_PANIC
#define XML_ATTR_QUORUM_PANIC
Definition: msg_xml.h:122
pe_flag_concurrent_fencing
#define pe_flag_concurrent_fencing
Definition: pe_types.h:102
XML_ATTR_TYPE
#define XML_ATTR_TYPE
Definition: msg_xml.h:138
XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY
#define XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY
Definition: msg_xml.h:403
pe__update_recheck_time
void pe__update_recheck_time(time_t recheck, pe_working_set_t *data_set)
Definition: utils.c:703
pe_resource_s::role
enum rsc_role_e role
Definition: pe_types.h:377
XML_TAG_UTILIZATION
#define XML_TAG_UTILIZATION
Definition: msg_xml.h:214
pe__cmp_node_name
gint pe__cmp_node_name(gconstpointer a, gconstpointer b)
Definition: utils.c:180
pcmk__config_warn
#define pcmk__config_warn(fmt...)
Definition: logging_internal.h:36
node_remote
Definition: pe_types.h:74
pe_resource_s::children
GList * children
Definition: pe_types.h:384
XML_RULE_ATTR_SCORE
#define XML_RULE_ATTR_SCORE
Definition: msg_xml.h:336
XML_BOOLEAN_FALSE
#define XML_BOOLEAN_FALSE
Definition: msg_xml.h:147
pe_flag_symmetric_cluster
#define pe_flag_symmetric_cluster
Definition: pe_types.h:96
pe_ticket_s::standby
gboolean standby
Definition: pe_types.h:456
pe_rsc_stop
#define pe_rsc_stop
Definition: pe_types.h:271
pe_working_set_s::priority_fencing_delay
int priority_fencing_delay
Definition: pe_types.h:197
first_named_child
xmlNode * first_named_child(const xmlNode *parent, const char *name)
Definition: xml.c:2930
pe_flag_remove_after_stop
#define pe_flag_remove_after_stop
Definition: pe_types.h:111
pe_resource_s::next_role
enum rsc_role_e next_role
Definition: pe_types.h:378
pe_flag_maintenance_mode
#define pe_flag_maintenance_mode
Definition: pe_types.h:97
pe_action_s
Definition: pe_types.h:402
pe__is_newer_op
int pe__is_newer_op(const xmlNode *xml_a, const xmlNode *xml_b, bool same_node_default)
Definition: pe_actions.c:1494
char2score
int char2score(const char *score)
Get the integer value of a score string.
Definition: scores.c:36
pcmk__config_err
#define pcmk__config_err(fmt...)
Definition: logging_internal.h:24
find_xml_node
xmlNode * find_xml_node(const xmlNode *root, const char *search_path, gboolean must_find)
Definition: xml.c:470
pe_proc_warn
#define pe_proc_warn(fmt...)
Definition: internal.h:60
pe_node_shared_s::remote_rsc
pe_resource_t * remote_rsc
Definition: pe_types.h:237
XML_TAG_TRANSIENT_NODEATTRS
#define XML_TAG_TRANSIENT_NODEATTRS
Definition: msg_xml.h:417
CRMD_ACTION_NOTIFY
#define CRMD_ACTION_NOTIFY
Definition: crm.h:185
unpack_resources
gboolean unpack_resources(xmlNode *xml_resources, pe_working_set_t *data_set)
Definition: unpack.c:762
pe_resource_s::meta
GHashTable * meta
Definition: pe_types.h:380
pe_rsc_unique
#define pe_rsc_unique
Definition: pe_types.h:262
pe_rule_eval_data
Definition: common.h:193
PCMK_OCF_NOT_RUNNING
Service safely stopped.
Definition: results.h:171
pe_resource_s::fns
resource_object_functions_t * fns
Definition: pe_types.h:340
XML_CIB_TAG_TAG
#define XML_CIB_TAG_TAG
Definition: msg_xml.h:444
XML_LRM_TAG_RESOURCE
#define XML_LRM_TAG_RESOURCE
Definition: msg_xml.h:267
crm_xml_add
const char * crm_xml_add(xmlNode *node, const char *name, const char *value)
Create an XML attribute with specified name and value.
Definition: nvpair.c:323
pe__copy_node
pe_node_t * pe__copy_node(const pe_node_t *this_node)
Definition: utils.c:89
no_quorum_freeze
Definition: pe_types.h:64
CRMD_ACTION_PROMOTE
#define CRMD_ACTION_PROMOTE
Definition: crm.h:180
pe__bundle_needs_remote_name
bool pe__bundle_needs_remote_name(pe_resource_t *rsc, pe_working_set_t *data_set)
Definition: bundle.c:690
unpack_tags
gboolean unpack_tags(xmlNode *xml_tags, pe_working_set_t *data_set)
Definition: unpack.c:828
calculate_active_ops
void calculate_active_ops(GList *sorted_op_list, int *start_index, int *stop_index)
Definition: unpack.c:2290
node_member
Definition: pe_types.h:73
pe_node_shared_s::pending
gboolean pending
Definition: pe_types.h:223
XML_NVPAIR_ATTR_NAME
#define XML_NVPAIR_ATTR_NAME
Definition: msg_xml.h:391
XML_NODE_IS_MAINTENANCE
#define XML_NODE_IS_MAINTENANCE
Definition: msg_xml.h:284
pe_tag_s::id
char * id
Definition: pe_types.h:461
sort_op_by_callid
gint sort_op_by_callid(gconstpointer a, gconstpointer b)
Definition: pe_actions.c:1627
XML_NODE_EXPECTED
#define XML_NODE_EXPECTED
Definition: msg_xml.h:279
XML_CIB_TAG_RSC_TEMPLATE
#define XML_CIB_TAG_RSC_TEMPLATE
Definition: msg_xml.h:222
PCMK_EXEC_NO_SECRETS
Necessary CIB secrets are unavailable.
Definition: results.h:323
pe_find_resource
pe_resource_t * pe_find_resource(GList *rsc_list, const char *id_rh)
Definition: status.c:391
get_effective_time
time_t get_effective_time(pe_working_set_t *data_set)
Definition: utils.c:421
CRM_LOG_ASSERT
#define CRM_LOG_ASSERT(expr)
Definition: logging.h:211
pe_pref
const char * pe_pref(GHashTable *options, const char *name)
Definition: common.c:307
resource_location
void resource_location(pe_resource_t *rsc, pe_node_t *node, int score, const char *tag, pe_working_set_t *data_set)
Definition: utils.c:385
PCMK_OCF_DEGRADED_PROMOTED
Service promoted but more likely to fail soon.
Definition: results.h:175
unpack_remote_nodes
gboolean unpack_remote_nodes(xmlNode *xml_resources, pe_working_set_t *data_set)
Definition: unpack.c:625
type
enum crm_ais_msg_types type
Definition: cpg.c:48
copy_in_properties
void copy_in_properties(xmlNode *target, xmlNode *src)
Definition: xml.c:545
CRMD_JOINSTATE_NACK
#define CRMD_JOINSTATE_NACK
Definition: crm.h:164
pcmk__str_null_matches
Definition: strings_internal.h:27
XML_CIB_TAG_LRM
#define XML_CIB_TAG_LRM
Definition: msg_xml.h:265
CRM_ATTR_CLUSTER_NAME
#define CRM_ATTR_CLUSTER_NAME
Definition: crm.h:118
pe_resource_s::partial_migration_target
pe_node_t * partial_migration_target
Definition: pe_types.h:371
pe_working_set_s::tickets
GHashTable * tickets
Definition: pe_types.h:159
PCMK_EXEC_TIMEOUT
Action did not complete in time.
Definition: results.h:314
pe_node_shared_s::remote_was_fenced
gboolean remote_was_fenced
Definition: pe_types.h:232
crm_element_value_int
int crm_element_value_int(const xmlNode *data, const char *name, int *dest)
Retrieve the integer value of an XML attribute.
Definition: nvpair.c:553
PCMK_EXEC_ERROR_HARD
Execution failed, do not retry on node.
Definition: results.h:317
pcmk__ends_with
bool pcmk__ends_with(const char *s, const char *match)
Definition: strings.c:536
pe_flag_have_quorum
#define pe_flag_have_quorum
Definition: pe_types.h:95
pe_working_set_s
Definition: pe_types.h:143
unpack_config
gboolean unpack_config(xmlNode *config, pe_working_set_t *data_set)
Definition: unpack.c:196
pe_resource_s::pending_task
char * pending_task
Definition: pe_types.h:353
get_xpath_object
xmlNode * get_xpath_object(const char *xpath, xmlNode *xml_obj, int error_level)
Definition: xpath.c:214
pe_proc_err
#define pe_proc_err(fmt...)
Definition: internal.h:59
pe_node_shared_s::remote_requires_reset
gboolean remote_requires_reset
Definition: pe_types.h:231
action_fail_response
action_fail_response
Definition: common.h:33
pe_node_shared_s::data_set
pe_working_set_t * data_set
Cluster that this node is part of.
Definition: pe_types.h:245
PCMK_EXEC_CANCELLED
Action was cancelled.
Definition: results.h:313
strndup
char * strndup(const char *str, size_t len)
pe_working_set_s::dc_uuid
char * dc_uuid
Definition: pe_types.h:148
PCMK_EXEC_NO_FENCE_DEVICE
No fence device is configured for target.
Definition: results.h:322
action
const char * action
Definition: pcmk_fence.c:30
pe__set_resource_flags
#define pe__set_resource_flags(resource, flags_to_set)
Definition: internal.h:74
pe_working_set_s::resources
GList * resources
Definition: pe_types.h:165
pe_working_set_s::stonith_timeout
int stonith_timeout
Definition: pe_types.h:155
XML_CIB_TAG_PROPSET
#define XML_CIB_TAG_PROPSET
Definition: msg_xml.h:208
pe_resource_s::lock_node
pe_node_t * lock_node
Definition: pe_types.h:391
pe_working_set_s::nodes
GList * nodes
Definition: pe_types.h:164
pcmk__effective_rc
int pcmk__effective_rc(int rc)
Definition: agents.c:71
pe_flag_stop_everything
#define pe_flag_stop_everything
Definition: pe_types.h:106
XML_LRM_ATTR_RSCID
#define XML_LRM_ATTR_RSCID
Definition: msg_xml.h:309
pe_node_shared_s::remote_maintenance
gboolean remote_maintenance
Definition: pe_types.h:233
CRMD_ACTION_START
#define CRMD_ACTION_START
Definition: crm.h:174
demote_action
#define demote_action(rsc, node, optional)
Definition: internal.h:441
pe_node_shared_s::is_dc
gboolean is_dc
Definition: pe_types.h:228
XML_LRM_ATTR_TASK_KEY
#define XML_LRM_ATTR_TASK_KEY
Definition: msg_xml.h:301
pe__is_remote_node
bool pe__is_remote_node(const pe_node_t *node)
Definition: remote.c:25
XML_TAG_ATTR_SETS
#define XML_TAG_ATTR_SETS
Definition: msg_xml.h:209
XML_LRM_ATTR_TASK
#define XML_LRM_ATTR_TASK
Definition: msg_xml.h:300
role2text
const char * role2text(enum rsc_role_e role)
Definition: common.c:454
pe_create_remote_xml
xmlNode * pe_create_remote_xml(xmlNode *parent, const char *uname, const char *container_id, const char *migrateable, const char *is_managed, const char *start_timeout, const char *server, const char *port)
Definition: remote.c:156
resource_object_functions_s::location
pe_node_t *(* location)(const pe_resource_t *, GList **, int)
Definition: pe_types.h:55
CRMD_ACTION_STOP
#define CRMD_ACTION_STOP
Definition: crm.h:177
pe__find_bundle_replica
pe_resource_t * pe__find_bundle_replica(const pe_resource_t *bundle, const pe_node_t *node)
Definition: bundle.c:1140
pe_node_s::weight
int weight
Definition: pe_types.h:249
pcmk__scan_ll
int pcmk__scan_ll(const char *text, long long *result, long long default_value)
Definition: strings.c:97
pe_flag_have_remote_nodes
#define pe_flag_have_remote_nodes
Definition: pe_types.h:118
pe_resource_s::dangling_migrations
GList * dangling_migrations
Definition: pe_types.h:385
CRMD_JOINSTATE_DOWN
#define CRMD_JOINSTATE_DOWN
Definition: crm.h:161
PCMK_EXEC_MAX
Maximum value for this enum.
Definition: results.h:326
crm_warn
#define crm_warn(fmt, args...)
Definition: logging.h:360
pe_resource_s::remote_reconnect_ms
guint remote_reconnect_ms
Definition: pe_types.h:352
XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT
#define XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT
Definition: msg_xml.h:402
CRMD_ACTION_DEMOTE
#define CRMD_ACTION_DEMOTE
Definition: crm.h:182
clone_zero
char * clone_zero(const char *last_rsc_id)
Definition: unpack.c:1669
pe_rsc_allow_migrate
#define pe_rsc_allow_migrate
Definition: pe_types.h:287
pe_rsc_orphan_container_filler
#define pe_rsc_orphan_container_filler
Definition: pe_types.h:259
crm_element_value_ms
int crm_element_value_ms(const xmlNode *data, const char *name, guint *dest)
Retrieve the millisecond value of an XML attribute.
Definition: nvpair.c:610
pe_resource_s
Definition: pe_types.h:328
pe_rsc_failed
#define pe_rsc_failed
Definition: pe_types.h:276
crm_debug
#define crm_debug(fmt, args...)
Definition: logging.h:364
pe__target_rc_from_xml
int pe__target_rc_from_xml(xmlNode *xml_op)
Definition: unpack.c:3786
uber_parent
pe_resource_t * uber_parent(pe_resource_t *rsc)
Definition: complex.c:912
XML_RSC_ATTR_CONTAINER
#define XML_RSC_ATTR_CONTAINER
Definition: msg_xml.h:244
util.h
Utility functions.
PCMK_EXEC_UNKNOWN
Used only to initialize variables.
Definition: results.h:310
XML_ATTR_ID
#define XML_ATTR_ID
Definition: msg_xml.h:134
crm_element_value
const char * crm_element_value(const xmlNode *data, const char *name)
Retrieve the value of an XML attribute.
Definition: nvpair.c:517
pe_node_attribute_raw
const char * pe_node_attribute_raw(const pe_node_t *node, const char *name)
Definition: common.c:562
XML_CIB_TAG_RESOURCE
#define XML_CIB_TAG_RESOURCE
Definition: msg_xml.h:217
pe_rsc_is_container
#define pe_rsc_is_container
Definition: pe_types.h:291
XML_BOOLEAN_TRUE
#define XML_BOOLEAN_TRUE
Definition: msg_xml.h:146
XML_CIB_TAG_STATE
#define XML_CIB_TAG_STATE
Definition: msg_xml.h:204
pe_base_name_end
const char * pe_base_name_end(const char *id)
Definition: unpack.c:1607
pe__is_guest_node
bool pe__is_guest_node(const pe_node_t *node)
Definition: remote.c:33
PCMK_OCF_INVALID_PARAM
Parameter invalid (in local context)
Definition: results.h:166
pe_node_shared_s::unpacked
gboolean unpacked
Definition: pe_types.h:234
pe_resource_s::failure_timeout
int failure_timeout
Definition: pe_types.h:350
crm_internal.h
pe_check_last_failure
Definition: pe_types.h:206
pe_find_clone
match only clone instances
Definition: pe_types.h:87
pe_rsc_start_pending
#define pe_rsc_start_pending
Definition: pe_types.h:278
PCMK_OCF_NOT_CONFIGURED
Parameter invalid (inherently)
Definition: results.h:170
pe_warn_once
#define pe_warn_once(pe_wo_bit, fmt...)
Definition: internal.h:172
ticket_new
pe_ticket_t * ticket_new(const char *ticket_id, pe_working_set_t *data_set)
Definition: utils.c:533
CRM_ATTR_UNAME
#define CRM_ATTR_UNAME
Definition: crm.h:113
XML_NODE_IS_PEER
#define XML_NODE_IS_PEER
Definition: msg_xml.h:281
crm_trace
#define crm_trace(fmt, args...)
Definition: logging.h:365
CRMD_JOINSTATE_MEMBER
#define CRMD_JOINSTATE_MEMBER
Definition: crm.h:163
do_crm_log
#define do_crm_log(level, fmt, args...)
Log a message.
Definition: logging.h:168
op_digest_cache_s::rc
enum rsc_digest_cmp_val rc
Definition: internal.h:517
pcmk__g_strcat
void pcmk__g_strcat(GString *buffer,...) G_GNUC_NULL_TERMINATED
Definition: strings.c:1214
xml_contains_remote_node
bool xml_contains_remote_node(xmlNode *xml)
Definition: remote.c:84
crm_strdup_printf
char * crm_strdup_printf(char const *format,...) G_GNUC_PRINTF(1
pcmk__str_casei
Definition: strings_internal.h:26
pcmk_is_set
#define pcmk_is_set(g, f)
Convenience alias for pcmk_all_flags_set(), to check single flag.
Definition: util.h:121
PCMK_OCF_INSUFFICIENT_PRIV
Insufficient privileges.
Definition: results.h:168
add_node_copy
xmlNode * add_node_copy(xmlNode *new_parent, xmlNode *xml_node)
Definition: xml.c:727
pe_flag_startup_fencing
#define pe_flag_startup_fencing
Definition: pe_types.h:113
pe_working_set_s::stonith_action
const char * stonith_action
Definition: pe_types.h:150
pe_node_s::details
struct pe_node_shared_s * details
Definition: pe_types.h:252
get_target_role
gboolean get_target_role(pe_resource_t *rsc, enum rsc_role_e *role)
Definition: utils.c:436
crm_log_xml_debug
#define crm_log_xml_debug(xml, text)
Definition: logging.h:372
pe__shutdown_requested
bool pe__shutdown_requested(pe_node_t *node)
Definition: utils.c:688
pe_rsc_needs_fencing
#define pe_rsc_needs_fencing
Definition: pe_types.h:294
order_actions
gboolean order_actions(pe_action_t *lh_action, pe_action_t *rh_action, enum pe_ordering order)
Definition: utils.c:474
pe_resource_s::flags
unsigned long long flags
Definition: pe_types.h:355
pe_node_shared_s::uname
const char * uname
Definition: pe_types.h:216
pe_rsc_promotable
#define pe_rsc_promotable
Definition: pe_types.h:264
data_set
pe_working_set_t * data_set
Definition: native_find_rsc_test.c:23
XML_TAG_META_SETS
#define XML_TAG_META_SETS
Definition: msg_xml.h:210
pcmk__str_update
void pcmk__str_update(char **str, const char *value)
Definition: strings.c:1190
xml.h
Wrappers for and extensions to libxml2.
pe_working_set_s::config_hash
GHashTable * config_hash
Definition: pe_types.h:158
XML_ATTR_UNAME
#define XML_ATTR_UNAME
Definition: msg_xml.h:157
pcmk_xe_mask_probe_failure
bool pcmk_xe_mask_probe_failure(xmlNode *xml_op)
Definition: operations.c:562
pe_resource_s::clone_name
char * clone_name
Definition: pe_types.h:330
add_tag_ref
gboolean add_tag_ref(GHashTable *tags, const char *tag_name, const char *obj_ref)
Definition: utils.c:639
XML_RSC_ATTR_MANAGED
#define XML_RSC_ATTR_MANAGED
Definition: msg_xml.h:235
create_xml_node
xmlNode * create_xml_node(xmlNode *parent, const char *name)
Definition: xml.c:749
pe_resource_s::lock_time
time_t lock_time
Definition: pe_types.h:392
xml_internal.h
PCMK_EXEC_DONE
Action completed, result is known.
Definition: results.h:312
pe_ticket_s
Definition: pe_types.h:452
pe_flag_stonith_enabled
#define pe_flag_stonith_enabled
Definition: pe_types.h:99
unpack_status
gboolean unpack_status(xmlNode *status, pe_working_set_t *data_set)
Definition: unpack.c:1253
crm_element_value_epoch
int crm_element_value_epoch(const xmlNode *xml, const char *name, time_t *dest)
Retrieve the seconds-since-epoch value of an XML attribute.
Definition: nvpair.c:638
pe_ticket_s::last_granted
time_t last_granted
Definition: pe_types.h:455
pe__create_clone_child
pe_resource_t * pe__create_clone_child(pe_resource_t *rsc, pe_working_set_t *data_set)
Definition: clone.c:185
XML_RSC_ATTR_TARGET_ROLE
#define XML_RSC_ATTR_TARGET_ROLE
Definition: msg_xml.h:236
XML_LRM_ATTR_MIGRATE_TARGET
#define XML_LRM_ATTR_MIGRATE_TARGET
Definition: msg_xml.h:326
PCMK_EXEC_ERROR_FATAL
Execution failed, do not retry anywhere.
Definition: results.h:318
CIB_OPTIONS_FIRST
#define CIB_OPTIONS_FIRST
Definition: msg_xml.h:88
pe_node_shared_s::standby
gboolean standby
Definition: pe_types.h:221
XML_RSC_ATTR_REMOTE_NODE
#define XML_RSC_ATTR_REMOTE_NODE
Definition: msg_xml.h:247
XML_LRM_ATTR_EXIT_REASON
#define XML_LRM_ATTR_EXIT_REASON
Definition: msg_xml.h:318
XML_LRM_ATTR_RESTART_DIGEST
#define XML_LRM_ATTR_RESTART_DIGEST
Definition: msg_xml.h:316
pe__free_digests
void pe__free_digests(gpointer ptr)
Definition: pe_digest.c:34
no_quorum_ignore
Definition: pe_types.h:66
pe_node_shared_s::expected_up
gboolean expected_up
Definition: pe_types.h:227
PCMK_OCF_NOT_INSTALLED
Dependencies not available locally.
Definition: results.h:169
no_quorum_demote
Definition: pe_types.h:68
pe_node_s
Definition: pe_types.h:248
pe_resource_s::variant
enum pe_obj_types variant
Definition: pe_types.h:338
pcmk__str_any_of
bool pcmk__str_any_of(const char *s,...) G_GNUC_NULL_TERMINATED
Definition: strings.c:952
pe_working_set_s::input
xmlNode * input
Definition: pe_types.h:144
pe_ticket_s::granted
gboolean granted
Definition: pe_types.h:454
XML_CIB_TAG_NODE
#define XML_CIB_TAG_NODE
Definition: msg_xml.h:205
pe_working_set_s::placement_strategy
const char * placement_strategy
Definition: pe_types.h:151
id
uint32_t id
Definition: cpg.c:45
pe_node_shared_s::id
const char * id
Definition: pe_types.h:215
pcmk__op_key
char * pcmk__op_key(const char *rsc_id, const char *op_type, guint interval_ms)
Generate an operation key (RESOURCE_ACTION_INTERVAL)
Definition: operations.c:45
pe_ticket_s::id
char * id
Definition: pe_types.h:453
pe_working_set_s::shutdown_lock
guint shutdown_lock
Definition: pe_types.h:196
PCMK_OCF_UNKNOWN_ERROR
Unspecified error.
Definition: results.h:165
no_quorum_stop
Definition: pe_types.h:65
pe_tag_s::refs
GList * refs
Definition: pe_types.h:462
target
const char * target
Definition: pcmk_fence.c:29
pe_resource_s::fillers
GList * fillers
Definition: pe_types.h:388
pe_can_fence
bool pe_can_fence(pe_working_set_t *data_set, pe_node_t *node)
Definition: utils.c:36
pe_node_shared_s::running_rsc
GList * running_rsc
Definition: pe_types.h:238
msg_xml.h
XML_ATTR_TRANSITION_KEY
#define XML_ATTR_TRANSITION_KEY
Definition: msg_xml.h:413
pe_node_shared_s::rsc_discovery_enabled
gboolean rsc_discovery_enabled
Definition: pe_types.h:230
CRM_XS
#define CRM_XS
Definition: logging.h:55
pe_working_set_s::dc_node
pe_node_t * dc_node
Definition: pe_types.h:149
pe_fence_node
void pe_fence_node(pe_working_set_t *data_set, pe_node_t *node, const char *reason, bool priority_delay)
Schedule a fence action for a node.
Definition: unpack.c:95
pe_create_node
pe_node_t * pe_create_node(const char *id, const char *uname, const char *type, const char *score, pe_working_set_t *data_set)
Definition: unpack.c:381
PCMK_OCF_UNIMPLEMENT_FEATURE
Requested action not implemented.
Definition: results.h:167
pe_node_shared_s
Definition: pe_types.h:214
pe_working_set_s::localhost
const char * localhost
Definition: pe_types.h:186
pe_flag_quick_location
#define pe_flag_quick_location
Definition: pe_types.h:120
PCMK_OCF_DEGRADED
Service active but more likely to fail soon.
Definition: results.h:174
pcmk_xe_is_probe
bool pcmk_xe_is_probe(xmlNode *xml_op)
Definition: operations.c:551
pe_resource_s::is_remote_node
gboolean is_remote_node
Definition: pe_types.h:358
pe_resource_s::pending_node
pe_node_t * pending_node
Definition: pe_types.h:390
fail2text
const char * fail2text(enum action_fail_response fail)
Definition: common.c:313
PCMK_EXEC_NOT_SUPPORTED
Agent does not implement requested action.
Definition: results.h:315
pcmk__strkey_table
GHashTable * pcmk__strkey_table(GDestroyNotify key_destroy_func, GDestroyNotify value_destroy_func)
Definition: strings.c:611
result
pcmk__action_result_t result
Definition: pcmk_fence.c:35
pe__cmp_rsc_priority
G_GNUC_INTERNAL gint pe__cmp_rsc_priority(gconstpointer a, gconstpointer b)
Definition: utils.c:325
pe__unpack_node_health_scores
void pe__unpack_node_health_scores(pe_working_set_t *data_set)
Definition: pe_health.c:23
XML_LRM_TAG_RESOURCES
#define XML_LRM_TAG_RESOURCES
Definition: msg_xml.h:266
crm_err
#define crm_err(fmt, args...)
Definition: logging.h:359
XML_CIB_TAG_TICKET_STATE
#define XML_CIB_TAG_TICKET_STATE
Definition: msg_xml.h:441
set_config_flag
#define set_config_flag(data_set, option, flag)
Definition: unpack.c:34
CRM_ASSERT
#define CRM_ASSERT(expr)
Definition: results.h:42
xpath_search
xmlXPathObjectPtr xpath_search(xmlNode *xml_top, const char *path)
Definition: xpath.c:139
PCMK_OCF_OK
Success.
Definition: results.h:164
pe__clear_failcount
pe_action_t * pe__clear_failcount(pe_resource_t *rsc, pe_node_t *node, const char *reason, pe_working_set_t *data_set)
Schedule a controller operation to clear a fail count.
Definition: failcounts.c:360
pe_rule_eval_data::node_hash
GHashTable * node_hash
Definition: common.h:194
crm_parse_interval_spec
char guint crm_parse_interval_spec(const char *input)
Parse milliseconds from a Pacemaker interval specification.
Definition: utils.c:271
decode_transition_key
gboolean decode_transition_key(const char *key, char **uuid, int *transition_id, int *action_id, int *target_rc)
Parse a transition key into its constituent parts.
Definition: operations.c:318
rsc_action_digest_cmp
op_digest_cache_t * rsc_action_digest_cmp(pe_resource_t *rsc, xmlNode *xml_op, pe_node_t *node, pe_working_set_t *data_set)
Definition: pe_digest.c:380
pe__set_next_role
void pe__set_next_role(pe_resource_t *rsc, enum rsc_role_e role, const char *why)
Definition: complex.c:1120
XML_ATTR_HAVE_WATCHDOG
#define XML_ATTR_HAVE_WATCHDOG
Definition: msg_xml.h:124
XML_NODE_ATTR_RSC_DISCOVERY
#define XML_NODE_ATTR_RSC_DISCOVERY
Definition: msg_xml.h:394
unpack_nodes
gboolean unpack_nodes(xmlNode *xml_nodes, pe_working_set_t *data_set)
Definition: unpack.c:538
pe_tag_s
Definition: pe_types.h:460
CRMD_ACTION_METADATA
#define CRMD_ACTION_METADATA
Definition: crm.h:189
XML_LRM_ATTR_INTERVAL_MS
#define XML_LRM_ATTR_INTERVAL_MS
Definition: msg_xml.h:298
XML_LRM_ATTR_CALLID
#define XML_LRM_ATTR_CALLID
Definition: msg_xml.h:312
CRMD_ACTION_MIGRATE
#define CRMD_ACTION_MIGRATE
Definition: crm.h:171
XML_NVPAIR_ATTR_VALUE
#define XML_NVPAIR_ATTR_VALUE
Definition: msg_xml.h:392
pe_node_shared_s::utilization
GHashTable * utilization
Definition: pe_types.h:242
native_add_running
void native_add_running(pe_resource_t *rsc, pe_node_t *node, pe_working_set_t *data_set, gboolean failed)
Definition: native.c:89
pe_node_shared_s::shutdown
gboolean shutdown
Definition: pe_types.h:226
uname
char uname[MAX_NAME]
Definition: cpg.c:50
pe__clear_resource_flags
#define pe__clear_resource_flags(resource, flags_to_clear)
Definition: internal.h:80
XML_LRM_ATTR_OPSTATUS
#define XML_LRM_ATTR_OPSTATUS
Definition: msg_xml.h:310
no_quorum_suicide
Definition: pe_types.h:67
pe__set_working_set_flags
#define pe__set_working_set_flags(working_set, flags_to_set)
Definition: internal.h:62
CRMD_JOINSTATE_PENDING
#define CRMD_JOINSTATE_PENDING
Definition: crm.h:162
op_digest_cache_s
Definition: internal.h:516
rsc_role_e
rsc_role_e
Possible roles that a resource can be in.
Definition: common.h:92
pe_resource_s::running_on
GList * running_on
Definition: pe_types.h:373
pe_rsc_block
#define pe_rsc_block
Definition: pe_types.h:258
pe_status_private.h
CRM_TRACE_INIT_DATA
CRM_TRACE_INIT_DATA(pe_status)
PCMK_EXEC_NOT_INSTALLED
Agent or dependency not available locally.
Definition: results.h:319
pe_node_shared_s::maintenance
gboolean maintenance
Definition: pe_types.h:229
XML_LRM_ATTR_RC
#define XML_LRM_ATTR_RC
Definition: msg_xml.h:311
pcmk_ok
#define pcmk_ok
Definition: results.h:68
pe_node_shared_s::digest_cache
GHashTable * digest_cache
cache of calculated resource digests
Definition: pe_types.h:243
pe__is_universal_clone
bool pe__is_universal_clone(pe_resource_t *rsc, pe_working_set_t *data_set)
Definition: clone.c:1158
pe_rsc_failure_ignored
#define pe_rsc_failure_ignored
Definition: pe_types.h:289
XML_NODE_JOIN_STATE
#define XML_NODE_JOIN_STATE
Definition: msg_xml.h:278
PCMK_EXEC_PENDING
Action is in progress.
Definition: results.h:311
destroy_ticket
void destroy_ticket(gpointer data)
Definition: utils.c:521
XML_CIB_TAG_STATUS
#define XML_CIB_TAG_STATUS
Definition: msg_xml.h:185
XML_CIB_TAG_OBJ_REF
#define XML_CIB_TAG_OBJ_REF
Definition: msg_xml.h:445
XML_NODE_IN_CLUSTER
#define XML_NODE_IN_CLUSTER
Definition: msg_xml.h:280
crm_is_true
gboolean crm_is_true(const char *s)
Definition: strings.c:416
pe_flag_have_stonith_resource
#define pe_flag_have_stonith_resource
Definition: pe_types.h:100
CRM_ATTR_SITE_NAME
#define CRM_ATTR_SITE_NAME
Definition: crm.h:119
XML_CIB_TAG_GROUP
#define XML_CIB_TAG_GROUP
Definition: msg_xml.h:218
pe_flag_enable_unfencing
#define pe_flag_enable_unfencing
Definition: pe_types.h:101
find_operations
GList * find_operations(const char *rsc, const char *node, gboolean active_filter, pe_working_set_t *data_set)
Definition: unpack.c:4310
XML_LRM_TAG_RSC_OP
#define XML_LRM_TAG_RSC_OP
Definition: msg_xml.h:268
pe_rsc_trace
#define pe_rsc_trace(rsc, fmt, args...)
Definition: internal.h:47
pe_wo
uint32_t pe_wo
Definition: unpack.c:67
ID
#define ID(x)
Definition: msg_xml.h:468
pe_working_set_s::flags
unsigned long long flags
Definition: pe_types.h:153
pe_err
#define pe_err(fmt...)
Definition: internal.h:49
parent
const char * parent
Definition: cib.c:25
pe_action_optional
Definition: pe_types.h:301
pcmk__epoch2str
const char * pcmk__epoch2str(const time_t *when)
Definition: iso8601.c:1730
PCMK_EXEC_INVALID
Action cannot be attempted (e.g. shutdown)
Definition: results.h:321
ONLINESTATUS
#define ONLINESTATUS
Definition: util.h:39
pe__resource_actions
GList * pe__resource_actions(const pe_resource_t *rsc, const pe_node_t *node, const char *task, bool require_node)
Find all actions of given type for a resource.
Definition: pe_actions.c:1398
pe_node_shared_s::standby_onfail
gboolean standby_onfail
Definition: pe_types.h:222
pe__clear_resource_history
pe_action_t * pe__clear_resource_history(pe_resource_t *rsc, pe_node_t *node, pe_working_set_t *data_set)
Definition: pe_actions.c:1473
XML_LRM_ATTR_MIGRATE_SOURCE
#define XML_LRM_ATTR_MIGRATE_SOURCE
Definition: msg_xml.h:325
resource_object_functions_s::find_rsc
pe_resource_t *(* find_rsc)(pe_resource_t *parent, const char *search, const pe_node_t *node, int flags)
Definition: pe_types.h:46
custom_action
pe_action_t * custom_action(pe_resource_t *rsc, char *key, const char *task, const pe_node_t *on_node, gboolean optional, gboolean foo, pe_working_set_t *data_set)
Create or update an action object.
Definition: pe_actions.c:940
freeXpathObject
void freeXpathObject(xmlXPathObjectPtr xpathObj)
Definition: xpath.c:39
CRM_ATTR_ID
#define CRM_ATTR_ID
Definition: crm.h:114
pe_node_shared_s::unclean
gboolean unclean
Definition: pe_types.h:224
pe__clear_working_set_flags
#define pe__clear_working_set_flags(working_set, flags_to_clear)
Definition: internal.h:68
XPATH_ENABLE_UNFENCING
#define XPATH_ENABLE_UNFENCING
Definition: unpack.c:175
pe_flag_start_failure_fatal
#define pe_flag_start_failure_fatal
Definition: pe_types.h:108
pe_node_shared_s::type
enum node_type type
Definition: pe_types.h:217
XML_CIB_TAG_TICKETS
#define XML_CIB_TAG_TICKETS
Definition: msg_xml.h:440
pe_working_set_s::now
crm_time_t * now
Definition: pe_types.h:145
PCMK_EXEC_ERROR
Execution failed, may be retried.
Definition: results.h:316
node_ping
Definition: pe_types.h:72
crm_info
#define crm_info(fmt, args...)
Definition: logging.h:362
pe_rsc_managed
#define pe_rsc_managed
Definition: pe_types.h:257
pe_rsc_orphan
#define pe_rsc_orphan
Definition: pe_types.h:256
pe_working_set_s::template_rsc_sets
GHashTable * template_rsc_sets
Definition: pe_types.h:185
pe_node_shared_s::online
gboolean online
Definition: pe_types.h:220
clone_strip
char * clone_strip(const char *last_rsc_id)
Definition: unpack.c:1647
pe_get_failcount
int pe_get_failcount(pe_node_t *node, pe_resource_t *rsc, time_t *last_failure, uint32_t flags, xmlNode *xml_op, pe_working_set_t *data_set)
Definition: failcounts.c:251
pe_working_set_s::stop_needed
GList * stop_needed
Definition: pe_types.h:193
pe_resource_s::parent
pe_resource_t * parent
Definition: pe_types.h:336
pe_fence_op
pe_action_t * pe_fence_op(pe_node_t *node, const char *op, bool optional, const char *reason, bool priority_delay, pe_working_set_t *data_set)
Definition: pe_actions.c:1081
pe_free_action
void pe_free_action(pe_action_t *action)
Definition: pe_actions.c:1193
pe_flag_shutdown_lock
#define pe_flag_shutdown_lock
Definition: pe_types.h:114
pe__unpack_resource
int pe__unpack_resource(xmlNode *xml_obj, pe_resource_t **rsc, pe_resource_t *parent, pe_working_set_t *data_set)
Definition: complex.c:562
pe_rsc_info
#define pe_rsc_info(rsc, fmt, args...)
Definition: internal.h:45
pe_resource_s::id
char * id
Definition: pe_types.h:329
pe_resource_s::allowed_nodes
GHashTable * allowed_nodes
Definition: pe_types.h:375
pe_flag_startup_probes
#define pe_flag_startup_probes
Definition: pe_types.h:116
CRMD_ACTION_STATUS
#define CRMD_ACTION_STATUS
Definition: crm.h:188
crm_next_same_xml
xmlNode * crm_next_same_xml(const xmlNode *sibling)
Get next instance of same XML tag.
Definition: xml.c:2956
pe__unpack_dataset_nvpairs
void pe__unpack_dataset_nvpairs(const xmlNode *xml_obj, const char *set_name, pe_rule_eval_data_t *rule_data, GHashTable *hash, const char *always_first, gboolean overwrite, pe_working_set_t *data_set)
Definition: utils.c:717
pe_flag_stop_rsc_orphans
#define pe_flag_stop_rsc_orphans
Definition: pe_types.h:104
Generated on Thu Dec 8 2022 10:06:23 for pacemaker by   doxygen 1.8.14