pacemaker  2.1.2-ada5c3b36
Scalable High-Availability cluster resource manager
unpack.c
Go to the documentation of this file.
1 /*
2  * Copyright 2004-2021 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>
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 &&
80  return TRUE;
81  }
82 
83  return FALSE;
84 }
85 
86 
95 void
97  const char *reason, bool priority_delay)
98 {
99  CRM_CHECK(node, return);
100 
101  /* A guest node is fenced by marking its container as failed */
102  if (pe__is_guest_node(node)) {
103  pe_resource_t *rsc = node->details->remote_rsc->container;
104 
105  if (!pcmk_is_set(rsc->flags, pe_rsc_failed)) {
106  if (!pcmk_is_set(rsc->flags, pe_rsc_managed)) {
107  crm_notice("Not fencing guest node %s "
108  "(otherwise would because %s): "
109  "its guest resource %s is unmanaged",
110  node->details->uname, reason, rsc->id);
111  } else {
112  crm_warn("Guest node %s will be fenced "
113  "(by recovering its guest resource %s): %s",
114  node->details->uname, rsc->id, reason);
115 
116  /* We don't mark the node as unclean because that would prevent the
117  * node from running resources. We want to allow it to run resources
118  * in this transition if the recovery succeeds.
119  */
120  node->details->remote_requires_reset = TRUE;
122  }
123  }
124 
125  } else if (is_dangling_guest_node(node)) {
126  crm_info("Cleaning up dangling connection for guest node %s: "
127  "fencing was already done because %s, "
128  "and guest resource no longer exists",
129  node->details->uname, reason);
132 
133  } else if (pe__is_remote_node(node)) {
134  pe_resource_t *rsc = node->details->remote_rsc;
135 
136  if ((rsc != NULL) && !pcmk_is_set(rsc->flags, pe_rsc_managed)) {
137  crm_notice("Not fencing remote node %s "
138  "(otherwise would because %s): connection is unmanaged",
139  node->details->uname, reason);
140  } else if(node->details->remote_requires_reset == FALSE) {
141  node->details->remote_requires_reset = TRUE;
142  crm_warn("Remote node %s %s: %s",
143  node->details->uname,
144  pe_can_fence(data_set, node)? "will be fenced" : "is unclean",
145  reason);
146  }
147  node->details->unclean = TRUE;
148  // No need to apply `priority-fencing-delay` for remote nodes
149  pe_fence_op(node, NULL, TRUE, reason, FALSE, data_set);
150 
151  } else if (node->details->unclean) {
152  crm_trace("Cluster node %s %s because %s",
153  node->details->uname,
154  pe_can_fence(data_set, node)? "would also be fenced" : "also is unclean",
155  reason);
156 
157  } else {
158  crm_warn("Cluster node %s %s: %s",
159  node->details->uname,
160  pe_can_fence(data_set, node)? "will be fenced" : "is unclean",
161  reason);
162  node->details->unclean = TRUE;
163  pe_fence_op(node, NULL, TRUE, reason, priority_delay, data_set);
164  }
165 }
166 
167 // @TODO xpaths can't handle templates, rules, or id-refs
168 
169 // nvpair with provides or requires set to unfencing
170 #define XPATH_UNFENCING_NVPAIR XML_CIB_TAG_NVPAIR \
171  "[(@" XML_NVPAIR_ATTR_NAME "='" PCMK_STONITH_PROVIDES "'" \
172  "or @" XML_NVPAIR_ATTR_NAME "='" XML_RSC_ATTR_REQUIRES "') " \
173  "and @" XML_NVPAIR_ATTR_VALUE "='unfencing']"
174 
175 // unfencing in rsc_defaults or any resource
176 #define XPATH_ENABLE_UNFENCING \
177  "/" XML_TAG_CIB "/" XML_CIB_TAG_CONFIGURATION "/" XML_CIB_TAG_RESOURCES \
178  "//" XML_TAG_META_SETS "/" XPATH_UNFENCING_NVPAIR \
179  "|/" XML_TAG_CIB "/" XML_CIB_TAG_CONFIGURATION "/" XML_CIB_TAG_RSCCONFIG \
180  "/" XML_TAG_META_SETS "/" XPATH_UNFENCING_NVPAIR
181 
182 static void
183 set_if_xpath(uint64_t flag, const char *xpath, pe_working_set_t *data_set)
184 {
185  xmlXPathObjectPtr result = NULL;
186 
187  if (!pcmk_is_set(data_set->flags, flag)) {
188  result = xpath_search(data_set->input, xpath);
189  if (result && (numXpathResults(result) > 0)) {
190  pe__set_working_set_flags(data_set, flag);
191  }
192  freeXpathObject(result);
193  }
194 }
195 
196 gboolean
197 unpack_config(xmlNode * config, pe_working_set_t * data_set)
198 {
199  const char *value = NULL;
200  GHashTable *config_hash = pcmk__strkey_table(free, free);
201 
202  pe_rule_eval_data_t rule_data = {
203  .node_hash = NULL,
204  .role = RSC_ROLE_UNKNOWN,
205  .now = data_set->now,
206  .match_data = NULL,
207  .rsc_data = NULL,
208  .op_data = NULL
209  };
210 
211  data_set->config_hash = config_hash;
212 
213  pe__unpack_dataset_nvpairs(config, XML_CIB_TAG_PROPSET, &rule_data, config_hash,
214  CIB_OPTIONS_FIRST, FALSE, data_set);
215 
216  verify_pe_options(data_set->config_hash);
217 
218  set_config_flag(data_set, "enable-startup-probes", pe_flag_startup_probes);
219  if (!pcmk_is_set(data_set->flags, pe_flag_startup_probes)) {
220  crm_info("Startup probes: disabled (dangerous)");
221  }
222 
223  value = pe_pref(data_set->config_hash, XML_ATTR_HAVE_WATCHDOG);
224  if (value && crm_is_true(value)) {
225  crm_info("Watchdog-based self-fencing will be performed via SBD if "
226  "fencing is required and stonith-watchdog-timeout is nonzero");
228  }
229 
230  /* Set certain flags via xpath here, so they can be used before the relevant
231  * configuration sections are unpacked.
232  */
233  set_if_xpath(pe_flag_enable_unfencing, XPATH_ENABLE_UNFENCING, data_set);
234 
235  value = pe_pref(data_set->config_hash, "stonith-timeout");
236  data_set->stonith_timeout = (int) crm_parse_interval_spec(value);
237  crm_debug("STONITH timeout: %d", data_set->stonith_timeout);
238 
239  set_config_flag(data_set, "stonith-enabled", pe_flag_stonith_enabled);
240  crm_debug("STONITH of failed nodes is %s",
241  pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)? "enabled" : "disabled");
242 
243  data_set->stonith_action = pe_pref(data_set->config_hash, "stonith-action");
244  if (!strcmp(data_set->stonith_action, "poweroff")) {
246  "Support for stonith-action of 'poweroff' is deprecated "
247  "and will be removed in a future release (use 'off' instead)");
248  data_set->stonith_action = "off";
249  }
250  crm_trace("STONITH will %s nodes", data_set->stonith_action);
251 
252  set_config_flag(data_set, "concurrent-fencing", pe_flag_concurrent_fencing);
253  crm_debug("Concurrent fencing is %s",
254  pcmk_is_set(data_set->flags, pe_flag_concurrent_fencing)? "enabled" : "disabled");
255 
256  value = pe_pref(data_set->config_hash,
258  if (value) {
259  data_set->priority_fencing_delay = crm_parse_interval_spec(value) / 1000;
260  crm_trace("Priority fencing delay is %ds", data_set->priority_fencing_delay);
261  }
262 
263  set_config_flag(data_set, "stop-all-resources", pe_flag_stop_everything);
264  crm_debug("Stop all active resources: %s",
265  pcmk__btoa(pcmk_is_set(data_set->flags, pe_flag_stop_everything)));
266 
267  set_config_flag(data_set, "symmetric-cluster", pe_flag_symmetric_cluster);
268  if (pcmk_is_set(data_set->flags, pe_flag_symmetric_cluster)) {
269  crm_debug("Cluster is symmetric" " - resources can run anywhere by default");
270  }
271 
272  value = pe_pref(data_set->config_hash, "no-quorum-policy");
273 
274  if (pcmk__str_eq(value, "ignore", pcmk__str_casei)) {
276 
277  } else if (pcmk__str_eq(value, "freeze", pcmk__str_casei)) {
279 
280  } else if (pcmk__str_eq(value, "demote", pcmk__str_casei)) {
282 
283  } else if (pcmk__str_eq(value, "suicide", pcmk__str_casei)) {
284  if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
285  int do_panic = 0;
286 
288  &do_panic);
289  if (do_panic || pcmk_is_set(data_set->flags, pe_flag_have_quorum)) {
291  } else {
292  crm_notice("Resetting no-quorum-policy to 'stop': cluster has never had quorum");
293  data_set->no_quorum_policy = no_quorum_stop;
294  }
295  } else {
296  pcmk__config_err("Resetting no-quorum-policy to 'stop' because "
297  "fencing is disabled");
298  data_set->no_quorum_policy = no_quorum_stop;
299  }
300 
301  } else {
302  data_set->no_quorum_policy = no_quorum_stop;
303  }
304 
305  switch (data_set->no_quorum_policy) {
306  case no_quorum_freeze:
307  crm_debug("On loss of quorum: Freeze resources");
308  break;
309  case no_quorum_stop:
310  crm_debug("On loss of quorum: Stop ALL resources");
311  break;
312  case no_quorum_demote:
313  crm_debug("On loss of quorum: "
314  "Demote promotable resources and stop other resources");
315  break;
316  case no_quorum_suicide:
317  crm_notice("On loss of quorum: Fence all remaining nodes");
318  break;
319  case no_quorum_ignore:
320  crm_notice("On loss of quorum: Ignore");
321  break;
322  }
323 
324  set_config_flag(data_set, "stop-orphan-resources", pe_flag_stop_rsc_orphans);
325  crm_trace("Orphan resources are %s",
326  pcmk_is_set(data_set->flags, pe_flag_stop_rsc_orphans)? "stopped" : "ignored");
327 
328  set_config_flag(data_set, "stop-orphan-actions", pe_flag_stop_action_orphans);
329  crm_trace("Orphan resource actions are %s",
330  pcmk_is_set(data_set->flags, pe_flag_stop_action_orphans)? "stopped" : "ignored");
331 
332  value = pe_pref(data_set->config_hash, "remove-after-stop");
333  if (value != NULL) {
334  if (crm_is_true(value)) {
336 #ifndef PCMK__COMPAT_2_0
338  "Support for the remove-after-stop cluster property is"
339  " deprecated and will be removed in a future release");
340 #endif
341  } else {
343  }
344  }
345 
346  set_config_flag(data_set, "maintenance-mode", pe_flag_maintenance_mode);
347  crm_trace("Maintenance mode: %s",
348  pcmk__btoa(pcmk_is_set(data_set->flags, pe_flag_maintenance_mode)));
349 
350  set_config_flag(data_set, "start-failure-is-fatal", pe_flag_start_failure_fatal);
351  crm_trace("Start failures are %s",
352  pcmk_is_set(data_set->flags, pe_flag_start_failure_fatal)? "always fatal" : "handled by failcount");
353 
354  if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
355  set_config_flag(data_set, "startup-fencing", pe_flag_startup_fencing);
356  }
357  if (pcmk_is_set(data_set->flags, pe_flag_startup_fencing)) {
358  crm_trace("Unseen nodes will be fenced");
359  } else {
360  pe_warn_once(pe_wo_blind, "Blind faith: not fencing unseen nodes");
361  }
362 
363  pcmk__score_red = char2score(pe_pref(data_set->config_hash, "node-health-red"));
364  pcmk__score_green = char2score(pe_pref(data_set->config_hash, "node-health-green"));
365  pcmk__score_yellow = char2score(pe_pref(data_set->config_hash, "node-health-yellow"));
366 
367  crm_debug("Node scores: 'red' = %s, 'yellow' = %s, 'green' = %s",
368  pe_pref(data_set->config_hash, "node-health-red"),
369  pe_pref(data_set->config_hash, "node-health-yellow"),
370  pe_pref(data_set->config_hash, "node-health-green"));
371 
372  data_set->placement_strategy = pe_pref(data_set->config_hash, "placement-strategy");
373  crm_trace("Placement strategy: %s", data_set->placement_strategy);
374 
375  set_config_flag(data_set, "shutdown-lock", pe_flag_shutdown_lock);
376  crm_trace("Resources will%s be locked to cleanly shut down nodes",
377  (pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)? "" : " not"));
378  if (pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)) {
379  value = pe_pref(data_set->config_hash,
381  data_set->shutdown_lock = crm_parse_interval_spec(value) / 1000;
382  crm_trace("Shutdown locks expire after %us", data_set->shutdown_lock);
383  }
384 
385  return TRUE;
386 }
387 
388 pe_node_t *
389 pe_create_node(const char *id, const char *uname, const char *type,
390  const char *score, pe_working_set_t * data_set)
391 {
392  pe_node_t *new_node = NULL;
393 
394  if (pe_find_node(data_set->nodes, uname) != NULL) {
395  pcmk__config_warn("More than one node entry has name '%s'", uname);
396  }
397 
398  new_node = calloc(1, sizeof(pe_node_t));
399  if (new_node == NULL) {
400  return NULL;
401  }
402 
403  new_node->weight = char2score(score);
404  new_node->fixed = FALSE;
405  new_node->details = calloc(1, sizeof(struct pe_node_shared_s));
406 
407  if (new_node->details == NULL) {
408  free(new_node);
409  return NULL;
410  }
411 
412  crm_trace("Creating node for entry %s/%s", uname, id);
413  new_node->details->id = id;
414  new_node->details->uname = uname;
415  new_node->details->online = FALSE;
416  new_node->details->shutdown = FALSE;
417  new_node->details->rsc_discovery_enabled = TRUE;
418  new_node->details->running_rsc = NULL;
419  new_node->details->type = node_ping;
420 
421  if (pcmk__str_eq(type, "remote", pcmk__str_casei)) {
422  new_node->details->type = node_remote;
424  } else if (pcmk__str_eq(type, "member", pcmk__str_null_matches | pcmk__str_casei)) {
425  new_node->details->type = node_member;
426  }
427 
428  new_node->details->attrs = pcmk__strkey_table(free, free);
429 
430  if (pe__is_guest_or_remote_node(new_node)) {
431  g_hash_table_insert(new_node->details->attrs, strdup(CRM_ATTR_KIND),
432  strdup("remote"));
433  } else {
434  g_hash_table_insert(new_node->details->attrs, strdup(CRM_ATTR_KIND),
435  strdup("cluster"));
436  }
437 
438  new_node->details->utilization = pcmk__strkey_table(free, free);
439  new_node->details->digest_cache = pcmk__strkey_table(free,
441 
442  data_set->nodes = g_list_insert_sorted(data_set->nodes, new_node, sort_node_uname);
443  return new_node;
444 }
445 
446 static const char *
447 expand_remote_rsc_meta(xmlNode *xml_obj, xmlNode *parent, pe_working_set_t *data)
448 {
449  xmlNode *attr_set = NULL;
450  xmlNode *attr = NULL;
451 
452  const char *container_id = ID(xml_obj);
453  const char *remote_name = NULL;
454  const char *remote_server = NULL;
455  const char *remote_port = NULL;
456  const char *connect_timeout = "60s";
457  const char *remote_allow_migrate=NULL;
458  const char *is_managed = NULL;
459 
460  for (attr_set = pcmk__xe_first_child(xml_obj); attr_set != NULL;
461  attr_set = pcmk__xe_next(attr_set)) {
462 
463  if (!pcmk__str_eq((const char *)attr_set->name, XML_TAG_META_SETS,
464  pcmk__str_casei)) {
465  continue;
466  }
467 
468  for (attr = pcmk__xe_first_child(attr_set); attr != NULL;
469  attr = pcmk__xe_next(attr)) {
470  const char *value = crm_element_value(attr, XML_NVPAIR_ATTR_VALUE);
471  const char *name = crm_element_value(attr, XML_NVPAIR_ATTR_NAME);
472 
473  if (pcmk__str_eq(name, XML_RSC_ATTR_REMOTE_NODE, pcmk__str_casei)) {
474  remote_name = value;
475  } else if (pcmk__str_eq(name, "remote-addr", pcmk__str_casei)) {
476  remote_server = value;
477  } else if (pcmk__str_eq(name, "remote-port", pcmk__str_casei)) {
478  remote_port = value;
479  } else if (pcmk__str_eq(name, "remote-connect-timeout", pcmk__str_casei)) {
480  connect_timeout = value;
481  } else if (pcmk__str_eq(name, "remote-allow-migrate", pcmk__str_casei)) {
482  remote_allow_migrate=value;
483  } else if (pcmk__str_eq(name, XML_RSC_ATTR_MANAGED, pcmk__str_casei)) {
484  is_managed = value;
485  }
486  }
487  }
488 
489  if (remote_name == NULL) {
490  return NULL;
491  }
492 
493  if (pe_find_resource(data->resources, remote_name) != NULL) {
494  return NULL;
495  }
496 
497  pe_create_remote_xml(parent, remote_name, container_id,
498  remote_allow_migrate, is_managed,
499  connect_timeout, remote_server, remote_port);
500  return remote_name;
501 }
502 
503 static void
504 handle_startup_fencing(pe_working_set_t *data_set, pe_node_t *new_node)
505 {
506  if ((new_node->details->type == node_remote) && (new_node->details->remote_rsc == NULL)) {
507  /* Ignore fencing for remote nodes that don't have a connection resource
508  * associated with them. This happens when remote node entries get left
509  * in the nodes section after the connection resource is removed.
510  */
511  return;
512  }
513 
514  if (pcmk_is_set(data_set->flags, pe_flag_startup_fencing)) {
515  // All nodes are unclean until we've seen their status entry
516  new_node->details->unclean = TRUE;
517 
518  } else {
519  // Blind faith ...
520  new_node->details->unclean = FALSE;
521  }
522 
523  /* We need to be able to determine if a node's status section
524  * exists or not separate from whether the node is unclean. */
525  new_node->details->unseen = TRUE;
526 }
527 
528 gboolean
529 unpack_nodes(xmlNode * xml_nodes, pe_working_set_t * data_set)
530 {
531  xmlNode *xml_obj = NULL;
532  pe_node_t *new_node = NULL;
533  const char *id = NULL;
534  const char *uname = NULL;
535  const char *type = NULL;
536  const char *score = NULL;
537 
538  pe_rule_eval_data_t rule_data = {
539  .node_hash = NULL,
540  .role = RSC_ROLE_UNKNOWN,
541  .now = data_set->now,
542  .match_data = NULL,
543  .rsc_data = NULL,
544  .op_data = NULL
545  };
546 
547  for (xml_obj = pcmk__xe_first_child(xml_nodes); xml_obj != NULL;
548  xml_obj = pcmk__xe_next(xml_obj)) {
549 
550  if (pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_NODE, pcmk__str_none)) {
551  new_node = NULL;
552 
553  id = crm_element_value(xml_obj, XML_ATTR_ID);
556  score = crm_element_value(xml_obj, XML_RULE_ATTR_SCORE);
557  crm_trace("Processing node %s/%s", uname, id);
558 
559  if (id == NULL) {
561  "> entry in configuration without id");
562  continue;
563  }
564  new_node = pe_create_node(id, uname, type, score, data_set);
565 
566  if (new_node == NULL) {
567  return FALSE;
568  }
569 
570 /* if(data_set->have_quorum == FALSE */
571 /* && data_set->no_quorum_policy == no_quorum_stop) { */
572 /* /\* start shutting resources down *\/ */
573 /* new_node->weight = -INFINITY; */
574 /* } */
575 
576  handle_startup_fencing(data_set, new_node);
577 
578  add_node_attrs(xml_obj, new_node, FALSE, data_set);
579  pe__unpack_dataset_nvpairs(xml_obj, XML_TAG_UTILIZATION, &rule_data,
580  new_node->details->utilization, NULL,
581  FALSE, data_set);
582 
583  crm_trace("Done with node %s", crm_element_value(xml_obj, XML_ATTR_UNAME));
584  }
585  }
586 
587  if (data_set->localhost && pe_find_node(data_set->nodes, data_set->localhost) == NULL) {
588  crm_info("Creating a fake local node");
589  pe_create_node(data_set->localhost, data_set->localhost, NULL, 0,
590  data_set);
591  }
592 
593  return TRUE;
594 }
595 
596 static void
597 setup_container(pe_resource_t * rsc, pe_working_set_t * data_set)
598 {
599  const char *container_id = NULL;
600 
601  if (rsc->children) {
602  g_list_foreach(rsc->children, (GFunc) setup_container, data_set);
603  return;
604  }
605 
606  container_id = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_CONTAINER);
607  if (container_id && !pcmk__str_eq(container_id, rsc->id, pcmk__str_casei)) {
608  pe_resource_t *container = pe_find_resource(data_set->resources, container_id);
609 
610  if (container) {
611  rsc->container = container;
613  container->fillers = g_list_append(container->fillers, rsc);
614  pe_rsc_trace(rsc, "Resource %s's container is %s", rsc->id, container_id);
615  } else {
616  pe_err("Resource %s: Unknown resource container (%s)", rsc->id, container_id);
617  }
618  }
619 }
620 
621 gboolean
622 unpack_remote_nodes(xmlNode * xml_resources, pe_working_set_t * data_set)
623 {
624  xmlNode *xml_obj = NULL;
625 
626  /* Create remote nodes and guest nodes from the resource configuration
627  * before unpacking resources.
628  */
629  for (xml_obj = pcmk__xe_first_child(xml_resources); xml_obj != NULL;
630  xml_obj = pcmk__xe_next(xml_obj)) {
631 
632  const char *new_node_id = NULL;
633 
634  /* Check for remote nodes, which are defined by ocf:pacemaker:remote
635  * primitives.
636  */
637  if (xml_contains_remote_node(xml_obj)) {
638  new_node_id = ID(xml_obj);
639  /* The "pe_find_node" check is here to make sure we don't iterate over
640  * an expanded node that has already been added to the node list. */
641  if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) {
642  crm_trace("Found remote node %s defined by resource %s",
643  new_node_id, ID(xml_obj));
644  pe_create_node(new_node_id, new_node_id, "remote", NULL,
645  data_set);
646  }
647  continue;
648  }
649 
650  /* Check for guest nodes, which are defined by special meta-attributes
651  * of a primitive of any type (for example, VirtualDomain or Xen).
652  */
653  if (pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_RESOURCE, pcmk__str_none)) {
654  /* This will add an ocf:pacemaker:remote primitive to the
655  * configuration for the guest node's connection, to be unpacked
656  * later.
657  */
658  new_node_id = expand_remote_rsc_meta(xml_obj, xml_resources, data_set);
659  if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) {
660  crm_trace("Found guest node %s in resource %s",
661  new_node_id, ID(xml_obj));
662  pe_create_node(new_node_id, new_node_id, "remote", NULL,
663  data_set);
664  }
665  continue;
666  }
667 
668  /* Check for guest nodes inside a group. Clones are currently not
669  * supported as guest nodes.
670  */
671  if (pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_GROUP, pcmk__str_none)) {
672  xmlNode *xml_obj2 = NULL;
673  for (xml_obj2 = pcmk__xe_first_child(xml_obj); xml_obj2 != NULL;
674  xml_obj2 = pcmk__xe_next(xml_obj2)) {
675 
676  new_node_id = expand_remote_rsc_meta(xml_obj2, xml_resources, data_set);
677 
678  if (new_node_id && pe_find_node(data_set->nodes, new_node_id) == NULL) {
679  crm_trace("Found guest node %s in resource %s inside group %s",
680  new_node_id, ID(xml_obj2), ID(xml_obj));
681  pe_create_node(new_node_id, new_node_id, "remote", NULL,
682  data_set);
683  }
684  }
685  }
686  }
687  return TRUE;
688 }
689 
690 /* Call this after all the nodes and resources have been
691  * unpacked, but before the status section is read.
692  *
693  * A remote node's online status is reflected by the state
694  * of the remote node's connection resource. We need to link
695  * the remote node to this connection resource so we can have
696  * easy access to the connection resource during the scheduler calculations.
697  */
698 static void
699 link_rsc2remotenode(pe_working_set_t *data_set, pe_resource_t *new_rsc)
700 {
701  pe_node_t *remote_node = NULL;
702 
703  if (new_rsc->is_remote_node == FALSE) {
704  return;
705  }
706 
707  if (pcmk_is_set(data_set->flags, pe_flag_quick_location)) {
708  /* remote_nodes and remote_resources are not linked in quick location calculations */
709  return;
710  }
711 
712  remote_node = pe_find_node(data_set->nodes, new_rsc->id);
713  CRM_CHECK(remote_node != NULL, return;);
714 
715  pe_rsc_trace(new_rsc, "Linking remote connection resource %s to node %s",
716  new_rsc->id, remote_node->details->uname);
717  remote_node->details->remote_rsc = new_rsc;
718 
719  if (new_rsc->container == NULL) {
720  /* Handle start-up fencing for remote nodes (as opposed to guest nodes)
721  * the same as is done for cluster nodes.
722  */
723  handle_startup_fencing(data_set, remote_node);
724 
725  } else {
726  /* pe_create_node() marks the new node as "remote" or "cluster"; now
727  * that we know the node is a guest node, update it correctly.
728  */
729  g_hash_table_replace(remote_node->details->attrs, strdup(CRM_ATTR_KIND),
730  strdup("container"));
731  }
732 }
733 
734 static void
735 destroy_tag(gpointer data)
736 {
737  pe_tag_t *tag = data;
738 
739  if (tag) {
740  free(tag->id);
741  g_list_free_full(tag->refs, free);
742  free(tag);
743  }
744 }
745 
758 gboolean
759 unpack_resources(xmlNode * xml_resources, pe_working_set_t * data_set)
760 {
761  xmlNode *xml_obj = NULL;
762  GList *gIter = NULL;
763 
764  data_set->template_rsc_sets = pcmk__strkey_table(free, destroy_tag);
765 
766  for (xml_obj = pcmk__xe_first_child(xml_resources); xml_obj != NULL;
767  xml_obj = pcmk__xe_next(xml_obj)) {
768 
769  pe_resource_t *new_rsc = NULL;
770 
771  if (pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_RSC_TEMPLATE, pcmk__str_none)) {
772  const char *template_id = ID(xml_obj);
773 
774  if (template_id && g_hash_table_lookup_extended(data_set->template_rsc_sets,
775  template_id, NULL, NULL) == FALSE) {
776  /* Record the template's ID for the knowledge of its existence anyway. */
777  g_hash_table_insert(data_set->template_rsc_sets, strdup(template_id), NULL);
778  }
779  continue;
780  }
781 
782  crm_trace("Beginning unpack... <%s id=%s... >", crm_element_name(xml_obj), ID(xml_obj));
783  if (common_unpack(xml_obj, &new_rsc, NULL, data_set) && (new_rsc != NULL)) {
784  data_set->resources = g_list_append(data_set->resources, new_rsc);
785  pe_rsc_trace(new_rsc, "Added resource %s", new_rsc->id);
786 
787  } else {
788  pcmk__config_err("Ignoring <%s> resource '%s' "
789  "because configuration is invalid",
790  crm_element_name(xml_obj), crm_str(ID(xml_obj)));
791  if (new_rsc != NULL && new_rsc->fns != NULL) {
792  new_rsc->fns->free(new_rsc);
793  }
794  }
795  }
796 
797  for (gIter = data_set->resources; gIter != NULL; gIter = gIter->next) {
798  pe_resource_t *rsc = (pe_resource_t *) gIter->data;
799 
800  setup_container(rsc, data_set);
801  link_rsc2remotenode(data_set, rsc);
802  }
803 
804  data_set->resources = g_list_sort(data_set->resources, sort_rsc_priority);
805  if (pcmk_is_set(data_set->flags, pe_flag_quick_location)) {
806  /* Ignore */
807 
808  } else if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)
810 
811  pcmk__config_err("Resource start-up disabled since no STONITH resources have been defined");
812  pcmk__config_err("Either configure some or disable STONITH with the stonith-enabled option");
813  pcmk__config_err("NOTE: Clusters with shared data need STONITH to ensure data integrity");
814  }
815 
816  return TRUE;
817 }
818 
819 gboolean
820 unpack_tags(xmlNode * xml_tags, pe_working_set_t * data_set)
821 {
822  xmlNode *xml_tag = NULL;
823 
824  data_set->tags = pcmk__strkey_table(free, destroy_tag);
825 
826  for (xml_tag = pcmk__xe_first_child(xml_tags); xml_tag != NULL;
827  xml_tag = pcmk__xe_next(xml_tag)) {
828 
829  xmlNode *xml_obj_ref = NULL;
830  const char *tag_id = ID(xml_tag);
831 
832  if (!pcmk__str_eq((const char *)xml_tag->name, XML_CIB_TAG_TAG, pcmk__str_none)) {
833  continue;
834  }
835 
836  if (tag_id == NULL) {
837  pcmk__config_err("Ignoring <%s> without " XML_ATTR_ID,
838  crm_element_name(xml_tag));
839  continue;
840  }
841 
842  for (xml_obj_ref = pcmk__xe_first_child(xml_tag); xml_obj_ref != NULL;
843  xml_obj_ref = pcmk__xe_next(xml_obj_ref)) {
844 
845  const char *obj_ref = ID(xml_obj_ref);
846 
847  if (!pcmk__str_eq((const char *)xml_obj_ref->name, XML_CIB_TAG_OBJ_REF, pcmk__str_none)) {
848  continue;
849  }
850 
851  if (obj_ref == NULL) {
852  pcmk__config_err("Ignoring <%s> for tag '%s' without " XML_ATTR_ID,
853  crm_element_name(xml_obj_ref), tag_id);
854  continue;
855  }
856 
857  if (add_tag_ref(data_set->tags, tag_id, obj_ref) == FALSE) {
858  return FALSE;
859  }
860  }
861  }
862 
863  return TRUE;
864 }
865 
866 /* The ticket state section:
867  * "/cib/status/tickets/ticket_state" */
868 static gboolean
869 unpack_ticket_state(xmlNode * xml_ticket, pe_working_set_t * data_set)
870 {
871  const char *ticket_id = NULL;
872  const char *granted = NULL;
873  const char *last_granted = NULL;
874  const char *standby = NULL;
875  xmlAttrPtr xIter = NULL;
876 
877  pe_ticket_t *ticket = NULL;
878 
879  ticket_id = ID(xml_ticket);
880  if (pcmk__str_empty(ticket_id)) {
881  return FALSE;
882  }
883 
884  crm_trace("Processing ticket state for %s", ticket_id);
885 
886  ticket = g_hash_table_lookup(data_set->tickets, ticket_id);
887  if (ticket == NULL) {
888  ticket = ticket_new(ticket_id, data_set);
889  if (ticket == NULL) {
890  return FALSE;
891  }
892  }
893 
894  for (xIter = xml_ticket->properties; xIter; xIter = xIter->next) {
895  const char *prop_name = (const char *)xIter->name;
896  const char *prop_value = crm_element_value(xml_ticket, prop_name);
897 
898  if (pcmk__str_eq(prop_name, XML_ATTR_ID, pcmk__str_none)) {
899  continue;
900  }
901  g_hash_table_replace(ticket->state, strdup(prop_name), strdup(prop_value));
902  }
903 
904  granted = g_hash_table_lookup(ticket->state, "granted");
905  if (granted && crm_is_true(granted)) {
906  ticket->granted = TRUE;
907  crm_info("We have ticket '%s'", ticket->id);
908  } else {
909  ticket->granted = FALSE;
910  crm_info("We do not have ticket '%s'", ticket->id);
911  }
912 
913  last_granted = g_hash_table_lookup(ticket->state, "last-granted");
914  if (last_granted) {
915  long long last_granted_ll;
916 
917  pcmk__scan_ll(last_granted, &last_granted_ll, 0LL);
918  ticket->last_granted = (time_t) last_granted_ll;
919  }
920 
921  standby = g_hash_table_lookup(ticket->state, "standby");
922  if (standby && crm_is_true(standby)) {
923  ticket->standby = TRUE;
924  if (ticket->granted) {
925  crm_info("Granted ticket '%s' is in standby-mode", ticket->id);
926  }
927  } else {
928  ticket->standby = FALSE;
929  }
930 
931  crm_trace("Done with ticket state for %s", ticket_id);
932 
933  return TRUE;
934 }
935 
936 static gboolean
937 unpack_tickets_state(xmlNode * xml_tickets, pe_working_set_t * data_set)
938 {
939  xmlNode *xml_obj = NULL;
940 
941  for (xml_obj = pcmk__xe_first_child(xml_tickets); xml_obj != NULL;
942  xml_obj = pcmk__xe_next(xml_obj)) {
943 
944  if (!pcmk__str_eq((const char *)xml_obj->name, XML_CIB_TAG_TICKET_STATE, pcmk__str_none)) {
945  continue;
946  }
947  unpack_ticket_state(xml_obj, data_set);
948  }
949 
950  return TRUE;
951 }
952 
953 static void
954 unpack_handle_remote_attrs(pe_node_t *this_node, xmlNode *state, pe_working_set_t * data_set)
955 {
956  const char *resource_discovery_enabled = NULL;
957  xmlNode *attrs = NULL;
958  pe_resource_t *rsc = NULL;
959 
960  if (!pcmk__str_eq((const char *)state->name, XML_CIB_TAG_STATE, pcmk__str_none)) {
961  return;
962  }
963 
964  if ((this_node == NULL) || !pe__is_guest_or_remote_node(this_node)) {
965  return;
966  }
967  crm_trace("Processing remote node id=%s, uname=%s", this_node->details->id, this_node->details->uname);
968 
970  &(this_node->details->remote_maintenance), 0);
971 
972  rsc = this_node->details->remote_rsc;
973  if (this_node->details->remote_requires_reset == FALSE) {
974  this_node->details->unclean = FALSE;
975  this_node->details->unseen = FALSE;
976  }
977  attrs = find_xml_node(state, XML_TAG_TRANSIENT_NODEATTRS, FALSE);
978  add_node_attrs(attrs, this_node, TRUE, data_set);
979 
980  if (pe__shutdown_requested(this_node)) {
981  crm_info("Node %s is shutting down", this_node->details->uname);
982  this_node->details->shutdown = TRUE;
983  }
984 
985  if (crm_is_true(pe_node_attribute_raw(this_node, "standby"))) {
986  crm_info("Node %s is in standby-mode", this_node->details->uname);
987  this_node->details->standby = TRUE;
988  }
989 
990  if (crm_is_true(pe_node_attribute_raw(this_node, "maintenance")) ||
991  ((rsc != NULL) && !pcmk_is_set(rsc->flags, pe_rsc_managed))) {
992  crm_info("Node %s is in maintenance-mode", this_node->details->uname);
993  this_node->details->maintenance = TRUE;
994  }
995 
996  resource_discovery_enabled = pe_node_attribute_raw(this_node, XML_NODE_ATTR_RSC_DISCOVERY);
997  if (resource_discovery_enabled && !crm_is_true(resource_discovery_enabled)) {
998  if (pe__is_remote_node(this_node)
999  && !pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
1000  crm_warn("Ignoring %s attribute on remote node %s because stonith is disabled",
1002  } else {
1003  /* This is either a remote node with fencing enabled, or a guest
1004  * node. We don't care whether fencing is enabled when fencing guest
1005  * nodes, because they are "fenced" by recovering their containing
1006  * resource.
1007  */
1008  crm_info("Node %s has resource discovery disabled", this_node->details->uname);
1009  this_node->details->rsc_discovery_enabled = FALSE;
1010  }
1011  }
1012 }
1013 
1022 static void
1023 unpack_transient_attributes(xmlNode *state, pe_node_t *node,
1024  pe_working_set_t *data_set)
1025 {
1026  const char *discovery = NULL;
1027  xmlNode *attrs = find_xml_node(state, XML_TAG_TRANSIENT_NODEATTRS, FALSE);
1028 
1029  add_node_attrs(attrs, node, TRUE, data_set);
1030 
1031  if (crm_is_true(pe_node_attribute_raw(node, "standby"))) {
1032  crm_info("Node %s is in standby-mode", node->details->uname);
1033  node->details->standby = TRUE;
1034  }
1035 
1036  if (crm_is_true(pe_node_attribute_raw(node, "maintenance"))) {
1037  crm_info("Node %s is in maintenance-mode", node->details->uname);
1038  node->details->maintenance = TRUE;
1039  }
1040 
1042  if ((discovery != NULL) && !crm_is_true(discovery)) {
1043  crm_warn("Ignoring %s attribute for node %s because disabling "
1044  "resource discovery is not allowed for cluster nodes",
1046  }
1047 }
1048 
1061 static void
1062 unpack_node_state(xmlNode *state, pe_working_set_t *data_set)
1063 {
1064  const char *id = NULL;
1065  const char *uname = NULL;
1066  pe_node_t *this_node = NULL;
1067 
1068  id = crm_element_value(state, XML_ATTR_ID);
1069  if (id == NULL) {
1070  crm_warn("Ignoring malformed " XML_CIB_TAG_STATE " entry without "
1071  XML_ATTR_ID);
1072  return;
1073  }
1074 
1076  if (uname == NULL) {
1077  crm_warn("Ignoring malformed " XML_CIB_TAG_STATE " entry without "
1078  XML_ATTR_UNAME);
1079  return;
1080  }
1081 
1082  this_node = pe_find_node_any(data_set->nodes, id, uname);
1083  if (this_node == NULL) {
1084  pcmk__config_warn("Ignoring recorded node state for '%s' because "
1085  "it is no longer in the configuration", uname);
1086  return;
1087  }
1088 
1089  if (pe__is_guest_or_remote_node(this_node)) {
1090  /* We can't determine the online status of Pacemaker Remote nodes until
1091  * after all resource history has been unpacked. In this first pass, we
1092  * do need to mark whether the node has been fenced, as this plays a
1093  * role during unpacking cluster node resource state.
1094  */
1096  &(this_node->details->remote_was_fenced), 0);
1097  return;
1098  }
1099 
1100  unpack_transient_attributes(state, this_node, data_set);
1101 
1102  /* Provisionally mark this cluster node as clean. We have at least seen it
1103  * in the current cluster's lifetime.
1104  */
1105  this_node->details->unclean = FALSE;
1106  this_node->details->unseen = FALSE;
1107 
1108  crm_trace("Determining online status of cluster node %s (id %s)",
1109  this_node->details->uname, id);
1110  determine_online_status(state, this_node, data_set);
1111 
1112  if (!pcmk_is_set(data_set->flags, pe_flag_have_quorum)
1113  && this_node->details->online
1114  && (data_set->no_quorum_policy == no_quorum_suicide)) {
1115  /* Everything else should flow from this automatically
1116  * (at least until the scheduler becomes able to migrate off
1117  * healthy resources)
1118  */
1119  pe_fence_node(data_set, this_node, "cluster does not have quorum",
1120  FALSE);
1121  }
1122 }
1123 
1141 static int
1142 unpack_node_history(xmlNode *status, bool fence, pe_working_set_t *data_set)
1143 {
1144  int rc = pcmk_rc_ok;
1145 
1146  // Loop through all node_state entries in CIB status
1147  for (xmlNode *state = first_named_child(status, XML_CIB_TAG_STATE);
1148  state != NULL; state = crm_next_same_xml(state)) {
1149 
1150  const char *id = ID(state);
1151  const char *uname = crm_element_value(state, XML_ATTR_UNAME);
1152  pe_node_t *this_node = NULL;
1153 
1154  if ((id == NULL) || (uname == NULL)) {
1155  // Warning already logged in first pass through status section
1156  crm_trace("Not unpacking resource history from malformed "
1157  XML_CIB_TAG_STATE " without id and/or uname");
1158  continue;
1159  }
1160 
1161  this_node = pe_find_node_any(data_set->nodes, id, uname);
1162  if (this_node == NULL) {
1163  // Warning already logged in first pass through status section
1164  crm_trace("Not unpacking resource history for node %s because "
1165  "no longer in configuration", id);
1166  continue;
1167  }
1168 
1169  if (this_node->details->unpacked) {
1170  crm_trace("Not unpacking resource history for node %s because "
1171  "already unpacked", id);
1172  continue;
1173  }
1174 
1175  if (fence) {
1176  // We're processing all remaining nodes
1177 
1178  } else if (pe__is_guest_node(this_node)) {
1179  /* We can unpack a guest node's history only after we've unpacked
1180  * other resource history to the point that we know that the node's
1181  * connection and containing resource are both up.
1182  */
1183  pe_resource_t *rsc = this_node->details->remote_rsc;
1184 
1185  if ((rsc == NULL) || (rsc->role != RSC_ROLE_STARTED)
1186  || (rsc->container->role != RSC_ROLE_STARTED)) {
1187  crm_trace("Not unpacking resource history for guest node %s "
1188  "because container and connection are not known to "
1189  "be up", id);
1190  continue;
1191  }
1192 
1193  } else if (pe__is_remote_node(this_node)) {
1194  /* We can unpack a remote node's history only after we've unpacked
1195  * other resource history to the point that we know that the node's
1196  * connection is up, with the exception of when shutdown locks are
1197  * in use.
1198  */
1199  pe_resource_t *rsc = this_node->details->remote_rsc;
1200 
1201  if ((rsc == NULL)
1202  || (!pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)
1203  && (rsc->role != RSC_ROLE_STARTED))) {
1204  crm_trace("Not unpacking resource history for remote node %s "
1205  "because connection is not known to be up", id);
1206  continue;
1207  }
1208 
1209  /* If fencing and shutdown locks are disabled and we're not processing
1210  * unseen nodes, then we don't want to unpack offline nodes until online
1211  * nodes have been unpacked. This allows us to number active clone
1212  * instances first.
1213  */
1214  } else if (!pcmk_any_flags_set(data_set->flags, pe_flag_stonith_enabled
1216  && !this_node->details->online) {
1217  crm_trace("Not unpacking resource history for offline "
1218  "cluster node %s", id);
1219  continue;
1220  }
1221 
1222  if (pe__is_guest_or_remote_node(this_node)) {
1223  determine_remote_online_status(data_set, this_node);
1224  unpack_handle_remote_attrs(this_node, state, data_set);
1225  }
1226 
1227  crm_trace("Unpacking resource history for %snode %s",
1228  (fence? "unseen " : ""), id);
1229 
1230  this_node->details->unpacked = TRUE;
1231  unpack_node_lrm(this_node, state, data_set);
1232 
1233  rc = EAGAIN; // Other node histories might depend on this one
1234  }
1235  return rc;
1236 }
1237 
1238 /* remove nodes that are down, stopping */
1239 /* create positive rsc_to_node constraints between resources and the nodes they are running on */
1240 /* anything else? */
1241 gboolean
1242 unpack_status(xmlNode * status, pe_working_set_t * data_set)
1243 {
1244  xmlNode *state = NULL;
1245 
1246  crm_trace("Beginning unpack");
1247 
1248  if (data_set->tickets == NULL) {
1249  data_set->tickets = pcmk__strkey_table(free, destroy_ticket);
1250  }
1251 
1252  for (state = pcmk__xe_first_child(status); state != NULL;
1253  state = pcmk__xe_next(state)) {
1254 
1255  if (pcmk__str_eq((const char *)state->name, XML_CIB_TAG_TICKETS, pcmk__str_none)) {
1256  unpack_tickets_state((xmlNode *) state, data_set);
1257 
1258  } else if (pcmk__str_eq((const char *)state->name, XML_CIB_TAG_STATE, pcmk__str_none)) {
1259  unpack_node_state(state, data_set);
1260  }
1261  }
1262 
1263  while (unpack_node_history(status, FALSE, data_set) == EAGAIN) {
1264  crm_trace("Another pass through node resource histories is needed");
1265  }
1266 
1267  // Now catch any nodes we didn't see
1268  unpack_node_history(status,
1270  data_set);
1271 
1272  /* Now that we know where resources are, we can schedule stops of containers
1273  * with failed bundle connections
1274  */
1275  if (data_set->stop_needed != NULL) {
1276  for (GList *item = data_set->stop_needed; item; item = item->next) {
1277  pe_resource_t *container = item->data;
1278  pe_node_t *node = pe__current_node(container);
1279 
1280  if (node) {
1281  stop_action(container, node, FALSE);
1282  }
1283  }
1284  g_list_free(data_set->stop_needed);
1285  data_set->stop_needed = NULL;
1286  }
1287 
1288  /* Now that we know status of all Pacemaker Remote connections and nodes,
1289  * we can stop connections for node shutdowns, and check the online status
1290  * of remote/guest nodes that didn't have any node history to unpack.
1291  */
1292  for (GList *gIter = data_set->nodes; gIter != NULL; gIter = gIter->next) {
1293  pe_node_t *this_node = gIter->data;
1294 
1295  if (!pe__is_guest_or_remote_node(this_node)) {
1296  continue;
1297  }
1298  if (this_node->details->shutdown
1299  && (this_node->details->remote_rsc != NULL)) {
1301  "remote shutdown");
1302  }
1303  if (!this_node->details->unpacked) {
1304  determine_remote_online_status(data_set, this_node);
1305  }
1306  }
1307 
1308  return TRUE;
1309 }
1310 
1311 static gboolean
1312 determine_online_status_no_fencing(pe_working_set_t * data_set, xmlNode * node_state,
1313  pe_node_t * this_node)
1314 {
1315  gboolean online = FALSE;
1316  const char *join = crm_element_value(node_state, XML_NODE_JOIN_STATE);
1317  const char *is_peer = crm_element_value(node_state, XML_NODE_IS_PEER);
1318  const char *in_cluster = crm_element_value(node_state, XML_NODE_IN_CLUSTER);
1319  const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED);
1320 
1321  if (!crm_is_true(in_cluster)) {
1322  crm_trace("Node is down: in_cluster=%s", crm_str(in_cluster));
1323 
1324  } else if (pcmk__str_eq(is_peer, ONLINESTATUS, pcmk__str_casei)) {
1325  if (pcmk__str_eq(join, CRMD_JOINSTATE_MEMBER, pcmk__str_casei)) {
1326  online = TRUE;
1327  } else {
1328  crm_debug("Node is not ready to run resources: %s", join);
1329  }
1330 
1331  } else if (this_node->details->expected_up == FALSE) {
1332  crm_trace("Controller is down: in_cluster=%s", crm_str(in_cluster));
1333  crm_trace("\tis_peer=%s, join=%s, expected=%s",
1334  crm_str(is_peer), crm_str(join), crm_str(exp_state));
1335 
1336  } else {
1337  /* mark it unclean */
1338  pe_fence_node(data_set, this_node, "peer is unexpectedly down", FALSE);
1339  crm_info("\tin_cluster=%s, is_peer=%s, join=%s, expected=%s",
1340  crm_str(in_cluster), crm_str(is_peer), crm_str(join), crm_str(exp_state));
1341  }
1342  return online;
1343 }
1344 
1345 static gboolean
1346 determine_online_status_fencing(pe_working_set_t * data_set, xmlNode * node_state,
1347  pe_node_t * this_node)
1348 {
1349  gboolean online = FALSE;
1350  gboolean do_terminate = FALSE;
1351  bool crmd_online = FALSE;
1352  const char *join = crm_element_value(node_state, XML_NODE_JOIN_STATE);
1353  const char *is_peer = crm_element_value(node_state, XML_NODE_IS_PEER);
1354  const char *in_cluster = crm_element_value(node_state, XML_NODE_IN_CLUSTER);
1355  const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED);
1356  const char *terminate = pe_node_attribute_raw(this_node, "terminate");
1357 
1358 /*
1359  - XML_NODE_IN_CLUSTER ::= true|false
1360  - XML_NODE_IS_PEER ::= online|offline
1361  - XML_NODE_JOIN_STATE ::= member|down|pending|banned
1362  - XML_NODE_EXPECTED ::= member|down
1363 */
1364 
1365  if (crm_is_true(terminate)) {
1366  do_terminate = TRUE;
1367 
1368  } else if (terminate != NULL && strlen(terminate) > 0) {
1369  /* could be a time() value */
1370  char t = terminate[0];
1371 
1372  if (t != '0' && isdigit(t)) {
1373  do_terminate = TRUE;
1374  }
1375  }
1376 
1377  crm_trace("%s: in_cluster=%s, is_peer=%s, join=%s, expected=%s, term=%d",
1378  this_node->details->uname, crm_str(in_cluster), crm_str(is_peer),
1379  crm_str(join), crm_str(exp_state), do_terminate);
1380 
1381  online = crm_is_true(in_cluster);
1382  crmd_online = pcmk__str_eq(is_peer, ONLINESTATUS, pcmk__str_casei);
1383  if (exp_state == NULL) {
1384  exp_state = CRMD_JOINSTATE_DOWN;
1385  }
1386 
1387  if (this_node->details->shutdown) {
1388  crm_debug("%s is shutting down", this_node->details->uname);
1389 
1390  /* Slightly different criteria since we can't shut down a dead peer */
1391  online = crmd_online;
1392 
1393  } else if (in_cluster == NULL) {
1394  pe_fence_node(data_set, this_node, "peer has not been seen by the cluster", FALSE);
1395 
1396  } else if (pcmk__str_eq(join, CRMD_JOINSTATE_NACK, pcmk__str_casei)) {
1397  pe_fence_node(data_set, this_node,
1398  "peer failed Pacemaker membership criteria", FALSE);
1399 
1400  } else if (do_terminate == FALSE && pcmk__str_eq(exp_state, CRMD_JOINSTATE_DOWN, pcmk__str_casei)) {
1401 
1402  if (crm_is_true(in_cluster) || crmd_online) {
1403  crm_info("- Node %s is not ready to run resources", this_node->details->uname);
1404  this_node->details->standby = TRUE;
1405  this_node->details->pending = TRUE;
1406 
1407  } else {
1408  crm_trace("%s is down or still coming up", this_node->details->uname);
1409  }
1410 
1411  } else if (do_terminate && pcmk__str_eq(join, CRMD_JOINSTATE_DOWN, pcmk__str_casei)
1412  && crm_is_true(in_cluster) == FALSE && !crmd_online) {
1413  crm_info("Node %s was just shot", this_node->details->uname);
1414  online = FALSE;
1415 
1416  } else if (crm_is_true(in_cluster) == FALSE) {
1417  // Consider `priority-fencing-delay` for lost nodes
1418  pe_fence_node(data_set, this_node, "peer is no longer part of the cluster", TRUE);
1419 
1420  } else if (!crmd_online) {
1421  pe_fence_node(data_set, this_node, "peer process is no longer available", FALSE);
1422 
1423  /* Everything is running at this point, now check join state */
1424  } else if (do_terminate) {
1425  pe_fence_node(data_set, this_node, "termination was requested", FALSE);
1426 
1427  } else if (pcmk__str_eq(join, CRMD_JOINSTATE_MEMBER, pcmk__str_casei)) {
1428  crm_info("Node %s is active", this_node->details->uname);
1429 
1431  crm_info("Node %s is not ready to run resources", this_node->details->uname);
1432  this_node->details->standby = TRUE;
1433  this_node->details->pending = TRUE;
1434 
1435  } else {
1436  pe_fence_node(data_set, this_node, "peer was in an unknown state", FALSE);
1437  crm_warn("%s: in-cluster=%s, is-peer=%s, join=%s, expected=%s, term=%d, shutdown=%d",
1438  this_node->details->uname, crm_str(in_cluster), crm_str(is_peer),
1439  crm_str(join), crm_str(exp_state), do_terminate, this_node->details->shutdown);
1440  }
1441 
1442  return online;
1443 }
1444 
1445 static void
1446 determine_remote_online_status(pe_working_set_t * data_set, pe_node_t * this_node)
1447 {
1448  pe_resource_t *rsc = this_node->details->remote_rsc;
1449  pe_resource_t *container = NULL;
1450  pe_node_t *host = NULL;
1451 
1452  /* If there is a node state entry for a (former) Pacemaker Remote node
1453  * but no resource creating that node, the node's connection resource will
1454  * be NULL. Consider it an offline remote node in that case.
1455  */
1456  if (rsc == NULL) {
1457  this_node->details->online = FALSE;
1458  goto remote_online_done;
1459  }
1460 
1461  container = rsc->container;
1462 
1463  if (container && pcmk__list_of_1(rsc->running_on)) {
1464  host = rsc->running_on->data;
1465  }
1466 
1467  /* If the resource is currently started, mark it online. */
1468  if (rsc->role == RSC_ROLE_STARTED) {
1469  crm_trace("%s node %s presumed ONLINE because connection resource is started",
1470  (container? "Guest" : "Remote"), this_node->details->id);
1471  this_node->details->online = TRUE;
1472  }
1473 
1474  /* consider this node shutting down if transitioning start->stop */
1475  if (rsc->role == RSC_ROLE_STARTED && rsc->next_role == RSC_ROLE_STOPPED) {
1476  crm_trace("%s node %s shutting down because connection resource is stopping",
1477  (container? "Guest" : "Remote"), this_node->details->id);
1478  this_node->details->shutdown = TRUE;
1479  }
1480 
1481  /* Now check all the failure conditions. */
1482  if(container && pcmk_is_set(container->flags, pe_rsc_failed)) {
1483  crm_trace("Guest node %s UNCLEAN because guest resource failed",
1484  this_node->details->id);
1485  this_node->details->online = FALSE;
1486  this_node->details->remote_requires_reset = TRUE;
1487 
1488  } else if (pcmk_is_set(rsc->flags, pe_rsc_failed)) {
1489  crm_trace("%s node %s OFFLINE because connection resource failed",
1490  (container? "Guest" : "Remote"), this_node->details->id);
1491  this_node->details->online = FALSE;
1492 
1493  } else if (rsc->role == RSC_ROLE_STOPPED
1494  || (container && container->role == RSC_ROLE_STOPPED)) {
1495 
1496  crm_trace("%s node %s OFFLINE because its resource is stopped",
1497  (container? "Guest" : "Remote"), this_node->details->id);
1498  this_node->details->online = FALSE;
1499  this_node->details->remote_requires_reset = FALSE;
1500 
1501  } else if (host && (host->details->online == FALSE)
1502  && host->details->unclean) {
1503  crm_trace("Guest node %s UNCLEAN because host is unclean",
1504  this_node->details->id);
1505  this_node->details->online = FALSE;
1506  this_node->details->remote_requires_reset = TRUE;
1507  }
1508 
1509 remote_online_done:
1510  crm_trace("Remote node %s online=%s",
1511  this_node->details->id, this_node->details->online ? "TRUE" : "FALSE");
1512 }
1513 
1514 static void
1515 determine_online_status(xmlNode * node_state, pe_node_t * this_node, pe_working_set_t * data_set)
1516 {
1517  gboolean online = FALSE;
1518  const char *exp_state = crm_element_value(node_state, XML_NODE_EXPECTED);
1519 
1520  CRM_CHECK(this_node != NULL, return);
1521 
1522  this_node->details->shutdown = FALSE;
1523  this_node->details->expected_up = FALSE;
1524 
1525  if (pe__shutdown_requested(this_node)) {
1526  this_node->details->shutdown = TRUE;
1527 
1528  } else if (pcmk__str_eq(exp_state, CRMD_JOINSTATE_MEMBER, pcmk__str_casei)) {
1529  this_node->details->expected_up = TRUE;
1530  }
1531 
1532  if (this_node->details->type == node_ping) {
1533  this_node->details->unclean = FALSE;
1534  online = FALSE; /* As far as resource management is concerned,
1535  * the node is safely offline.
1536  * Anyone caught abusing this logic will be shot
1537  */
1538 
1539  } else if (!pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
1540  online = determine_online_status_no_fencing(data_set, node_state, this_node);
1541 
1542  } else {
1543  online = determine_online_status_fencing(data_set, node_state, this_node);
1544  }
1545 
1546  if (online) {
1547  this_node->details->online = TRUE;
1548 
1549  } else {
1550  /* remove node from contention */
1551  this_node->fixed = TRUE;
1552  this_node->weight = -INFINITY;
1553  }
1554 
1555  if (online && this_node->details->shutdown) {
1556  /* don't run resources here */
1557  this_node->fixed = TRUE;
1558  this_node->weight = -INFINITY;
1559  }
1560 
1561  if (this_node->details->type == node_ping) {
1562  crm_info("Node %s is not a Pacemaker node", this_node->details->uname);
1563 
1564  } else if (this_node->details->unclean) {
1565  pe_proc_warn("Node %s is unclean", this_node->details->uname);
1566 
1567  } else if (this_node->details->online) {
1568  crm_info("Node %s is %s", this_node->details->uname,
1569  this_node->details->shutdown ? "shutting down" :
1570  this_node->details->pending ? "pending" :
1571  this_node->details->standby ? "standby" :
1572  this_node->details->maintenance ? "maintenance" : "online");
1573 
1574  } else {
1575  crm_trace("Node %s is offline", this_node->details->uname);
1576  }
1577 }
1578 
1587 const char *
1588 pe_base_name_end(const char *id)
1589 {
1590  if (!pcmk__str_empty(id)) {
1591  const char *end = id + strlen(id) - 1;
1592 
1593  for (const char *s = end; s > id; --s) {
1594  switch (*s) {
1595  case '0':
1596  case '1':
1597  case '2':
1598  case '3':
1599  case '4':
1600  case '5':
1601  case '6':
1602  case '7':
1603  case '8':
1604  case '9':
1605  break;
1606  case ':':
1607  return (s == end)? s : (s - 1);
1608  default:
1609  return end;
1610  }
1611  }
1612  return end;
1613  }
1614  return NULL;
1615 }
1616 
1627 char *
1628 clone_strip(const char *last_rsc_id)
1629 {
1630  const char *end = pe_base_name_end(last_rsc_id);
1631  char *basename = NULL;
1632 
1633  CRM_ASSERT(end);
1634  basename = strndup(last_rsc_id, end - last_rsc_id + 1);
1635  CRM_ASSERT(basename);
1636  return basename;
1637 }
1638 
1649 char *
1650 clone_zero(const char *last_rsc_id)
1651 {
1652  const char *end = pe_base_name_end(last_rsc_id);
1653  size_t base_name_len = end - last_rsc_id + 1;
1654  char *zero = NULL;
1655 
1656  CRM_ASSERT(end);
1657  zero = calloc(base_name_len + 3, sizeof(char));
1658  CRM_ASSERT(zero);
1659  memcpy(zero, last_rsc_id, base_name_len);
1660  zero[base_name_len] = ':';
1661  zero[base_name_len + 1] = '0';
1662  return zero;
1663 }
1664 
1665 static pe_resource_t *
1666 create_fake_resource(const char *rsc_id, xmlNode * rsc_entry, pe_working_set_t * data_set)
1667 {
1668  pe_resource_t *rsc = NULL;
1669  xmlNode *xml_rsc = create_xml_node(NULL, XML_CIB_TAG_RESOURCE);
1670 
1671  copy_in_properties(xml_rsc, rsc_entry);
1672  crm_xml_add(xml_rsc, XML_ATTR_ID, rsc_id);
1673  crm_log_xml_debug(xml_rsc, "Orphan resource");
1674 
1675  if (!common_unpack(xml_rsc, &rsc, NULL, data_set)) {
1676  return NULL;
1677  }
1678 
1679  if (xml_contains_remote_node(xml_rsc)) {
1680  pe_node_t *node;
1681 
1682  crm_debug("Detected orphaned remote node %s", rsc_id);
1683  node = pe_find_node(data_set->nodes, rsc_id);
1684  if (node == NULL) {
1685  node = pe_create_node(rsc_id, rsc_id, "remote", NULL, data_set);
1686  }
1687  link_rsc2remotenode(data_set, rsc);
1688 
1689  if (node) {
1690  crm_trace("Setting node %s as shutting down due to orphaned connection resource", rsc_id);
1691  node->details->shutdown = TRUE;
1692  }
1693  }
1694 
1695  if (crm_element_value(rsc_entry, XML_RSC_ATTR_CONTAINER)) {
1696  /* This orphaned rsc needs to be mapped to a container. */
1697  crm_trace("Detected orphaned container filler %s", rsc_id);
1699  }
1701  data_set->resources = g_list_append(data_set->resources, rsc);
1702  return rsc;
1703 }
1704 
1709 static pe_resource_t *
1710 create_anonymous_orphan(pe_resource_t *parent, const char *rsc_id,
1711  pe_node_t *node, pe_working_set_t *data_set)
1712 {
1713  pe_resource_t *top = pe__create_clone_child(parent, data_set);
1714 
1715  // find_rsc() because we might be a cloned group
1716  pe_resource_t *orphan = top->fns->find_rsc(top, rsc_id, NULL, pe_find_clone);
1717 
1718  pe_rsc_debug(parent, "Created orphan %s for %s: %s on %s",
1719  top->id, parent->id, rsc_id, node->details->uname);
1720  return orphan;
1721 }
1722 
1737 static pe_resource_t *
1738 find_anonymous_clone(pe_working_set_t * data_set, pe_node_t * node, pe_resource_t * parent,
1739  const char *rsc_id)
1740 {
1741  GList *rIter = NULL;
1742  pe_resource_t *rsc = NULL;
1743  pe_resource_t *inactive_instance = NULL;
1744  gboolean skip_inactive = FALSE;
1745 
1746  CRM_ASSERT(parent != NULL);
1747  CRM_ASSERT(pe_rsc_is_clone(parent));
1749 
1750  // Check for active (or partially active, for cloned groups) instance
1751  pe_rsc_trace(parent, "Looking for %s on %s in %s", rsc_id, node->details->uname, parent->id);
1752  for (rIter = parent->children; rsc == NULL && rIter; rIter = rIter->next) {
1753  GList *locations = NULL;
1754  pe_resource_t *child = rIter->data;
1755 
1756  /* Check whether this instance is already known to be active or pending
1757  * anywhere, at this stage of unpacking. Because this function is called
1758  * for a resource before the resource's individual operation history
1759  * entries are unpacked, locations will generally not contain the
1760  * desired node.
1761  *
1762  * However, there are three exceptions:
1763  * (1) when child is a cloned group and we have already unpacked the
1764  * history of another member of the group on the same node;
1765  * (2) when we've already unpacked the history of another numbered
1766  * instance on the same node (which can happen if globally-unique
1767  * was flipped from true to false); and
1768  * (3) when we re-run calculations on the same data set as part of a
1769  * simulation.
1770  */
1771  child->fns->location(child, &locations, 2);
1772  if (locations) {
1773  /* We should never associate the same numbered anonymous clone
1774  * instance with multiple nodes, and clone instances can't migrate,
1775  * so there must be only one location, regardless of history.
1776  */
1777  CRM_LOG_ASSERT(locations->next == NULL);
1778 
1779  if (((pe_node_t *)locations->data)->details == node->details) {
1780  /* This child instance is active on the requested node, so check
1781  * for a corresponding configured resource. We use find_rsc()
1782  * instead of child because child may be a cloned group, and we
1783  * need the particular member corresponding to rsc_id.
1784  *
1785  * If the history entry is orphaned, rsc will be NULL.
1786  */
1787  rsc = parent->fns->find_rsc(child, rsc_id, NULL, pe_find_clone);
1788  if (rsc) {
1789  /* If there are multiple instance history entries for an
1790  * anonymous clone in a single node's history (which can
1791  * happen if globally-unique is switched from true to
1792  * false), we want to consider the instances beyond the
1793  * first as orphans, even if there are inactive instance
1794  * numbers available.
1795  */
1796  if (rsc->running_on) {
1797  crm_notice("Active (now-)anonymous clone %s has "
1798  "multiple (orphan) instance histories on %s",
1799  parent->id, node->details->uname);
1800  skip_inactive = TRUE;
1801  rsc = NULL;
1802  } else {
1803  pe_rsc_trace(parent, "Resource %s, active", rsc->id);
1804  }
1805  }
1806  }
1807  g_list_free(locations);
1808 
1809  } else {
1810  pe_rsc_trace(parent, "Resource %s, skip inactive", child->id);
1811  if (!skip_inactive && !inactive_instance
1812  && !pcmk_is_set(child->flags, pe_rsc_block)) {
1813  // Remember one inactive instance in case we don't find active
1814  inactive_instance = parent->fns->find_rsc(child, rsc_id, NULL,
1815  pe_find_clone);
1816 
1817  /* ... but don't use it if it was already associated with a
1818  * pending action on another node
1819  */
1820  if (inactive_instance && inactive_instance->pending_node
1821  && (inactive_instance->pending_node->details != node->details)) {
1822  inactive_instance = NULL;
1823  }
1824  }
1825  }
1826  }
1827 
1828  if ((rsc == NULL) && !skip_inactive && (inactive_instance != NULL)) {
1829  pe_rsc_trace(parent, "Resource %s, empty slot", inactive_instance->id);
1830  rsc = inactive_instance;
1831  }
1832 
1833  /* If the resource has "requires" set to "quorum" or "nothing", and we don't
1834  * have a clone instance for every node, we don't want to consume a valid
1835  * instance number for unclean nodes. Such instances may appear to be active
1836  * according to the history, but should be considered inactive, so we can
1837  * start an instance elsewhere. Treat such instances as orphans.
1838  *
1839  * An exception is instances running on guest nodes -- since guest node
1840  * "fencing" is actually just a resource stop, requires shouldn't apply.
1841  *
1842  * @TODO Ideally, we'd use an inactive instance number if it is not needed
1843  * for any clean instances. However, we don't know that at this point.
1844  */
1845  if ((rsc != NULL) && !pcmk_is_set(rsc->flags, pe_rsc_needs_fencing)
1846  && (!node->details->online || node->details->unclean)
1847  && !pe__is_guest_node(node)
1848  && !pe__is_universal_clone(parent, data_set)) {
1849 
1850  rsc = NULL;
1851  }
1852 
1853  if (rsc == NULL) {
1854  rsc = create_anonymous_orphan(parent, rsc_id, node, data_set);
1855  pe_rsc_trace(parent, "Resource %s, orphan", rsc->id);
1856  }
1857  return rsc;
1858 }
1859 
1860 static pe_resource_t *
1861 unpack_find_resource(pe_working_set_t * data_set, pe_node_t * node, const char *rsc_id,
1862  xmlNode * rsc_entry)
1863 {
1864  pe_resource_t *rsc = NULL;
1865  pe_resource_t *parent = NULL;
1866 
1867  crm_trace("looking for %s", rsc_id);
1868  rsc = pe_find_resource(data_set->resources, rsc_id);
1869 
1870  if (rsc == NULL) {
1871  /* If we didn't find the resource by its name in the operation history,
1872  * check it again as a clone instance. Even when clone-max=0, we create
1873  * a single :0 orphan to match against here.
1874  */
1875  char *clone0_id = clone_zero(rsc_id);
1876  pe_resource_t *clone0 = pe_find_resource(data_set->resources, clone0_id);
1877 
1878  if (clone0 && !pcmk_is_set(clone0->flags, pe_rsc_unique)) {
1879  rsc = clone0;
1880  parent = uber_parent(clone0);
1881  crm_trace("%s found as %s (%s)", rsc_id, clone0_id, parent->id);
1882  } else {
1883  crm_trace("%s is not known as %s either (orphan)",
1884  rsc_id, clone0_id);
1885  }
1886  free(clone0_id);
1887 
1888  } else if (rsc->variant > pe_native) {
1889  crm_trace("Resource history for %s is orphaned because it is no longer primitive",
1890  rsc_id);
1891  return NULL;
1892 
1893  } else {
1894  parent = uber_parent(rsc);
1895  }
1896 
1897  if (pe_rsc_is_anon_clone(parent)) {
1898 
1899  if (pe_rsc_is_bundled(parent)) {
1900  rsc = pe__find_bundle_replica(parent->parent, node);
1901  } else {
1902  char *base = clone_strip(rsc_id);
1903 
1904  rsc = find_anonymous_clone(data_set, node, parent, base);
1905  free(base);
1906  CRM_ASSERT(rsc != NULL);
1907  }
1908  }
1909 
1910  if (rsc && !pcmk__str_eq(rsc_id, rsc->id, pcmk__str_casei)
1911  && !pcmk__str_eq(rsc_id, rsc->clone_name, pcmk__str_casei)) {
1912 
1913  free(rsc->clone_name);
1914  rsc->clone_name = strdup(rsc_id);
1915  pe_rsc_debug(rsc, "Internally renamed %s on %s to %s%s",
1916  rsc_id, node->details->uname, rsc->id,
1917  (pcmk_is_set(rsc->flags, pe_rsc_orphan)? " (ORPHAN)" : ""));
1918  }
1919  return rsc;
1920 }
1921 
1922 static pe_resource_t *
1923 process_orphan_resource(xmlNode * rsc_entry, pe_node_t * node, pe_working_set_t * data_set)
1924 {
1925  pe_resource_t *rsc = NULL;
1926  const char *rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID);
1927 
1928  crm_debug("Detected orphan resource %s on %s", rsc_id, node->details->uname);
1929  rsc = create_fake_resource(rsc_id, rsc_entry, data_set);
1930  if (rsc == NULL) {
1931  return NULL;
1932  }
1933 
1934  if (!pcmk_is_set(data_set->flags, pe_flag_stop_rsc_orphans)) {
1936 
1937  } else {
1938  CRM_CHECK(rsc != NULL, return NULL);
1939  pe_rsc_trace(rsc, "Added orphan %s", rsc->id);
1940  resource_location(rsc, NULL, -INFINITY, "__orphan_do_not_run__", data_set);
1941  }
1942  return rsc;
1943 }
1944 
1945 static void
1946 process_rsc_state(pe_resource_t * rsc, pe_node_t * node,
1947  enum action_fail_response on_fail,
1948  xmlNode * migrate_op, pe_working_set_t * data_set)
1949 {
1950  pe_node_t *tmpnode = NULL;
1951  char *reason = NULL;
1952  enum action_fail_response save_on_fail = action_fail_ignore;
1953 
1954  CRM_ASSERT(rsc);
1955  pe_rsc_trace(rsc, "Resource %s is %s on %s: on_fail=%s",
1956  rsc->id, role2text(rsc->role), node->details->uname, fail2text(on_fail));
1957 
1958  /* process current state */
1959  if (rsc->role != RSC_ROLE_UNKNOWN) {
1960  pe_resource_t *iter = rsc;
1961 
1962  while (iter) {
1963  if (g_hash_table_lookup(iter->known_on, node->details->id) == NULL) {
1964  pe_node_t *n = pe__copy_node(node);
1965 
1966  pe_rsc_trace(rsc, "%s (aka. %s) known on %s", rsc->id, rsc->clone_name,
1967  n->details->uname);
1968  g_hash_table_insert(iter->known_on, (gpointer) n->details->id, n);
1969  }
1970  if (pcmk_is_set(iter->flags, pe_rsc_unique)) {
1971  break;
1972  }
1973  iter = iter->parent;
1974  }
1975  }
1976 
1977  /* If a managed resource is believed to be running, but node is down ... */
1978  if (rsc->role > RSC_ROLE_STOPPED
1979  && node->details->online == FALSE
1980  && node->details->maintenance == FALSE
1981  && pcmk_is_set(rsc->flags, pe_rsc_managed)) {
1982 
1983  gboolean should_fence = FALSE;
1984 
1985  /* If this is a guest node, fence it (regardless of whether fencing is
1986  * enabled, because guest node fencing is done by recovery of the
1987  * container resource rather than by the fencer). Mark the resource
1988  * we're processing as failed. When the guest comes back up, its
1989  * operation history in the CIB will be cleared, freeing the affected
1990  * resource to run again once we are sure we know its state.
1991  */
1992  if (pe__is_guest_node(node)) {
1994  should_fence = TRUE;
1995 
1996  } else if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
1997  if (pe__is_remote_node(node) && node->details->remote_rsc
1999 
2000  /* Setting unseen means that fencing of the remote node will
2001  * occur only if the connection resource is not going to start
2002  * somewhere. This allows connection resources on a failed
2003  * cluster node to move to another node without requiring the
2004  * remote nodes to be fenced as well.
2005  */
2006  node->details->unseen = TRUE;
2007  reason = crm_strdup_printf("%s is active there (fencing will be"
2008  " revoked if remote connection can "
2009  "be re-established elsewhere)",
2010  rsc->id);
2011  }
2012  should_fence = TRUE;
2013  }
2014 
2015  if (should_fence) {
2016  if (reason == NULL) {
2017  reason = crm_strdup_printf("%s is thought to be active there", rsc->id);
2018  }
2019  pe_fence_node(data_set, node, reason, FALSE);
2020  }
2021  free(reason);
2022  }
2023 
2024  /* In order to calculate priority_fencing_delay correctly, save the failure information and pass it to native_add_running(). */
2025  save_on_fail = on_fail;
2026 
2027  if (node->details->unclean) {
2028  /* No extra processing needed
2029  * Also allows resources to be started again after a node is shot
2030  */
2031  on_fail = action_fail_ignore;
2032  }
2033 
2034  switch (on_fail) {
2035  case action_fail_ignore:
2036  /* nothing to do */
2037  break;
2038 
2039  case action_fail_demote:
2041  demote_action(rsc, node, FALSE);
2042  break;
2043 
2044  case action_fail_fence:
2045  /* treat it as if it is still running
2046  * but also mark the node as unclean
2047  */
2048  reason = crm_strdup_printf("%s failed there", rsc->id);
2049  pe_fence_node(data_set, node, reason, FALSE);
2050  free(reason);
2051  break;
2052 
2053  case action_fail_standby:
2054  node->details->standby = TRUE;
2055  node->details->standby_onfail = TRUE;
2056  break;
2057 
2058  case action_fail_block:
2059  /* is_managed == FALSE will prevent any
2060  * actions being sent for the resource
2061  */
2064  break;
2065 
2066  case action_fail_migrate:
2067  /* make sure it comes up somewhere else
2068  * or not at all
2069  */
2070  resource_location(rsc, node, -INFINITY, "__action_migration_auto__", data_set);
2071  break;
2072 
2073  case action_fail_stop:
2074  pe__set_next_role(rsc, RSC_ROLE_STOPPED, "on-fail=stop");
2075  break;
2076 
2077  case action_fail_recover:
2078  if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) {
2080  stop_action(rsc, node, FALSE);
2081  }
2082  break;
2083 
2086  if (rsc->container && pe_rsc_is_bundled(rsc)) {
2087  /* A bundle's remote connection can run on a different node than
2088  * the bundle's container. We don't necessarily know where the
2089  * container is running yet, so remember it and add a stop
2090  * action for it later.
2091  */
2092  data_set->stop_needed = g_list_prepend(data_set->stop_needed,
2093  rsc->container);
2094  } else if (rsc->container) {
2095  stop_action(rsc->container, node, FALSE);
2096  } else if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) {
2097  stop_action(rsc, node, FALSE);
2098  }
2099  break;
2100 
2103  if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
2104  tmpnode = NULL;
2105  if (rsc->is_remote_node) {
2106  tmpnode = pe_find_node(data_set->nodes, rsc->id);
2107  }
2108  if (tmpnode &&
2109  pe__is_remote_node(tmpnode) &&
2110  tmpnode->details->remote_was_fenced == 0) {
2111 
2112  /* The remote connection resource failed in a way that
2113  * should result in fencing the remote node.
2114  */
2115  pe_fence_node(data_set, tmpnode,
2116  "remote connection is unrecoverable", FALSE);
2117  }
2118  }
2119 
2120  /* require the stop action regardless if fencing is occurring or not. */
2121  if (rsc->role > RSC_ROLE_STOPPED) {
2122  stop_action(rsc, node, FALSE);
2123  }
2124 
2125  /* if reconnect delay is in use, prevent the connection from exiting the
2126  * "STOPPED" role until the failure is cleared by the delay timeout. */
2127  if (rsc->remote_reconnect_ms) {
2128  pe__set_next_role(rsc, RSC_ROLE_STOPPED, "remote reset");
2129  }
2130  break;
2131  }
2132 
2133  /* ensure a remote-node connection failure forces an unclean remote-node
2134  * to be fenced. By setting unseen = FALSE, the remote-node failure will
2135  * result in a fencing operation regardless if we're going to attempt to
2136  * reconnect to the remote-node in this transition or not. */
2137  if (pcmk_is_set(rsc->flags, pe_rsc_failed) && rsc->is_remote_node) {
2138  tmpnode = pe_find_node(data_set->nodes, rsc->id);
2139  if (tmpnode && tmpnode->details->unclean) {
2140  tmpnode->details->unseen = FALSE;
2141  }
2142  }
2143 
2144  if (rsc->role != RSC_ROLE_STOPPED && rsc->role != RSC_ROLE_UNKNOWN) {
2145  if (pcmk_is_set(rsc->flags, pe_rsc_orphan)) {
2146  if (pcmk_is_set(rsc->flags, pe_rsc_managed)) {
2147  pcmk__config_warn("Detected active orphan %s running on %s",
2148  rsc->id, node->details->uname);
2149  } else {
2150  pcmk__config_warn("Resource '%s' must be stopped manually on "
2151  "%s because cluster is configured not to "
2152  "stop active orphans",
2153  rsc->id, node->details->uname);
2154  }
2155  }
2156 
2157  native_add_running(rsc, node, data_set, (save_on_fail != action_fail_ignore));
2158  switch (on_fail) {
2159  case action_fail_ignore:
2160  break;
2161  case action_fail_demote:
2162  case action_fail_block:
2164  break;
2165  default:
2167  break;
2168  }
2169 
2170  } else if (rsc->clone_name && strchr(rsc->clone_name, ':') != NULL) {
2171  /* Only do this for older status sections that included instance numbers
2172  * Otherwise stopped instances will appear as orphans
2173  */
2174  pe_rsc_trace(rsc, "Resetting clone_name %s for %s (stopped)", rsc->clone_name, rsc->id);
2175  free(rsc->clone_name);
2176  rsc->clone_name = NULL;
2177 
2178  } else {
2179  GList *possible_matches = pe__resource_actions(rsc, node, RSC_STOP,
2180  FALSE);
2181  GList *gIter = possible_matches;
2182 
2183  for (; gIter != NULL; gIter = gIter->next) {
2184  pe_action_t *stop = (pe_action_t *) gIter->data;
2185 
2187  }
2188 
2189  g_list_free(possible_matches);
2190  }
2191 }
2192 
2193 /* create active recurring operations as optional */
2194 static void
2195 process_recurring(pe_node_t * node, pe_resource_t * rsc,
2196  int start_index, int stop_index,
2197  GList *sorted_op_list, pe_working_set_t * data_set)
2198 {
2199  int counter = -1;
2200  const char *task = NULL;
2201  const char *status = NULL;
2202  GList *gIter = sorted_op_list;
2203 
2204  CRM_ASSERT(rsc);
2205  pe_rsc_trace(rsc, "%s: Start index %d, stop index = %d", rsc->id, start_index, stop_index);
2206 
2207  for (; gIter != NULL; gIter = gIter->next) {
2208  xmlNode *rsc_op = (xmlNode *) gIter->data;
2209 
2210  guint interval_ms = 0;
2211  char *key = NULL;
2212  const char *id = ID(rsc_op);
2213 
2214  counter++;
2215 
2216  if (node->details->online == FALSE) {
2217  pe_rsc_trace(rsc, "Skipping %s/%s: node is offline", rsc->id, node->details->uname);
2218  break;
2219 
2220  /* Need to check if there's a monitor for role="Stopped" */
2221  } else if (start_index < stop_index && counter <= stop_index) {
2222  pe_rsc_trace(rsc, "Skipping %s/%s: resource is not active", id, node->details->uname);
2223  continue;
2224 
2225  } else if (counter < start_index) {
2226  pe_rsc_trace(rsc, "Skipping %s/%s: old %d", id, node->details->uname, counter);
2227  continue;
2228  }
2229 
2230  crm_element_value_ms(rsc_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms);
2231  if (interval_ms == 0) {
2232  pe_rsc_trace(rsc, "Skipping %s/%s: non-recurring", id, node->details->uname);
2233  continue;
2234  }
2235 
2236  status = crm_element_value(rsc_op, XML_LRM_ATTR_OPSTATUS);
2237  if (pcmk__str_eq(status, "-1", pcmk__str_casei)) {
2238  pe_rsc_trace(rsc, "Skipping %s/%s: status", id, node->details->uname);
2239  continue;
2240  }
2241  task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK);
2242  /* create the action */
2243  key = pcmk__op_key(rsc->id, task, interval_ms);
2244  pe_rsc_trace(rsc, "Creating %s/%s", key, node->details->uname);
2245  custom_action(rsc, key, task, node, TRUE, TRUE, data_set);
2246  }
2247 }
2248 
2249 void
2250 calculate_active_ops(GList *sorted_op_list, int *start_index, int *stop_index)
2251 {
2252  int counter = -1;
2253  int implied_monitor_start = -1;
2254  int implied_clone_start = -1;
2255  const char *task = NULL;
2256  const char *status = NULL;
2257  GList *gIter = sorted_op_list;
2258 
2259  *stop_index = -1;
2260  *start_index = -1;
2261 
2262  for (; gIter != NULL; gIter = gIter->next) {
2263  xmlNode *rsc_op = (xmlNode *) gIter->data;
2264 
2265  counter++;
2266 
2267  task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK);
2268  status = crm_element_value(rsc_op, XML_LRM_ATTR_OPSTATUS);
2269 
2270  if (pcmk__str_eq(task, CRMD_ACTION_STOP, pcmk__str_casei)
2271  && pcmk__str_eq(status, "0", pcmk__str_casei)) {
2272  *stop_index = counter;
2273 
2274  } else if (pcmk__strcase_any_of(task, CRMD_ACTION_START, CRMD_ACTION_MIGRATED, NULL)) {
2275  *start_index = counter;
2276 
2277  } else if ((implied_monitor_start <= *stop_index) && pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_casei)) {
2278  const char *rc = crm_element_value(rsc_op, XML_LRM_ATTR_RC);
2279 
2280  if (pcmk__strcase_any_of(rc, "0", "8", NULL)) {
2281  implied_monitor_start = counter;
2282  }
2283  } else if (pcmk__strcase_any_of(task, CRMD_ACTION_PROMOTE, CRMD_ACTION_DEMOTE, NULL)) {
2284  implied_clone_start = counter;
2285  }
2286  }
2287 
2288  if (*start_index == -1) {
2289  if (implied_clone_start != -1) {
2290  *start_index = implied_clone_start;
2291  } else if (implied_monitor_start != -1) {
2292  *start_index = implied_monitor_start;
2293  }
2294  }
2295 }
2296 
2297 // If resource history entry has shutdown lock, remember lock node and time
2298 static void
2299 unpack_shutdown_lock(xmlNode *rsc_entry, pe_resource_t *rsc, pe_node_t *node,
2300  pe_working_set_t *data_set)
2301 {
2302  time_t lock_time = 0; // When lock started (i.e. node shutdown time)
2303 
2305  &lock_time) == pcmk_ok) && (lock_time != 0)) {
2306 
2307  if ((data_set->shutdown_lock > 0)
2308  && (get_effective_time(data_set)
2309  > (lock_time + data_set->shutdown_lock))) {
2310  pe_rsc_info(rsc, "Shutdown lock for %s on %s expired",
2311  rsc->id, node->details->uname);
2312  pe__clear_resource_history(rsc, node, data_set);
2313  } else {
2314  rsc->lock_node = node;
2315  rsc->lock_time = lock_time;
2316  }
2317  }
2318 }
2319 
2330 static pe_resource_t *
2331 unpack_lrm_resource(pe_node_t *node, xmlNode *lrm_resource,
2332  pe_working_set_t *data_set)
2333 {
2334  GList *gIter = NULL;
2335  int stop_index = -1;
2336  int start_index = -1;
2337  enum rsc_role_e req_role = RSC_ROLE_UNKNOWN;
2338 
2339  const char *task = NULL;
2340  const char *rsc_id = ID(lrm_resource);
2341 
2342  pe_resource_t *rsc = NULL;
2343  GList *op_list = NULL;
2344  GList *sorted_op_list = NULL;
2345 
2346  xmlNode *migrate_op = NULL;
2347  xmlNode *rsc_op = NULL;
2348  xmlNode *last_failure = NULL;
2349 
2351  enum rsc_role_e saved_role = RSC_ROLE_UNKNOWN;
2352 
2353  if (rsc_id == NULL) {
2354  crm_warn("Ignoring malformed " XML_LRM_TAG_RESOURCE
2355  " entry without id");
2356  return NULL;
2357  }
2358  crm_trace("Unpacking " XML_LRM_TAG_RESOURCE " for %s on %s",
2359  rsc_id, node->details->uname);
2360 
2361  // Build a list of individual lrm_rsc_op entries, so we can sort them
2362  for (rsc_op = first_named_child(lrm_resource, XML_LRM_TAG_RSC_OP);
2363  rsc_op != NULL; rsc_op = crm_next_same_xml(rsc_op)) {
2364 
2365  op_list = g_list_prepend(op_list, rsc_op);
2366  }
2367 
2368  if (!pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)) {
2369  if (op_list == NULL) {
2370  // If there are no operations, there is nothing to do
2371  return NULL;
2372  }
2373  }
2374 
2375  /* find the resource */
2376  rsc = unpack_find_resource(data_set, node, rsc_id, lrm_resource);
2377  if (rsc == NULL) {
2378  if (op_list == NULL) {
2379  // If there are no operations, there is nothing to do
2380  return NULL;
2381  } else {
2382  rsc = process_orphan_resource(lrm_resource, node, data_set);
2383  }
2384  }
2385  CRM_ASSERT(rsc != NULL);
2386 
2387  // Check whether the resource is "shutdown-locked" to this node
2388  if (pcmk_is_set(data_set->flags, pe_flag_shutdown_lock)) {
2389  unpack_shutdown_lock(lrm_resource, rsc, node, data_set);
2390  }
2391 
2392  /* process operations */
2393  saved_role = rsc->role;
2394  rsc->role = RSC_ROLE_UNKNOWN;
2395  sorted_op_list = g_list_sort(op_list, sort_op_by_callid);
2396 
2397  for (gIter = sorted_op_list; gIter != NULL; gIter = gIter->next) {
2398  xmlNode *rsc_op = (xmlNode *) gIter->data;
2399 
2400  task = crm_element_value(rsc_op, XML_LRM_ATTR_TASK);
2401  if (pcmk__str_eq(task, CRMD_ACTION_MIGRATED, pcmk__str_casei)) {
2402  migrate_op = rsc_op;
2403  }
2404 
2405  unpack_rsc_op(rsc, node, rsc_op, &last_failure, &on_fail, data_set);
2406  }
2407 
2408  /* create active recurring operations as optional */
2409  calculate_active_ops(sorted_op_list, &start_index, &stop_index);
2410  process_recurring(node, rsc, start_index, stop_index, sorted_op_list, data_set);
2411 
2412  /* no need to free the contents */
2413  g_list_free(sorted_op_list);
2414 
2415  process_rsc_state(rsc, node, on_fail, migrate_op, data_set);
2416 
2417  if (get_target_role(rsc, &req_role)) {
2418  if (rsc->next_role == RSC_ROLE_UNKNOWN || req_role < rsc->next_role) {
2420 
2421  } else if (req_role > rsc->next_role) {
2422  pe_rsc_info(rsc, "%s: Not overwriting calculated next role %s"
2423  " with requested next role %s",
2424  rsc->id, role2text(rsc->next_role), role2text(req_role));
2425  }
2426  }
2427 
2428  if (saved_role > rsc->role) {
2429  rsc->role = saved_role;
2430  }
2431 
2432  return rsc;
2433 }
2434 
2435 static void
2436 handle_orphaned_container_fillers(xmlNode * lrm_rsc_list, pe_working_set_t * data_set)
2437 {
2438  xmlNode *rsc_entry = NULL;
2439  for (rsc_entry = pcmk__xe_first_child(lrm_rsc_list); rsc_entry != NULL;
2440  rsc_entry = pcmk__xe_next(rsc_entry)) {
2441 
2442  pe_resource_t *rsc;
2443  pe_resource_t *container;
2444  const char *rsc_id;
2445  const char *container_id;
2446 
2447  if (!pcmk__str_eq((const char *)rsc_entry->name, XML_LRM_TAG_RESOURCE, pcmk__str_casei)) {
2448  continue;
2449  }
2450 
2451  container_id = crm_element_value(rsc_entry, XML_RSC_ATTR_CONTAINER);
2452  rsc_id = crm_element_value(rsc_entry, XML_ATTR_ID);
2453  if (container_id == NULL || rsc_id == NULL) {
2454  continue;
2455  }
2456 
2457  container = pe_find_resource(data_set->resources, container_id);
2458  if (container == NULL) {
2459  continue;
2460  }
2461 
2462  rsc = pe_find_resource(data_set->resources, rsc_id);
2463  if (rsc == NULL ||
2465  rsc->container != NULL) {
2466  continue;
2467  }
2468 
2469  pe_rsc_trace(rsc, "Mapped container of orphaned resource %s to %s",
2470  rsc->id, container_id);
2471  rsc->container = container;
2472  container->fillers = g_list_append(container->fillers, rsc);
2473  }
2474 }
2475 
2484 static void
2485 unpack_node_lrm(pe_node_t *node, xmlNode *xml, pe_working_set_t *data_set)
2486 {
2487  bool found_orphaned_container_filler = false;
2488 
2489  // Drill down to lrm_resources section
2490  xml = find_xml_node(xml, XML_CIB_TAG_LRM, FALSE);
2491  if (xml == NULL) {
2492  return;
2493  }
2494  xml = find_xml_node(xml, XML_LRM_TAG_RESOURCES, FALSE);
2495  if (xml == NULL) {
2496  return;
2497  }
2498 
2499  // Unpack each lrm_resource entry
2500  for (xmlNode *rsc_entry = first_named_child(xml, XML_LRM_TAG_RESOURCE);
2501  rsc_entry != NULL; rsc_entry = crm_next_same_xml(rsc_entry)) {
2502 
2503  pe_resource_t *rsc = unpack_lrm_resource(node, rsc_entry, data_set);
2504 
2505  if ((rsc != NULL)
2507  found_orphaned_container_filler = true;
2508  }
2509  }
2510 
2511  /* Now that all resource state has been unpacked for this node, map any
2512  * orphaned container fillers to their container resource.
2513  */
2514  if (found_orphaned_container_filler) {
2515  handle_orphaned_container_fillers(xml, data_set);
2516  }
2517 }
2518 
2519 static void
2520 set_active(pe_resource_t * rsc)
2521 {
2522  pe_resource_t *top = uber_parent(rsc);
2523 
2524  if (top && pcmk_is_set(top->flags, pe_rsc_promotable)) {
2525  rsc->role = RSC_ROLE_UNPROMOTED;
2526  } else {
2527  rsc->role = RSC_ROLE_STARTED;
2528  }
2529 }
2530 
2531 static void
2532 set_node_score(gpointer key, gpointer value, gpointer user_data)
2533 {
2534  pe_node_t *node = value;
2535  int *score = user_data;
2536 
2537  node->weight = *score;
2538 }
2539 
2540 #define STATUS_PATH_MAX 1024
2541 static xmlNode *
2542 find_lrm_op(const char *resource, const char *op, const char *node, const char *source,
2543  bool success_only, pe_working_set_t *data_set)
2544 {
2545  int offset = 0;
2546  char xpath[STATUS_PATH_MAX];
2547  xmlNode *xml = NULL;
2548 
2549  offset += snprintf(xpath + offset, STATUS_PATH_MAX - offset, "//node_state[@uname='%s']", node);
2550  offset +=
2551  snprintf(xpath + offset, STATUS_PATH_MAX - offset, "//" XML_LRM_TAG_RESOURCE "[@id='%s']",
2552  resource);
2553 
2554  /* Need to check against transition_magic too? */
2555  if (source && pcmk__str_eq(op, CRMD_ACTION_MIGRATE, pcmk__str_casei)) {
2556  offset +=
2557  snprintf(xpath + offset, STATUS_PATH_MAX - offset,
2558  "/" XML_LRM_TAG_RSC_OP "[@operation='%s' and @migrate_target='%s']", op,
2559  source);
2560  } else if (source && pcmk__str_eq(op, CRMD_ACTION_MIGRATED, pcmk__str_casei)) {
2561  offset +=
2562  snprintf(xpath + offset, STATUS_PATH_MAX - offset,
2563  "/" XML_LRM_TAG_RSC_OP "[@operation='%s' and @migrate_source='%s']", op,
2564  source);
2565  } else {
2566  offset +=
2567  snprintf(xpath + offset, STATUS_PATH_MAX - offset,
2568  "/" XML_LRM_TAG_RSC_OP "[@operation='%s']", op);
2569  }
2570 
2571  CRM_LOG_ASSERT(offset > 0);
2572  xml = get_xpath_object(xpath, data_set->input, LOG_DEBUG);
2573 
2574  if (xml && success_only) {
2575  int rc = PCMK_OCF_UNKNOWN_ERROR;
2576  int status = PCMK_EXEC_ERROR;
2577 
2580  if ((rc != PCMK_OCF_OK) || (status != PCMK_EXEC_DONE)) {
2581  return NULL;
2582  }
2583  }
2584  return xml;
2585 }
2586 
2587 static int
2588 pe__call_id(xmlNode *op_xml)
2589 {
2590  int id = 0;
2591 
2592  if (op_xml) {
2594  }
2595  return id;
2596 }
2597 
2614 static bool
2615 stop_happened_after(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op,
2616  pe_working_set_t *data_set)
2617 {
2618  xmlNode *stop_op = find_lrm_op(rsc->id, CRMD_ACTION_STOP,
2619  node->details->uname, NULL, TRUE, data_set);
2620 
2621  return (stop_op && (pe__call_id(stop_op) > pe__call_id(xml_op)));
2622 }
2623 
2624 static void
2625 unpack_migrate_to_success(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op,
2626  pe_working_set_t *data_set)
2627 {
2628  /* A successful migration sequence is:
2629  * migrate_to on source node
2630  * migrate_from on target node
2631  * stop on source node
2632  *
2633  * If a migrate_to is followed by a stop, the entire migration (successful
2634  * or failed) is complete, and we don't care what happened on the target.
2635  *
2636  * If no migrate_from has happened, the migration is considered to be
2637  * "partial". If the migrate_from failed, make sure the resource gets
2638  * stopped on both source and target (if up).
2639  *
2640  * If the migrate_to and migrate_from both succeeded (which also implies the
2641  * resource is no longer running on the source), but there is no stop, the
2642  * migration is considered to be "dangling". Schedule a stop on the source
2643  * in this case.
2644  */
2645  int from_rc = 0;
2646  int from_status = 0;
2647  pe_node_t *target_node = NULL;
2648  pe_node_t *source_node = NULL;
2649  xmlNode *migrate_from = NULL;
2650  const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE);
2651  const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
2652 
2653  // Sanity check
2654  CRM_CHECK(source && target && !strcmp(source, node->details->uname), return);
2655 
2656  if (stop_happened_after(rsc, node, xml_op, data_set)) {
2657  return;
2658  }
2659 
2660  // Clones are not allowed to migrate, so role can't be promoted
2661  rsc->role = RSC_ROLE_STARTED;
2662 
2663  target_node = pe_find_node(data_set->nodes, target);
2664  source_node = pe_find_node(data_set->nodes, source);
2665 
2666  // Check whether there was a migrate_from action on the target
2667  migrate_from = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATED, target,
2668  source, FALSE, data_set);
2669  if (migrate_from) {
2670  crm_element_value_int(migrate_from, XML_LRM_ATTR_RC, &from_rc);
2671  crm_element_value_int(migrate_from, XML_LRM_ATTR_OPSTATUS, &from_status);
2672  pe_rsc_trace(rsc, "%s op on %s exited with status=%d, rc=%d",
2673  ID(migrate_from), target, from_status, from_rc);
2674  }
2675 
2676  if (migrate_from && from_rc == PCMK_OCF_OK
2677  && (from_status == PCMK_EXEC_DONE)) {
2678  /* The migrate_to and migrate_from both succeeded, so mark the migration
2679  * as "dangling". This will be used to schedule a stop action on the
2680  * source without affecting the target.
2681  */
2682  pe_rsc_trace(rsc, "Detected dangling migration op: %s on %s", ID(xml_op),
2683  source);
2684  rsc->role = RSC_ROLE_STOPPED;
2685  rsc->dangling_migrations = g_list_prepend(rsc->dangling_migrations, node);
2686 
2687  } else if (migrate_from && (from_status != PCMK_EXEC_PENDING)) { // Failed
2688  if (target_node && target_node->details->online) {
2689  pe_rsc_trace(rsc, "Marking active on %s %p %d", target, target_node,
2690  target_node->details->online);
2691  native_add_running(rsc, target_node, data_set, TRUE);
2692  }
2693 
2694  } else { // Pending, or complete but erased
2695  if (target_node && target_node->details->online) {
2696  pe_rsc_trace(rsc, "Marking active on %s %p %d", target, target_node,
2697  target_node->details->online);
2698 
2699  native_add_running(rsc, target_node, data_set, FALSE);
2700  if (source_node && source_node->details->online) {
2701  /* This is a partial migration: the migrate_to completed
2702  * successfully on the source, but the migrate_from has not
2703  * completed. Remember the source and target; if the newly
2704  * chosen target remains the same when we schedule actions
2705  * later, we may continue with the migration.
2706  */
2707  rsc->partial_migration_target = target_node;
2708  rsc->partial_migration_source = source_node;
2709  }
2710  } else {
2711  /* Consider it failed here - forces a restart, prevents migration */
2714  }
2715  }
2716 }
2717 
2718 // Is there an action_name in node_name's rsc history newer than call_id?
2719 static bool
2720 newer_op(pe_resource_t *rsc, const char *action_name, const char *node_name,
2721  int call_id, pe_working_set_t *data_set)
2722 {
2723  xmlNode *action = find_lrm_op(rsc->id, action_name, node_name, NULL, TRUE,
2724  data_set);
2725 
2726  return pe__call_id(action) > call_id;
2727 }
2728 
2729 static void
2730 unpack_migrate_to_failure(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op,
2731  pe_working_set_t *data_set)
2732 {
2733  int target_stop_id = 0;
2734  int target_migrate_from_id = 0;
2735  xmlNode *target_stop = NULL;
2736  xmlNode *target_migrate_from = NULL;
2737  const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE);
2738  const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
2739 
2740  // Sanity check
2741  CRM_CHECK(source && target && !strcmp(source, node->details->uname), return);
2742 
2743  /* If a migration failed, we have to assume the resource is active. Clones
2744  * are not allowed to migrate, so role can't be promoted.
2745  */
2746  rsc->role = RSC_ROLE_STARTED;
2747 
2748  // Check for stop on the target
2749  target_stop = find_lrm_op(rsc->id, CRMD_ACTION_STOP, target, NULL,
2750  TRUE, data_set);
2751  target_stop_id = pe__call_id(target_stop);
2752 
2753  // Check for migrate_from on the target
2754  target_migrate_from = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATED, target,
2755  source, TRUE, data_set);
2756  target_migrate_from_id = pe__call_id(target_migrate_from);
2757 
2758  if ((target_stop == NULL) || (target_stop_id < target_migrate_from_id)) {
2759  /* There was no stop on the target, or a stop that happened before a
2760  * migrate_from, so assume the resource is still active on the target
2761  * (if it is up).
2762  */
2763  pe_node_t *target_node = pe_find_node(data_set->nodes, target);
2764 
2765  pe_rsc_trace(rsc, "stop (%d) + migrate_from (%d)",
2766  target_stop_id, target_migrate_from_id);
2767  if (target_node && target_node->details->online) {
2768  native_add_running(rsc, target_node, data_set, FALSE);
2769  }
2770 
2771  } else if (target_migrate_from == NULL) {
2772  /* We know there was a stop on the target, but there may not have been a
2773  * migrate_from (the stop could have happened before migrate_from was
2774  * scheduled or attempted).
2775  *
2776  * That means this could be a "dangling" migration. But first, check
2777  * whether there is a newer successful stop, start, or migrate_from on
2778  * the source node -- it's possible the failed migration was followed by
2779  * a successful stop, full restart, or migration in the reverse
2780  * direction, in which case we don't want to force a stop.
2781  */
2782  int source_migrate_to_id = pe__call_id(xml_op);
2783 
2784  if (newer_op(rsc, CRMD_ACTION_MIGRATED, source, source_migrate_to_id,
2785  data_set)
2786  || newer_op(rsc, CRMD_ACTION_START, source, source_migrate_to_id,
2787  data_set)
2788  || newer_op(rsc, CRMD_ACTION_STOP, source, source_migrate_to_id,
2789  data_set)) {
2790  return;
2791  }
2792 
2793  // Mark node as having dangling migration so we can force a stop later
2794  rsc->dangling_migrations = g_list_prepend(rsc->dangling_migrations, node);
2795  }
2796 }
2797 
2798 static void
2799 unpack_migrate_from_failure(pe_resource_t *rsc, pe_node_t *node,
2800  xmlNode *xml_op, pe_working_set_t *data_set)
2801 {
2802  xmlNode *source_stop = NULL;
2803  xmlNode *source_migrate_to = NULL;
2804  const char *source = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_SOURCE);
2805  const char *target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
2806 
2807  // Sanity check
2808  CRM_CHECK(source && target && !strcmp(target, node->details->uname), return);
2809 
2810  /* If a migration failed, we have to assume the resource is active. Clones
2811  * are not allowed to migrate, so role can't be promoted.
2812  */
2813  rsc->role = RSC_ROLE_STARTED;
2814 
2815  // Check for a stop on the source
2816  source_stop = find_lrm_op(rsc->id, CRMD_ACTION_STOP, source, NULL,
2817  TRUE, data_set);
2818 
2819  // Check for a migrate_to on the source
2820  source_migrate_to = find_lrm_op(rsc->id, CRMD_ACTION_MIGRATE,
2821  source, target, TRUE, data_set);
2822 
2823  if ((source_stop == NULL)
2824  || (pe__call_id(source_stop) < pe__call_id(source_migrate_to))) {
2825  /* There was no stop on the source, or a stop that happened before
2826  * migrate_to, so assume the resource is still active on the source (if
2827  * it is up).
2828  */
2829  pe_node_t *source_node = pe_find_node(data_set->nodes, source);
2830 
2831  if (source_node && source_node->details->online) {
2832  native_add_running(rsc, source_node, data_set, TRUE);
2833  }
2834  }
2835 }
2836 
2837 static void
2838 record_failed_op(xmlNode *op, const pe_node_t *node,
2839  const pe_resource_t *rsc, pe_working_set_t *data_set)
2840 {
2841  xmlNode *xIter = NULL;
2842  const char *op_key = crm_element_value(op, XML_LRM_ATTR_TASK_KEY);
2843 
2844  if (node->details->online == FALSE) {
2845  return;
2846  }
2847 
2848  for (xIter = data_set->failed->children; xIter; xIter = xIter->next) {
2849  const char *key = crm_element_value(xIter, XML_LRM_ATTR_TASK_KEY);
2850  const char *uname = crm_element_value(xIter, XML_ATTR_UNAME);
2851 
2852  if(pcmk__str_eq(op_key, key, pcmk__str_casei) && pcmk__str_eq(uname, node->details->uname, pcmk__str_casei)) {
2853  crm_trace("Skipping duplicate entry %s on %s", op_key, node->details->uname);
2854  return;
2855  }
2856  }
2857 
2858  crm_trace("Adding entry %s on %s", op_key, node->details->uname);
2859  crm_xml_add(op, XML_ATTR_UNAME, node->details->uname);
2860  crm_xml_add(op, XML_LRM_ATTR_RSCID, rsc->id);
2861  add_node_copy(data_set->failed, op);
2862 }
2863 
2864 static const char *get_op_key(xmlNode *xml_op)
2865 {
2866  const char *key = crm_element_value(xml_op, XML_LRM_ATTR_TASK_KEY);
2867  if(key == NULL) {
2868  key = ID(xml_op);
2869  }
2870  return key;
2871 }
2872 
2873 static const char *
2874 last_change_str(xmlNode *xml_op)
2875 {
2876  time_t when;
2877  const char *when_s = NULL;
2878 
2880  &when) == pcmk_ok) {
2881  when_s = pcmk__epoch2str(&when);
2882  if (when_s) {
2883  // Skip day of week to make message shorter
2884  when_s = strchr(when_s, ' ');
2885  if (when_s) {
2886  ++when_s;
2887  }
2888  }
2889  }
2890  return ((when_s && *when_s)? when_s : "unknown time");
2891 }
2892 
2905 static int
2906 cmp_on_fail(enum action_fail_response first, enum action_fail_response second)
2907 {
2908  switch (first) {
2909  case action_fail_demote:
2910  switch (second) {
2911  case action_fail_ignore:
2912  return 1;
2913  case action_fail_demote:
2914  return 0;
2915  default:
2916  return -1;
2917  }
2918  break;
2919 
2921  switch (second) {
2922  case action_fail_ignore:
2923  case action_fail_demote:
2924  case action_fail_recover:
2925  return 1;
2927  return 0;
2928  default:
2929  return -1;
2930  }
2931  break;
2932 
2934  switch (second) {
2935  case action_fail_ignore:
2936  case action_fail_demote:
2937  case action_fail_recover:
2939  return 1;
2941  return 0;
2942  default:
2943  return -1;
2944  }
2945  break;
2946 
2947  default:
2948  break;
2949  }
2950  switch (second) {
2951  case action_fail_demote:
2952  return (first == action_fail_ignore)? -1 : 1;
2953 
2955  switch (first) {
2956  case action_fail_ignore:
2957  case action_fail_demote:
2958  case action_fail_recover:
2959  return -1;
2960  default:
2961  return 1;
2962  }
2963  break;
2964 
2966  switch (first) {
2967  case action_fail_ignore:
2968  case action_fail_demote:
2969  case action_fail_recover:
2971  return -1;
2972  default:
2973  return 1;
2974  }
2975  break;
2976 
2977  default:
2978  break;
2979  }
2980  return first - second;
2981 }
2982 
2983 static void
2984 unpack_rsc_op_failure(pe_resource_t * rsc, pe_node_t * node, int rc, xmlNode * xml_op, xmlNode ** last_failure,
2985  enum action_fail_response * on_fail, pe_working_set_t * data_set)
2986 {
2987  guint interval_ms = 0;
2988  bool is_probe = false;
2989  pe_action_t *action = NULL;
2990 
2991  const char *key = get_op_key(xml_op);
2992  const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK);
2993  const char *exit_reason = crm_element_value(xml_op,
2995 
2996  CRM_ASSERT(rsc);
2997  CRM_CHECK(task != NULL, return);
2998 
2999  *last_failure = xml_op;
3000 
3001  crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms);
3002  if ((interval_ms == 0) && !strcmp(task, CRMD_ACTION_STATUS)) {
3003  is_probe = true;
3004  }
3005 
3006  if (exit_reason == NULL) {
3007  exit_reason = "";
3008  }
3009 
3011  && (rc == PCMK_OCF_NOT_INSTALLED)) {
3012  crm_trace("Unexpected result (%s%s%s) was recorded for "
3013  "%s of %s on %s at %s " CRM_XS " rc=%d id=%s",
3014  services_ocf_exitcode_str(rc),
3015  (*exit_reason? ": " : ""), exit_reason,
3016  (is_probe? "probe" : task), rsc->id, node->details->uname,
3017  last_change_str(xml_op), rc, ID(xml_op));
3018  } else {
3019  crm_warn("Unexpected result (%s%s%s) was recorded for "
3020  "%s of %s on %s at %s " CRM_XS " rc=%d id=%s",
3021  services_ocf_exitcode_str(rc),
3022  (*exit_reason? ": " : ""), exit_reason,
3023  (is_probe? "probe" : task), rsc->id, node->details->uname,
3024  last_change_str(xml_op), rc, ID(xml_op));
3025 
3026  if (is_probe && (rc != PCMK_OCF_OK)
3027  && (rc != PCMK_OCF_NOT_RUNNING)
3028  && (rc != PCMK_OCF_RUNNING_PROMOTED)) {
3029 
3030  /* A failed (not just unexpected) probe result could mean the user
3031  * didn't know resources will be probed even where they can't run.
3032  */
3033  crm_notice("If it is not possible for %s to run on %s, see "
3034  "the resource-discovery option for location constraints",
3035  rsc->id, node->details->uname);
3036  }
3037 
3038  record_failed_op(xml_op, node, rsc, data_set);
3039  }
3040 
3041  action = custom_action(rsc, strdup(key), task, NULL, TRUE, FALSE, data_set);
3042  if (cmp_on_fail(*on_fail, action->on_fail) < 0) {
3043  pe_rsc_trace(rsc, "on-fail %s -> %s for %s (%s)", fail2text(*on_fail),
3044  fail2text(action->on_fail), action->uuid, key);
3045  *on_fail = action->on_fail;
3046  }
3047 
3048  if (!strcmp(task, CRMD_ACTION_STOP)) {
3049  resource_location(rsc, node, -INFINITY, "__stop_fail__", data_set);
3050 
3051  } else if (!strcmp(task, CRMD_ACTION_MIGRATE)) {
3052  unpack_migrate_to_failure(rsc, node, xml_op, data_set);
3053 
3054  } else if (!strcmp(task, CRMD_ACTION_MIGRATED)) {
3055  unpack_migrate_from_failure(rsc, node, xml_op, data_set);
3056 
3057  } else if (!strcmp(task, CRMD_ACTION_PROMOTE)) {
3058  rsc->role = RSC_ROLE_PROMOTED;
3059 
3060  } else if (!strcmp(task, CRMD_ACTION_DEMOTE)) {
3061  if (action->on_fail == action_fail_block) {
3062  rsc->role = RSC_ROLE_PROMOTED;
3064  "demote with on-fail=block");
3065 
3066  } else if(rc == PCMK_OCF_NOT_RUNNING) {
3067  rsc->role = RSC_ROLE_STOPPED;
3068 
3069  } else {
3070  /* Staying in the promoted role would put the scheduler and
3071  * controller into a loop. Setting the role to unpromoted is not
3072  * dangerous because the resource will be stopped as part of
3073  * recovery, and any promotion will be ordered after that stop.
3074  */
3075  rsc->role = RSC_ROLE_UNPROMOTED;
3076  }
3077  }
3078 
3079  if(is_probe && rc == PCMK_OCF_NOT_INSTALLED) {
3080  /* leave stopped */
3081  pe_rsc_trace(rsc, "Leaving %s stopped", rsc->id);
3082  rsc->role = RSC_ROLE_STOPPED;
3083 
3084  } else if (rsc->role < RSC_ROLE_STARTED) {
3085  pe_rsc_trace(rsc, "Setting %s active", rsc->id);
3086  set_active(rsc);
3087  }
3088 
3089  pe_rsc_trace(rsc, "Resource %s: role=%s, unclean=%s, on_fail=%s, fail_role=%s",
3090  rsc->id, role2text(rsc->role),
3091  pcmk__btoa(node->details->unclean),
3092  fail2text(action->on_fail), role2text(action->fail_role));
3093 
3094  if (action->fail_role != RSC_ROLE_STARTED && rsc->next_role < action->fail_role) {
3095  pe__set_next_role(rsc, action->fail_role, "failure");
3096  }
3097 
3098  if (action->fail_role == RSC_ROLE_STOPPED) {
3099  int score = -INFINITY;
3100 
3101  pe_resource_t *fail_rsc = rsc;
3102 
3103  if (fail_rsc->parent) {
3104  pe_resource_t *parent = uber_parent(fail_rsc);
3105 
3106  if (pe_rsc_is_clone(parent)
3107  && !pcmk_is_set(parent->flags, pe_rsc_unique)) {
3108  /* For clone resources, if a child fails on an operation
3109  * with on-fail = stop, all the resources fail. Do this by preventing
3110  * the parent from coming up again. */
3111  fail_rsc = parent;
3112  }
3113  }
3114  crm_notice("%s will not be started under current conditions",
3115  fail_rsc->id);
3116  /* make sure it doesn't come up again */
3117  if (fail_rsc->allowed_nodes != NULL) {
3118  g_hash_table_destroy(fail_rsc->allowed_nodes);
3119  }
3120  fail_rsc->allowed_nodes = pe__node_list2table(data_set->nodes);
3121  g_hash_table_foreach(fail_rsc->allowed_nodes, set_node_score, &score);
3122  }
3123 
3125 }
3126 
3146 static int
3147 determine_op_status(
3148  pe_resource_t *rsc, int rc, int target_rc, pe_node_t * node, xmlNode * xml_op, enum action_fail_response * on_fail, pe_working_set_t * data_set)
3149 {
3150  guint interval_ms = 0;
3151  bool is_probe = false;
3152  int result = PCMK_EXEC_DONE;
3153  const char *key = get_op_key(xml_op);
3154  const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK);
3155  const char *exit_reason = crm_element_value(xml_op,
3157 
3158  CRM_ASSERT(rsc);
3159  CRM_CHECK(task != NULL, return PCMK_EXEC_ERROR);
3160 
3161  if (exit_reason == NULL) {
3162  exit_reason = "";
3163  }
3164 
3165  crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms);
3166  if ((interval_ms == 0) && !strcmp(task, CRMD_ACTION_STATUS)) {
3167  is_probe = true;
3168  task = "probe";
3169  }
3170 
3171  if (target_rc < 0) {
3172  /* Pre-1.0 Pacemaker versions, and Pacemaker 1.1.6 or earlier with
3173  * Heartbeat 2.0.7 or earlier as the cluster layer, did not include the
3174  * target_rc in the transition key, which (along with the similar case
3175  * of a corrupted transition key in the CIB) will be reported to this
3176  * function as -1. Pacemaker 2.0+ does not support rolling upgrades from
3177  * those versions or processing of saved CIB files from those versions,
3178  * so we do not need to care much about this case.
3179  */
3180  result = PCMK_EXEC_ERROR;
3181  crm_warn("Expected result not found for %s on %s (corrupt or obsolete CIB?)",
3182  key, node->details->uname);
3183 
3184  } else if (target_rc != rc) {
3185  result = PCMK_EXEC_ERROR;
3186  pe_rsc_debug(rsc, "%s on %s: expected %d (%s), got %d (%s%s%s)",
3187  key, node->details->uname,
3188  target_rc, services_ocf_exitcode_str(target_rc),
3189  rc, services_ocf_exitcode_str(rc),
3190  (*exit_reason? ": " : ""), exit_reason);
3191  }
3192 
3193  switch (rc) {
3194  case PCMK_OCF_OK:
3195  if (is_probe && (target_rc == PCMK_OCF_NOT_RUNNING)) {
3196  result = PCMK_EXEC_DONE;
3197  pe_rsc_info(rsc, "Probe found %s active on %s at %s",
3198  rsc->id, node->details->uname,
3199  last_change_str(xml_op));
3200  }
3201  break;
3202 
3203  case PCMK_OCF_NOT_RUNNING:
3204  if (is_probe || (target_rc == rc)
3205  || !pcmk_is_set(rsc->flags, pe_rsc_managed)) {
3206 
3207  result = PCMK_EXEC_DONE;
3208  rsc->role = RSC_ROLE_STOPPED;
3209 
3210  /* clear any previous failure actions */
3211  *on_fail = action_fail_ignore;
3212  pe__set_next_role(rsc, RSC_ROLE_UNKNOWN, "not running");
3213  }
3214  break;
3215 
3217  if (is_probe && (rc != target_rc)) {
3218  result = PCMK_EXEC_DONE;
3219  pe_rsc_info(rsc,
3220  "Probe found %s active and promoted on %s at %s",
3221  rsc->id, node->details->uname,
3222  last_change_str(xml_op));
3223  }
3224  rsc->role = RSC_ROLE_PROMOTED;
3225  break;
3226 
3229  rsc->role = RSC_ROLE_PROMOTED;
3230  result = PCMK_EXEC_ERROR;
3231  break;
3232 
3234  result = PCMK_EXEC_ERROR_FATAL;
3235  break;
3236 
3238  if (interval_ms > 0) {
3239  result = PCMK_EXEC_NOT_SUPPORTED;
3240  break;
3241  }
3242  // fall through
3246  if (!pe_can_fence(data_set, node)
3247  && !strcmp(task, CRMD_ACTION_STOP)) {
3248  /* If a stop fails and we can't fence, there's nothing else we can do */
3249  pe_proc_err("No further recovery can be attempted for %s "
3250  "because %s on %s failed (%s%s%s) at %s "
3251  CRM_XS " rc=%d id=%s", rsc->id, task,
3252  node->details->uname, services_ocf_exitcode_str(rc),
3253  (*exit_reason? ": " : ""), exit_reason,
3254  last_change_str(xml_op), rc, ID(xml_op));
3257  }
3258  result = PCMK_EXEC_ERROR_HARD;
3259  break;
3260 
3261  default:
3262  if (result == PCMK_EXEC_DONE) {
3263  crm_info("Treating unknown exit status %d from %s of %s "
3264  "on %s at %s as failure",
3265  rc, task, rsc->id, node->details->uname,
3266  last_change_str(xml_op));
3267  result = PCMK_EXEC_ERROR;
3268  }
3269  break;
3270  }
3271  return result;
3272 }
3273 
3274 // return TRUE if start or monitor last failure but parameters changed
3275 static bool
3276 should_clear_for_param_change(xmlNode *xml_op, const char *task,
3277  pe_resource_t *rsc, pe_node_t *node,
3278  pe_working_set_t *data_set)
3279 {
3280  if (!strcmp(task, "start") || !strcmp(task, "monitor")) {
3281 
3282  if (pe__bundle_needs_remote_name(rsc, data_set)) {
3283  /* We haven't allocated resources yet, so we can't reliably
3284  * substitute addr parameters for the REMOTE_CONTAINER_HACK.
3285  * When that's needed, defer the check until later.
3286  */
3287  pe__add_param_check(xml_op, rsc, node, pe_check_last_failure,
3288  data_set);
3289 
3290  } else {
3291  op_digest_cache_t *digest_data = NULL;
3292 
3293  digest_data = rsc_action_digest_cmp(rsc, xml_op, node, data_set);
3294  switch (digest_data->rc) {
3295  case RSC_DIGEST_UNKNOWN:
3296  crm_trace("Resource %s history entry %s on %s"
3297  " has no digest to compare",
3298  rsc->id, get_op_key(xml_op), node->details->id);
3299  break;
3300  case RSC_DIGEST_MATCH:
3301  break;
3302  default:
3303  return TRUE;
3304  }
3305  }
3306  }
3307  return FALSE;
3308 }
3309 
3310 // Order action after fencing of remote node, given connection rsc
3311 static void
3312 order_after_remote_fencing(pe_action_t *action, pe_resource_t *remote_conn,
3313  pe_working_set_t *data_set)
3314 {
3315  pe_node_t *remote_node = pe_find_node(data_set->nodes, remote_conn->id);
3316 
3317  if (remote_node) {
3318  pe_action_t *fence = pe_fence_op(remote_node, NULL, TRUE, NULL,
3319  FALSE, data_set);
3320 
3322  }
3323 }
3324 
3325 static bool
3326 should_ignore_failure_timeout(pe_resource_t *rsc, xmlNode *xml_op,
3327  const char *task, guint interval_ms,
3328  bool is_last_failure, pe_working_set_t *data_set)
3329 {
3330  /* Clearing failures of recurring monitors has special concerns. The
3331  * executor reports only changes in the monitor result, so if the
3332  * monitor is still active and still getting the same failure result,
3333  * that will go undetected after the failure is cleared.
3334  *
3335  * Also, the operation history will have the time when the recurring
3336  * monitor result changed to the given code, not the time when the
3337  * result last happened.
3338  *
3339  * @TODO We probably should clear such failures only when the failure
3340  * timeout has passed since the last occurrence of the failed result.
3341  * However we don't record that information. We could maybe approximate
3342  * that by clearing only if there is a more recent successful monitor or
3343  * stop result, but we don't even have that information at this point
3344  * since we are still unpacking the resource's operation history.
3345  *
3346  * This is especially important for remote connection resources with a
3347  * reconnect interval, so in that case, we skip clearing failures
3348  * if the remote node hasn't been fenced.
3349  */
3350  if (rsc->remote_reconnect_ms
3352  && (interval_ms != 0) && pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_casei)) {
3353 
3354  pe_node_t *remote_node = pe_find_node(data_set->nodes, rsc->id);
3355 
3356  if (remote_node && !remote_node->details->remote_was_fenced) {
3357  if (is_last_failure) {
3358  crm_info("Waiting to clear monitor failure for remote node %s"
3359  " until fencing has occurred", rsc->id);
3360  }
3361  return TRUE;
3362  }
3363  }
3364  return FALSE;
3365 }
3366 
3389 static bool
3390 check_operation_expiry(pe_resource_t *rsc, pe_node_t *node, int rc,
3391  xmlNode *xml_op, pe_working_set_t *data_set)
3392 {
3393  bool expired = FALSE;
3394  bool is_last_failure = pcmk__ends_with(ID(xml_op), "_last_failure_0");
3395  time_t last_run = 0;
3396  guint interval_ms = 0;
3397  int unexpired_fail_count = 0;
3398  const char *task = crm_element_value(xml_op, XML_LRM_ATTR_TASK);
3399  const char *clear_reason = NULL;
3400 
3401  crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms);
3402 
3403  if ((rsc->failure_timeout > 0)
3405  &last_run) == 0)) {
3406 
3407  // Resource has a failure-timeout, and history entry has a timestamp
3408 
3409  time_t now = get_effective_time(data_set);
3410  time_t last_failure = 0;
3411 
3412  // Is this particular operation history older than the failure timeout?
3413  if ((now >= (last_run + rsc->failure_timeout))
3414  && !should_ignore_failure_timeout(rsc, xml_op, task, interval_ms,
3415  is_last_failure, data_set)) {
3416  expired = TRUE;
3417  }
3418 
3419  // Does the resource as a whole have an unexpired fail count?
3420  unexpired_fail_count = pe_get_failcount(node, rsc, &last_failure,
3421  pe_fc_effective, xml_op,
3422  data_set);
3423 
3424  // Update scheduler recheck time according to *last* failure
3425  crm_trace("%s@%lld is %sexpired @%lld with unexpired_failures=%d timeout=%ds"
3426  " last-failure@%lld",
3427  ID(xml_op), (long long) last_run, (expired? "" : "not "),
3428  (long long) now, unexpired_fail_count, rsc->failure_timeout,
3429  (long long) last_failure);
3430  last_failure += rsc->failure_timeout + 1;
3431  if (unexpired_fail_count && (now < last_failure)) {
3432  pe__update_recheck_time(last_failure, data_set);
3433  }
3434  }
3435 
3436  if (expired) {
3437  if (pe_get_failcount(node, rsc, NULL, pe_fc_default, xml_op, data_set)) {
3438 
3439  // There is a fail count ignoring timeout
3440 
3441  if (unexpired_fail_count == 0) {
3442  // There is no fail count considering timeout
3443  clear_reason = "it expired";
3444 
3445  } else {
3446  /* This operation is old, but there is an unexpired fail count.
3447  * In a properly functioning cluster, this should only be
3448  * possible if this operation is not a failure (otherwise the
3449  * fail count should be expired too), so this is really just a
3450  * failsafe.
3451  */
3452  expired = FALSE;
3453  }
3454 
3455  } else if (is_last_failure && rsc->remote_reconnect_ms) {
3456  /* Clear any expired last failure when reconnect interval is set,
3457  * even if there is no fail count.
3458  */
3459  clear_reason = "reconnect interval is set";
3460  }
3461  }
3462 
3463  if (!expired && is_last_failure
3464  && should_clear_for_param_change(xml_op, task, rsc, node, data_set)) {
3465  clear_reason = "resource parameters have changed";
3466  }
3467 
3468  if (clear_reason != NULL) {
3469  // Schedule clearing of the fail count
3470  pe_action_t *clear_op = pe__clear_failcount(rsc, node, clear_reason,
3471  data_set);
3472 
3473  if (pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)
3474  && rsc->remote_reconnect_ms) {
3475  /* If we're clearing a remote connection due to a reconnect
3476  * interval, we want to wait until any scheduled fencing
3477  * completes.
3478  *
3479  * We could limit this to remote_node->details->unclean, but at
3480  * this point, that's always true (it won't be reliable until
3481  * after unpack_node_history() is done).
3482  */
3483  crm_info("Clearing %s failure will wait until any scheduled "
3484  "fencing of %s completes", task, rsc->id);
3485  order_after_remote_fencing(clear_op, rsc, data_set);
3486  }
3487  }
3488 
3489  if (expired && (interval_ms == 0) && pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_casei)) {
3490  switch(rc) {
3491  case PCMK_OCF_OK:
3492  case PCMK_OCF_NOT_RUNNING:
3494  case PCMK_OCF_DEGRADED:
3496  // Don't expire probes that return these values
3497  expired = FALSE;
3498  break;
3499  }
3500  }
3501 
3502  return expired;
3503 }
3504 
3505 int pe__target_rc_from_xml(xmlNode *xml_op)
3506 {
3507  int target_rc = 0;
3508  const char *key = crm_element_value(xml_op, XML_ATTR_TRANSITION_KEY);
3509 
3510  if (key == NULL) {
3511  return -1;
3512  }
3513  decode_transition_key(key, NULL, NULL, NULL, &target_rc);
3514  return target_rc;
3515 }
3516 
3517 static enum action_fail_response
3518 get_action_on_fail(pe_resource_t *rsc, const char *key, const char *task, pe_working_set_t * data_set)
3519 {
3521  pe_action_t *action = custom_action(rsc, strdup(key), task, NULL, TRUE, FALSE, data_set);
3522 
3523  result = action->on_fail;
3525 
3526  return result;
3527 }
3528 
3529 static void
3530 update_resource_state(pe_resource_t * rsc, pe_node_t * node, xmlNode * xml_op, const char * task, int rc,
3531  xmlNode * last_failure, enum action_fail_response * on_fail, pe_working_set_t * data_set)
3532 {
3533  gboolean clear_past_failure = FALSE;
3534 
3535  CRM_ASSERT(rsc);
3536  CRM_ASSERT(xml_op);
3537 
3538  if (rc == PCMK_OCF_NOT_RUNNING) {
3539  clear_past_failure = TRUE;
3540 
3541  } else if (rc == PCMK_OCF_NOT_INSTALLED) {
3542  rsc->role = RSC_ROLE_STOPPED;
3543 
3544  } else if (pcmk__str_eq(task, CRMD_ACTION_STATUS, pcmk__str_casei)) {
3545  if (last_failure) {
3546  const char *op_key = get_op_key(xml_op);
3547  const char *last_failure_key = get_op_key(last_failure);
3548 
3549  if (pcmk__str_eq(op_key, last_failure_key, pcmk__str_casei)) {
3550  clear_past_failure = TRUE;
3551  }
3552  }
3553 
3554  if (rsc->role < RSC_ROLE_STARTED) {
3555  set_active(rsc);
3556  }
3557 
3558  } else if (pcmk__str_eq(task, CRMD_ACTION_START, pcmk__str_casei)) {
3559  rsc->role = RSC_ROLE_STARTED;
3560  clear_past_failure = TRUE;
3561 
3562  } else if (pcmk__str_eq(task, CRMD_ACTION_STOP, pcmk__str_casei)) {
3563  rsc->role = RSC_ROLE_STOPPED;
3564  clear_past_failure = TRUE;
3565 
3566  } else if (pcmk__str_eq(task, CRMD_ACTION_PROMOTE, pcmk__str_casei)) {
3567  rsc->role = RSC_ROLE_PROMOTED;
3568  clear_past_failure = TRUE;
3569 
3570  } else if (pcmk__str_eq(task, CRMD_ACTION_DEMOTE, pcmk__str_casei)) {
3571 
3572  if (*on_fail == action_fail_demote) {
3573  // Demote clears an error only if on-fail=demote
3574  clear_past_failure = TRUE;
3575  }
3576  rsc->role = RSC_ROLE_UNPROMOTED;
3577 
3578  } else if (pcmk__str_eq(task, CRMD_ACTION_MIGRATED, pcmk__str_casei)) {
3579  rsc->role = RSC_ROLE_STARTED;
3580  clear_past_failure = TRUE;
3581 
3582  } else if (pcmk__str_eq(task, CRMD_ACTION_MIGRATE, pcmk__str_casei)) {
3583  unpack_migrate_to_success(rsc, node, xml_op, data_set);
3584 
3585  } else if (rsc->role < RSC_ROLE_STARTED) {
3586  pe_rsc_trace(rsc, "%s active on %s", rsc->id, node->details->uname);
3587  set_active(rsc);
3588  }
3589 
3590  /* clear any previous failure actions */
3591  if (clear_past_failure) {
3592  switch (*on_fail) {
3593  case action_fail_stop:
3594  case action_fail_fence:
3595  case action_fail_migrate:
3596  case action_fail_standby:
3597  pe_rsc_trace(rsc, "%s.%s is not cleared by a completed stop",
3598  rsc->id, fail2text(*on_fail));
3599  break;
3600 
3601  case action_fail_block:
3602  case action_fail_ignore:
3603  case action_fail_demote:
3604  case action_fail_recover:
3606  *on_fail = action_fail_ignore;
3607  pe__set_next_role(rsc, RSC_ROLE_UNKNOWN, "clear past failures");
3608  break;
3610  if (rsc->remote_reconnect_ms == 0) {
3611  /* With no reconnect interval, the connection is allowed to
3612  * start again after the remote node is fenced and
3613  * completely stopped. (With a reconnect interval, we wait
3614  * for the failure to be cleared entirely before attempting
3615  * to reconnect.)
3616  */
3617  *on_fail = action_fail_ignore;
3619  "clear past failures and reset remote");
3620  }
3621  break;
3622  }
3623  }
3624 }
3625 
3646 static int
3647 remap_monitor_rc(int rc, xmlNode *xml_op, const pe_node_t *node,
3648  const pe_resource_t *rsc, pe_working_set_t *data_set)
3649 {
3650  int remapped_rc = pcmk__effective_rc(rc);
3651 
3652  if (rc != remapped_rc) {
3653  crm_trace("Remapping monitor result %d to %d", rc, remapped_rc);
3654  if (!node->details->shutdown || node->details->online) {
3655  record_failed_op(xml_op, node, rsc, data_set);
3656  }
3657  }
3658  return remapped_rc;
3659 }
3660 
3661 static void
3662 unpack_rsc_op(pe_resource_t *rsc, pe_node_t *node, xmlNode *xml_op,
3663  xmlNode **last_failure, enum action_fail_response *on_fail,
3664  pe_working_set_t *data_set)
3665 {
3666  int rc = 0;
3667  int task_id = 0;
3668  int target_rc = 0;
3669  int status = PCMK_EXEC_UNKNOWN;
3670  guint interval_ms = 0;
3671  const char *task = NULL;
3672  const char *task_key = NULL;
3673  const char *exit_reason = NULL;
3674  bool expired = FALSE;
3675  pe_resource_t *parent = rsc;
3676  enum action_fail_response failure_strategy = action_fail_recover;
3677 
3678  CRM_CHECK(rsc && node && xml_op, return);
3679 
3680  target_rc = pe__target_rc_from_xml(xml_op);
3681  task_key = get_op_key(xml_op);
3682  task = crm_element_value(xml_op, XML_LRM_ATTR_TASK);
3683  exit_reason = crm_element_value(xml_op, XML_LRM_ATTR_EXIT_REASON);
3684  if (exit_reason == NULL) {
3685  exit_reason = "";
3686  }
3687 
3689  crm_element_value_int(xml_op, XML_LRM_ATTR_CALLID, &task_id);
3690  crm_element_value_int(xml_op, XML_LRM_ATTR_OPSTATUS, &status);
3691  crm_element_value_ms(xml_op, XML_LRM_ATTR_INTERVAL_MS, &interval_ms);
3692 
3693  CRM_CHECK(task != NULL, return);
3694  CRM_CHECK((status >= PCMK_EXEC_PENDING) && (status <= PCMK_EXEC_MAX),
3695  return);
3696 
3697  if (!strcmp(task, CRMD_ACTION_NOTIFY) ||
3698  !strcmp(task, CRMD_ACTION_METADATA)) {
3699  /* safe to ignore these */
3700  return;
3701  }
3702 
3703  if (!pcmk_is_set(rsc->flags, pe_rsc_unique)) {
3704  parent = uber_parent(rsc);
3705  }
3706 
3707  pe_rsc_trace(rsc, "Unpacking task %s/%s (call_id=%d, status=%d, rc=%d) on %s (role=%s)",
3708  task_key, task, task_id, status, rc, node->details->uname, role2text(rsc->role));
3709 
3710  if (node->details->unclean) {
3711  pe_rsc_trace(rsc, "Node %s (where %s is running) is unclean."
3712  " Further action depends on the value of the stop's on-fail attribute",
3713  node->details->uname, rsc->id);
3714  }
3715 
3716  /* It should be possible to call remap_monitor_rc() first then call
3717  * check_operation_expiry() only if rc != target_rc, because there should
3718  * never be a fail count without at least one unexpected result in the
3719  * resource history. That would be more efficient by avoiding having to call
3720  * check_operation_expiry() for expected results.
3721  *
3722  * However, we do have such configurations in the scheduler regression
3723  * tests, even if it shouldn't be possible with the current code. It's
3724  * probably a good idea anyway, but that would require updating the test
3725  * inputs to something currently possible.
3726  */
3727 
3728  if ((status != PCMK_EXEC_NOT_INSTALLED)
3729  && check_operation_expiry(rsc, node, rc, xml_op, data_set)) {
3730  expired = TRUE;
3731  }
3732 
3733  if (!strcmp(task, CRMD_ACTION_STATUS)) {
3734  rc = remap_monitor_rc(rc, xml_op, node, rsc, data_set);
3735  }
3736 
3737  if (expired && (rc != target_rc)) {
3738  const char *magic = crm_element_value(xml_op, XML_ATTR_TRANSITION_MAGIC);
3739 
3740  if (interval_ms == 0) {
3741  crm_notice("Ignoring expired %s failure on %s "
3742  CRM_XS " actual=%d expected=%d magic=%s",
3743  task_key, node->details->uname, rc, target_rc, magic);
3744  goto done;
3745 
3746  } else if(node->details->online && node->details->unclean == FALSE) {
3747  /* Reschedule the recurring monitor. CancelXmlOp() won't work at
3748  * this stage, so as a hacky workaround, forcibly change the restart
3749  * digest so check_action_definition() does what we want later.
3750  *
3751  * @TODO We should skip this if there is a newer successful monitor.
3752  * Also, this causes rescheduling only if the history entry
3753  * has an op-digest (which the expire-non-blocked-failure
3754  * scheduler regression test doesn't, but that may not be a
3755  * realistic scenario in production).
3756  */
3757  crm_notice("Rescheduling %s after failure expired on %s "
3758  CRM_XS " actual=%d expected=%d magic=%s",
3759  task_key, node->details->uname, rc, target_rc, magic);
3760  crm_xml_add(xml_op, XML_LRM_ATTR_RESTART_DIGEST, "calculated-failure-timeout");
3761  goto done;
3762  }
3763  }
3764 
3765  /* If the executor reported an operation status of anything but done or
3766  * error, consider that final. But for done or error, we know better whether
3767  * it should be treated as a failure or not, because we know the expected
3768  * result.
3769  */
3770  if(status == PCMK_EXEC_DONE || status == PCMK_EXEC_ERROR) {
3771  status = determine_op_status(rsc, rc, target_rc, node, xml_op, on_fail, data_set);
3772  pe_rsc_trace(rsc, "Remapped %s status to %d", task_key, status);
3773  }
3774 
3775  switch (status) {
3776  case PCMK_EXEC_CANCELLED:
3777  // Should never happen
3778  pe_err("Resource history contains cancellation '%s' "
3779  "(%s of %s on %s at %s)",
3780  ID(xml_op), task, rsc->id, node->details->uname,
3781  last_change_str(xml_op));
3782  goto done;
3783 
3784  case PCMK_EXEC_PENDING:
3785  if (!strcmp(task, CRMD_ACTION_START)) {
3787  set_active(rsc);
3788 
3789  } else if (!strcmp(task, CRMD_ACTION_PROMOTE)) {
3790  rsc->role = RSC_ROLE_PROMOTED;
3791 
3792  } else if (!strcmp(task, CRMD_ACTION_MIGRATE) && node->details->unclean) {
3793  /* If a pending migrate_to action is out on a unclean node,
3794  * we have to force the stop action on the target. */
3795  const char *migrate_target = crm_element_value(xml_op, XML_LRM_ATTR_MIGRATE_TARGET);
3796  pe_node_t *target = pe_find_node(data_set->nodes, migrate_target);
3797  if (target) {
3798  stop_action(rsc, target, FALSE);
3799  }
3800  }
3801 
3802  if (rsc->pending_task == NULL) {
3803  if ((interval_ms != 0) || strcmp(task, CRMD_ACTION_STATUS)) {
3804  rsc->pending_task = strdup(task);
3805  rsc->pending_node = node;
3806  } else {
3807  /* Pending probes are not printed, even if pending
3808  * operations are requested. If someone ever requests that
3809  * behavior, enable the below and the corresponding part of
3810  * native.c:native_pending_task().
3811  */
3812 #if 0
3813  rsc->pending_task = strdup("probe");
3814  rsc->pending_node = node;
3815 #endif
3816  }
3817  }
3818  goto done;
3819 
3820  case PCMK_EXEC_DONE:
3821  pe_rsc_trace(rsc, "%s of %s on %s completed at %s " CRM_XS " id=%s",
3822  task, rsc->id, node->details->uname,
3823  last_change_str(xml_op), ID(xml_op));
3824  update_resource_state(rsc, node, xml_op, task, rc, *last_failure, on_fail, data_set);
3825  goto done;
3826 
3828  failure_strategy = get_action_on_fail(rsc, task_key, task, data_set);
3829  if (failure_strategy == action_fail_ignore) {
3830  crm_warn("Cannot ignore failed %s of %s on %s: "
3831  "Resource agent doesn't exist "
3832  CRM_XS " status=%d rc=%d id=%s",
3833  task, rsc->id, node->details->uname, status, rc,
3834  ID(xml_op));
3835  /* Also for printing it as "FAILED" by marking it as pe_rsc_failed later */
3836  *on_fail = action_fail_migrate;
3837  }
3838  resource_location(parent, node, -INFINITY, "hard-error", data_set);
3839  unpack_rsc_op_failure(rsc, node, rc, xml_op, last_failure, on_fail, data_set);
3840  goto done;
3841 
3843  if (pe__is_guest_or_remote_node(node)
3845  /* We should never get into a situation where a managed remote
3846  * connection resource is considered OK but a resource action
3847  * behind the connection gets a "not connected" status. But as a
3848  * fail-safe in case a bug or unusual circumstances do lead to
3849  * that, ensure the remote connection is considered failed.
3850  */
3853  }
3854  break; // Not done, do error handling
3855 
3856  case PCMK_EXEC_ERROR:
3857  case PCMK_EXEC_ERROR_HARD:
3858  case PCMK_EXEC_ERROR_FATAL:
3859  case PCMK_EXEC_TIMEOUT:
3861  case PCMK_EXEC_INVALID:
3862  break; // Not done, do error handling
3863 
3864  /* These should only be possible in fence action results, not operation
3865  * history, but have some handling in place as a fail-safe.
3866  */
3868  case PCMK_EXEC_NO_SECRETS:
3869  status = PCMK_EXEC_ERROR_HARD;
3870  break; // Not done, do error handling
3871  }
3872 
3873  failure_strategy = get_action_on_fail(rsc, task_key, task, data_set);
3874  if ((failure_strategy == action_fail_ignore)
3875  || (failure_strategy == action_fail_restart_container
3876  && !strcmp(task, CRMD_ACTION_STOP))) {
3877 
3878  crm_warn("Pretending failed %s (%s%s%s) of %s on %s at %s "
3879  "succeeded " CRM_XS " rc=%d id=%s",
3880  task, services_ocf_exitcode_str(rc),
3881  (*exit_reason? ": " : ""), exit_reason, rsc->id,
3882  node->details->uname, last_change_str(xml_op), rc,
3883  ID(xml_op));
3884 
3885  update_resource_state(rsc, node, xml_op, task, target_rc, *last_failure,
3886  on_fail, data_set);
3887  crm_xml_add(xml_op, XML_ATTR_UNAME, node->details->uname);
3889 
3890  record_failed_op(xml_op, node, rsc, data_set);
3891 
3892  if ((failure_strategy == action_fail_restart_container)
3893  && cmp_on_fail(*on_fail, action_fail_recover) <= 0) {
3894  *on_fail = failure_strategy;
3895  }
3896 
3897  } else {
3898  unpack_rsc_op_failure(rsc, node, rc, xml_op, last_failure, on_fail,
3899  data_set);
3900 
3901  if (status == PCMK_EXEC_ERROR_HARD) {
3902  do_crm_log(rc != PCMK_OCF_NOT_INSTALLED?LOG_ERR:LOG_NOTICE,
3903  "Preventing %s from restarting on %s because "
3904  "of hard failure (%s%s%s)" CRM_XS " rc=%d id=%s",
3905  parent->id, node->details->uname,
3906  services_ocf_exitcode_str(rc),
3907  (*exit_reason? ": " : ""), exit_reason,
3908  rc, ID(xml_op));
3909  resource_location(parent, node, -INFINITY, "hard-error", data_set);
3910 
3911  } else if (status == PCMK_EXEC_ERROR_FATAL) {
3912  crm_err("Preventing %s from restarting anywhere because "
3913  "of fatal failure (%s%s%s) " CRM_XS " rc=%d id=%s",
3914  parent->id, services_ocf_exitcode_str(rc),
3915  (*exit_reason? ": " : ""), exit_reason,
3916  rc, ID(xml_op));
3917  resource_location(parent, NULL, -INFINITY, "fatal-error", data_set);
3918  }
3919  }
3920 
3921 done:
3922  pe_rsc_trace(rsc, "Resource %s after %s: role=%s, next=%s",
3923  rsc->id, task, role2text(rsc->role),
3924  role2text(rsc->next_role));
3925 }
3926 
3927 static void
3928 add_node_attrs(xmlNode *xml_obj, pe_node_t *node, bool overwrite,
3929  pe_working_set_t *data_set)
3930 {
3931  const char *cluster_name = NULL;
3932 
3933  pe_rule_eval_data_t rule_data = {
3934  .node_hash = NULL,
3935  .role = RSC_ROLE_UNKNOWN,
3936  .now = data_set->now,
3937  .match_data = NULL,
3938  .rsc_data = NULL,
3939  .op_data = NULL
3940  };
3941 
3942  g_hash_table_insert(node->details->attrs,
3943  strdup(CRM_ATTR_UNAME), strdup(node->details->uname));
3944 
3945  g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_ID),
3946  strdup(node->details->id));
3947  if (pcmk__str_eq(node->details->id, data_set->dc_uuid, pcmk__str_casei)) {
3948  data_set->dc_node = node;
3949  node->details->is_dc = TRUE;
3950  g_hash_table_insert(node->details->attrs,
3951  strdup(CRM_ATTR_IS_DC), strdup(XML_BOOLEAN_TRUE));
3952  } else {
3953  g_hash_table_insert(node->details->attrs,
3954  strdup(CRM_ATTR_IS_DC), strdup(XML_BOOLEAN_FALSE));
3955  }
3956 
3957  cluster_name = g_hash_table_lookup(data_set->config_hash, "cluster-name");
3958  if (cluster_name) {
3959  g_hash_table_insert(node->details->attrs, strdup(CRM_ATTR_CLUSTER_NAME),
3960  strdup(cluster_name));
3961  }
3962 
3963  pe__unpack_dataset_nvpairs(xml_obj, XML_TAG_ATTR_SETS, &rule_data,
3964  node->details->attrs, NULL, overwrite, data_set);
3965 
3966  if (pe_node_attribute_raw(node, CRM_ATTR_SITE_NAME) == NULL) {
3967  const char *site_name = pe_node_attribute_raw(node, "site-name");
3968 
3969  if (site_name) {
3970  g_hash_table_insert(node->details->attrs,
3971  strdup(CRM_ATTR_SITE_NAME),
3972  strdup(site_name));
3973 
3974  } else if (cluster_name) {
3975  /* Default to cluster-name if unset */
3976  g_hash_table_insert(node->details->attrs,
3977  strdup(CRM_ATTR_SITE_NAME),
3978  strdup(cluster_name));
3979  }
3980  }
3981 }
3982 
3983 static GList *
3984 extract_operations(const char *node, const char *rsc, xmlNode * rsc_entry, gboolean active_filter)
3985 {
3986  int counter = -1;
3987  int stop_index = -1;
3988  int start_index = -1;
3989 
3990  xmlNode *rsc_op = NULL;
3991 
3992  GList *gIter = NULL;
3993  GList *op_list = NULL;
3994  GList *sorted_op_list = NULL;
3995 
3996  /* extract operations */
3997  op_list = NULL;
3998  sorted_op_list = NULL;
3999 
4000  for (rsc_op = pcmk__xe_first_child(rsc_entry);
4001  rsc_op != NULL; rsc_op = pcmk__xe_next(rsc_op)) {
4002 
4003  if (pcmk__str_eq((const char *)rsc_op->name, XML_LRM_TAG_RSC_OP,
4004  pcmk__str_none)) {
4005  crm_xml_add(rsc_op, "resource", rsc);
4006  crm_xml_add(rsc_op, XML_ATTR_UNAME, node);
4007  op_list = g_list_prepend(op_list, rsc_op);
4008  }
4009  }
4010 
4011  if (op_list == NULL) {
4012  /* if there are no operations, there is nothing to do */
4013  return NULL;
4014  }
4015 
4016  sorted_op_list = g_list_sort(op_list, sort_op_by_callid);
4017 
4018  /* create active recurring operations as optional */
4019  if (active_filter == FALSE) {
4020  return sorted_op_list;
4021  }
4022 
4023  op_list = NULL;
4024 
4025  calculate_active_ops(sorted_op_list, &start_index, &stop_index);
4026 
4027  for (gIter = sorted_op_list; gIter != NULL; gIter = gIter->next) {
4028  xmlNode *rsc_op = (xmlNode *) gIter->data;
4029 
4030  counter++;
4031 
4032  if (start_index < stop_index) {
4033  crm_trace("Skipping %s: not active", ID(rsc_entry));
4034  break;
4035 
4036  } else if (counter < start_index) {
4037  crm_trace("Skipping %s: old", ID(rsc_op));
4038  continue;
4039  }
4040  op_list = g_list_append(op_list, rsc_op);
4041  }
4042 
4043  g_list_free(sorted_op_list);
4044  return op_list;
4045 }
4046 
4047 GList *
4048 find_operations(const char *rsc, const char *node, gboolean active_filter,
4049  pe_working_set_t * data_set)
4050 {
4051  GList *output = NULL;
4052  GList *intermediate = NULL;
4053 
4054  xmlNode *tmp = NULL;
4055  xmlNode *status = find_xml_node(data_set->input, XML_CIB_TAG_STATUS, TRUE);
4056 
4057  pe_node_t *this_node = NULL;
4058 
4059  xmlNode *node_state = NULL;
4060 
4061  for (node_state = pcmk__xe_first_child(status); node_state != NULL;
4062  node_state = pcmk__xe_next(node_state)) {
4063 
4064  if (pcmk__str_eq((const char *)node_state->name, XML_CIB_TAG_STATE, pcmk__str_none)) {
4065  const char *uname = crm_element_value(node_state, XML_ATTR_UNAME);
4066 
4067  if (node != NULL && !pcmk__str_eq(uname, node, pcmk__str_casei)) {
4068  continue;
4069  }
4070 
4071  this_node = pe_find_node(data_set->nodes, uname);
4072  if(this_node == NULL) {
4073  CRM_LOG_ASSERT(this_node != NULL);
4074  continue;
4075 
4076  } else if (pe__is_guest_or_remote_node(this_node)) {
4077  determine_remote_online_status(data_set, this_node);
4078 
4079  } else {
4080  determine_online_status(node_state, this_node, data_set);
4081  }
4082 
4083  if (this_node->details->online
4084  || pcmk_is_set(data_set->flags, pe_flag_stonith_enabled)) {
4085  /* offline nodes run no resources...
4086  * unless stonith is enabled in which case we need to
4087  * make sure rsc start events happen after the stonith
4088  */
4089  xmlNode *lrm_rsc = NULL;
4090 
4091  tmp = find_xml_node(node_state, XML_CIB_TAG_LRM, FALSE);
4092  tmp = find_xml_node(tmp, XML_LRM_TAG_RESOURCES, FALSE);
4093 
4094  for (lrm_rsc = pcmk__xe_first_child(tmp); lrm_rsc != NULL;
4095  lrm_rsc = pcmk__xe_next(lrm_rsc)) {
4096 
4097  if (pcmk__str_eq((const char *)lrm_rsc->name,
4099 
4100  const char *rsc_id = crm_element_value(lrm_rsc, XML_ATTR_ID);
4101 
4102  if (rsc != NULL && !pcmk__str_eq(rsc_id, rsc, pcmk__str_casei)) {
4103  continue;
4104  }
4105 
4106  intermediate = extract_operations(uname, rsc_id, lrm_rsc, active_filter);
4107  output = g_list_concat(output, intermediate);
4108  }
4109  }
4110  }
4111  }
4112  }
4113 
4114  return output;
4115 }
GHashTable * tags
Definition: pe_types.h:180
Services API.
pcmk__cpg_host_t host
Definition: cpg.c:49
#define CRM_CHECK(expr, failure_action)
Definition: logging.h:225
#define XML_RSC_OP_LAST_CHANGE
Definition: msg_xml.h:317
void verify_pe_options(GHashTable *options)
Definition: common.c:302
pe_node_t * pe_find_node(GList *node_list, const char *uname)
Definition: status.c:434
xmlNode * find_xml_node(xmlNode *cib, const char *node_path, gboolean must_find)
Definition: xml.c:445
enum pe_quorum_policy no_quorum_policy
Definition: pe_types.h:149
#define RSC_STOP
Definition: crm.h:204
A dumping ground.
pe_node_t * pe_find_node_any(GList *node_list, const char *id, const char *uname)
Definition: status.c:406
Service failed and possibly in promoted role.
Definition: results.h:171
GHashTable * pe__node_list2table(GList *list)
Definition: utils.c:204
#define crm_notice(fmt, args...)
Definition: logging.h:359
GHashTable * known_on
Definition: pe_types.h:367
#define CRMD_ACTION_MIGRATED
Definition: crm.h:174
xmlNode * failed
Definition: pe_types.h:165
No connection to executor.
Definition: results.h:316
#define pe_flag_stop_action_orphans
Definition: pe_types.h:104
bool pe__is_guest_or_remote_node(const pe_node_t *node)
Definition: remote.c:41
GHashTable * attrs
Definition: pe_types.h:234
#define pe_rsc_debug(rsc, fmt, args...)
Definition: internal.h:19
gboolean unseen
Definition: pe_types.h:218
#define XML_CONFIG_ATTR_SHUTDOWN_LOCK
Definition: msg_xml.h:387
gboolean fixed
Definition: pe_types.h:242
char data[0]
Definition: cpg.c:55
#define INFINITY
Definition: crm.h:99
gint sort_rsc_priority(gconstpointer a, gconstpointer b)
Definition: utils.c:362
Service active and promoted.
Definition: results.h:170
#define pe__set_action_flags(action, flags_to_set)
Definition: internal.h:59
#define CRM_ATTR_KIND
Definition: crm.h:116
#define XML_NODE_IS_FENCED
Definition: msg_xml.h:280
#define XML_ATTR_TRANSITION_MAGIC
Definition: msg_xml.h:398
GHashTable * state
Definition: pe_types.h:461
int pcmk__scan_min_int(const char *text, int *result, int minimum)
Definition: strings.c:127
#define CRM_ATTR_IS_DC
Definition: crm.h:118
#define stop_action(rsc, node, optional)
Definition: internal.h:377
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_t * container
Definition: pe_types.h:380
bool pcmk__strcase_any_of(const char *s,...) G_GNUC_NULL_TERMINATED
Definition: strings.c:931
pe_node_t * partial_migration_source
Definition: pe_types.h:365
#define XML_ATTR_QUORUM_PANIC
Definition: msg_xml.h:117
#define pe_flag_concurrent_fencing
Definition: pe_types.h:101
#define XML_ATTR_TYPE
Definition: msg_xml.h:132
#define XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY
Definition: msg_xml.h:389
void pe__update_recheck_time(time_t recheck, pe_working_set_t *data_set)
Definition: utils.c:2396
enum rsc_role_e role
Definition: pe_types.h:370
#define XML_TAG_UTILIZATION
Definition: msg_xml.h:211
#define pcmk__config_warn(fmt...)
GList * children
Definition: pe_types.h:377
#define XML_RULE_ATTR_SCORE
Definition: msg_xml.h:333
#define XML_BOOLEAN_FALSE
Definition: msg_xml.h:141
#define pe_flag_symmetric_cluster
Definition: pe_types.h:95
gboolean standby
Definition: pe_types.h:460
#define pe_rsc_stop
Definition: pe_types.h:262
int priority_fencing_delay
Definition: pe_types.h:190
xmlNode * first_named_child(const xmlNode *parent, const char *name)
Definition: xml.c:2789
#define pe_flag_remove_after_stop
Definition: pe_types.h:110
enum rsc_role_e next_role
Definition: pe_types.h:371
#define pe_flag_maintenance_mode
Definition: pe_types.h:96
int char2score(const char *score)
Definition: utils.c:61
#define pcmk__config_err(fmt...)
#define pe_proc_warn(fmt...)
Definition: internal.h:33
pe_resource_t * remote_rsc
Definition: pe_types.h:230
#define XML_TAG_TRANSIENT_NODEATTRS
Definition: msg_xml.h:403
#define CRMD_ACTION_NOTIFY
Definition: crm.h:187
gboolean unpack_resources(xmlNode *xml_resources, pe_working_set_t *data_set)
Definition: unpack.c:759
GHashTable * meta
Definition: pe_types.h:373
#define pe_rsc_unique
Definition: pe_types.h:254
Service safely stopped.
Definition: results.h:169
resource_object_functions_t * fns
Definition: pe_types.h:333
#define XML_CIB_TAG_TAG
Definition: msg_xml.h:430
#define XML_LRM_TAG_RESOURCE
Definition: msg_xml.h:264
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_node_t * pe__copy_node(const pe_node_t *this_node)
Definition: utils.c:142
#define CRMD_ACTION_PROMOTE
Definition: crm.h:182
bool pe__bundle_needs_remote_name(pe_resource_t *rsc, pe_working_set_t *data_set)
Definition: bundle.c:943
gboolean unpack_tags(xmlNode *xml_tags, pe_working_set_t *data_set)
Definition: unpack.c:820
void calculate_active_ops(GList *sorted_op_list, int *start_index, int *stop_index)
Definition: unpack.c:2250
gboolean pending
Definition: pe_types.h:216
#define XML_NVPAIR_ATTR_NAME
Definition: msg_xml.h:377
#define XML_NODE_IS_MAINTENANCE
Definition: msg_xml.h:281
char * id
Definition: pe_types.h:465
gint sort_op_by_callid(gconstpointer a, gconstpointer b)
Definition: utils.c:1735
#define XML_NODE_EXPECTED
Definition: msg_xml.h:276
#define XML_CIB_TAG_RSC_TEMPLATE
Definition: msg_xml.h:219
Necessary CIB secrets are unavailable.
Definition: results.h:319
pe_resource_t * pe_find_resource(GList *rsc_list, const char *id_rh)
Definition: status.c:382
time_t get_effective_time(pe_working_set_t *data_set)
Definition: utils.c:1852
#define CRM_LOG_ASSERT(expr)
Definition: logging.h:209
const char * pe_pref(GHashTable *options, const char *name)
Definition: common.c:308
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:1691
Service promoted but more likely to fail soon.
Definition: results.h:173
gboolean unpack_remote_nodes(xmlNode *xml_resources, pe_working_set_t *data_set)
Definition: unpack.c:622
enum crm_ais_msg_types type
Definition: cpg.c:48
void copy_in_properties(xmlNode *target, xmlNode *src)
Definition: xml.c:520
#define CRMD_JOINSTATE_NACK
Definition: crm.h:166
#define XML_CIB_TAG_LRM
Definition: msg_xml.h:262
#define CRM_ATTR_CLUSTER_NAME
Definition: crm.h:119
pe_node_t * partial_migration_target
Definition: pe_types.h:364
GHashTable * tickets
Definition: pe_types.h:152
Action did not complete in time.
Definition: results.h:310
gboolean remote_was_fenced
Definition: pe_types.h:225
int crm_element_value_int(const xmlNode *data, const char *name, int *dest)
Retrieve the integer value of an XML attribute.
Definition: nvpair.c:565
Execution failed, do not retry on node.
Definition: results.h:313
bool pcmk__ends_with(const char *s, const char *match)
Definition: strings.c:536
#define pe_flag_have_quorum
Definition: pe_types.h:94
gboolean unpack_config(xmlNode *config, pe_working_set_t *data_set)
Definition: unpack.c:197
char * pending_task
Definition: pe_types.h:346
xmlNode * get_xpath_object(const char *xpath, xmlNode *xml_obj, int error_level)
Definition: xpath.c:214
#define pe_proc_err(fmt...)
Definition: internal.h:32
gboolean remote_requires_reset
Definition: pe_types.h:224
action_fail_response
Definition: common.h:34
Action was cancelled.
Definition: results.h:309
char * strndup(const char *str, size_t len)
char * dc_uuid
Definition: pe_types.h:141
No fence device is configured for target.
Definition: results.h:318
const char * action
Definition: pcmk_fence.c:30
#define pe__set_resource_flags(resource, flags_to_set)
Definition: internal.h:47
GList * resources
Definition: pe_types.h:158
#define XML_CIB_TAG_PROPSET
Definition: msg_xml.h:202
pe_node_t * lock_node
Definition: pe_types.h:384
GList * nodes
Definition: pe_types.h:157
int pcmk__effective_rc(int rc)
Definition: agents.c:71
#define pe_flag_stop_everything
Definition: pe_types.h:105
void pe__unpack_dataset_nvpairs(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:2410
#define XML_LRM_ATTR_RSCID
Definition: msg_xml.h:306
gboolean remote_maintenance
Definition: pe_types.h:226
#define CRMD_ACTION_START
Definition: crm.h:176
#define demote_action(rsc, node, optional)
Definition: internal.h:403
gboolean is_dc
Definition: pe_types.h:221
#define XML_LRM_ATTR_TASK_KEY
Definition: msg_xml.h:298
bool pe__is_remote_node(const pe_node_t *node)
Definition: remote.c:25
#define XML_TAG_ATTR_SETS
Definition: msg_xml.h:203
#define XML_LRM_ATTR_TASK
Definition: msg_xml.h:297
const char * role2text(enum rsc_role_e role)
Definition: common.c:459
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
pe_node_t *(* location)(const pe_resource_t *, GList **, int)
Definition: pe_types.h:54
#define CRMD_ACTION_STOP
Definition: crm.h:179
pe_resource_t * pe__find_bundle_replica(const pe_resource_t *bundle, const pe_node_t *node)
Definition: bundle.c:1380
#define STATUS_PATH_MAX
Definition: unpack.c:2540
int weight
Definition: pe_types.h:241
int pcmk__scan_ll(const char *text, long long *result, long long default_value)
Definition: strings.c:97
#define pe_flag_have_remote_nodes
Definition: pe_types.h:117
GList * dangling_migrations
Definition: pe_types.h:378
#define CRMD_JOINSTATE_DOWN
Definition: crm.h:163
Maximum value for this enum.
Definition: results.h:322
#define crm_warn(fmt, args...)
Definition: logging.h:358
guint remote_reconnect_ms
Definition: pe_types.h:345
#define XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT
Definition: msg_xml.h:388
#define CRMD_ACTION_DEMOTE
Definition: crm.h:184
char * clone_zero(const char *last_rsc_id)
Definition: unpack.c:1650
#define pe_rsc_allow_migrate
Definition: pe_types.h:273
#define pe_rsc_orphan_container_filler
Definition: pe_types.h:251
int rc
Definition: pcmk_fence.c:35
int crm_element_value_ms(const xmlNode *data, const char *name, guint *dest)
Retrieve the millisecond value of an XML attribute.
Definition: nvpair.c:622
#define pe_rsc_failed
Definition: pe_types.h:267
#define crm_debug(fmt, args...)
Definition: logging.h:362
int pe__target_rc_from_xml(xmlNode *xml_op)
Definition: unpack.c:3505
pe_resource_t * uber_parent(pe_resource_t *rsc)
Definition: complex.c:903
#define XML_RSC_ATTR_CONTAINER
Definition: msg_xml.h:241
Utility functions.
Used only to initialize variables.
Definition: results.h:306
#define XML_ATTR_ID
Definition: msg_xml.h:129
const char * crm_element_value(const xmlNode *data, const char *name)
Retrieve the value of an XML attribute.
Definition: nvpair.c:529
#define XML_CIB_TAG_RESOURCE
Definition: msg_xml.h:214
#define pe_rsc_is_container
Definition: pe_types.h:277
#define XML_BOOLEAN_TRUE
Definition: msg_xml.h:140
#define XML_CIB_TAG_STATE
Definition: msg_xml.h:198
const char * pe_base_name_end(const char *id)
Definition: unpack.c:1588
bool pe__is_guest_node(const pe_node_t *node)
Definition: remote.c:33
Parameter invalid (in local context)
Definition: results.h:164
gboolean unpacked
Definition: pe_types.h:227
int failure_timeout
Definition: pe_types.h:343
match only clone instances
Definition: pe_types.h:86
#define pe_rsc_start_pending
Definition: pe_types.h:269
Parameter invalid (inherently)
Definition: results.h:168
#define pe_warn_once(pe_wo_bit, fmt...)
Definition: internal.h:154
pe_ticket_t * ticket_new(const char *ticket_id, pe_working_set_t *data_set)
Definition: utils.c:1976
#define CRM_ATTR_UNAME
Definition: crm.h:114
#define XML_NODE_IS_PEER
Definition: msg_xml.h:278
#define crm_trace(fmt, args...)
Definition: logging.h:363
#define CRMD_JOINSTATE_MEMBER
Definition: crm.h:165
#define do_crm_log(level, fmt, args...)
Log a message.
Definition: logging.h:166
enum rsc_digest_cmp_val rc
Definition: internal.h:478
bool xml_contains_remote_node(xmlNode *xml)
Definition: remote.c:84
char * crm_strdup_printf(char const *format,...) G_GNUC_PRINTF(1
#define pcmk_is_set(g, f)
Convenience alias for pcmk_all_flags_set(), to check single flag.
Definition: util.h:114
Insufficient privileges.
Definition: results.h:166
xmlNode * add_node_copy(xmlNode *new_parent, xmlNode *xml_node)
Definition: xml.c:674
#define pe_flag_startup_fencing
Definition: pe_types.h:112
const char * stonith_action
Definition: pe_types.h:143
struct pe_node_shared_s * details
Definition: pe_types.h:244
gboolean get_target_role(pe_resource_t *rsc, enum rsc_role_e *role)
Definition: utils.c:1867
#define crm_log_xml_debug(xml, text)
Definition: logging.h:370
bool pe__shutdown_requested(pe_node_t *node)
Definition: utils.c:2381
#define pe_rsc_needs_fencing
Definition: pe_types.h:280
gboolean order_actions(pe_action_t *lh_action, pe_action_t *rh_action, enum pe_ordering order)
Definition: utils.c:1905
unsigned long long flags
Definition: pe_types.h:348
const char * uname
Definition: pe_types.h:209
#define pe_rsc_promotable
Definition: pe_types.h:256
#define XML_TAG_META_SETS
Definition: msg_xml.h:204
Wrappers for and extensions to libxml2.
GHashTable * config_hash
Definition: pe_types.h:151
#define XML_ATTR_UNAME
Definition: msg_xml.h:151
char * clone_name
Definition: pe_types.h:323
gboolean add_tag_ref(GHashTable *tags, const char *tag_name, const char *obj_ref)
Definition: utils.c:2277
#define XML_RSC_ATTR_MANAGED
Definition: msg_xml.h:232
xmlNode * create_xml_node(xmlNode *parent, const char *name)
Definition: xml.c:696
time_t lock_time
Definition: pe_types.h:385
Action completed, result is known.
Definition: results.h:308
#define pe_flag_stonith_enabled
Definition: pe_types.h:98
gboolean unpack_status(xmlNode *status, pe_working_set_t *data_set)
Definition: unpack.c:1242
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:650
const char * pe_node_attribute_raw(pe_node_t *node, const char *name)
Definition: common.c:635
time_t last_granted
Definition: pe_types.h:459
pe_resource_t * pe__create_clone_child(pe_resource_t *rsc, pe_working_set_t *data_set)
Definition: clone.c:102
#define XML_RSC_ATTR_TARGET_ROLE
Definition: msg_xml.h:233
void(* free)(pe_resource_t *)
Definition: pe_types.h:55
#define XML_LRM_ATTR_MIGRATE_TARGET
Definition: msg_xml.h:323
Execution failed, do not retry anywhere.
Definition: results.h:314
#define CIB_OPTIONS_FIRST
Definition: msg_xml.h:82
gboolean standby
Definition: pe_types.h:214
#define XML_RSC_ATTR_REMOTE_NODE
Definition: msg_xml.h:244
#define XML_LRM_ATTR_EXIT_REASON
Definition: msg_xml.h:315
#define XML_LRM_ATTR_RESTART_DIGEST
Definition: msg_xml.h:313
void pe__free_digests(gpointer ptr)
Definition: pe_digest.c:34
gboolean expected_up
Definition: pe_types.h:220
Dependencies not available locally.
Definition: results.h:167
enum pe_obj_types variant
Definition: pe_types.h:331
xmlNode * input
Definition: pe_types.h:137
gboolean granted
Definition: pe_types.h:458
#define XML_CIB_TAG_NODE
Definition: msg_xml.h:199
const char * placement_strategy
Definition: pe_types.h:144
const char * pcmk__epoch2str(time_t *when)
Definition: iso8601.c:1714
uint32_t id
Definition: cpg.c:45
const char * id
Definition: pe_types.h:208
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
char * id
Definition: pe_types.h:457
guint shutdown_lock
Definition: pe_types.h:189
Unspecified error.
Definition: results.h:163
GList * refs
Definition: pe_types.h:466
const char * target
Definition: pcmk_fence.c:29
GList * fillers
Definition: pe_types.h:381
bool pe_can_fence(pe_working_set_t *data_set, pe_node_t *node)
Definition: utils.c:90
int pcmk__score_green
Definition: utils.c:57
GList * running_rsc
Definition: pe_types.h:231
#define XML_ATTR_TRANSITION_KEY
Definition: msg_xml.h:399
gboolean rsc_discovery_enabled
Definition: pe_types.h:223
#define CRM_XS
Definition: logging.h:54
pe_node_t * dc_node
Definition: pe_types.h:142
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:96
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:389
Requested action not implemented.
Definition: results.h:165
const char * localhost
Definition: pe_types.h:179
#define pe_flag_quick_location
Definition: pe_types.h:119
Service active but more likely to fail soon.
Definition: results.h:172
gboolean is_remote_node
Definition: pe_types.h:351
pe_node_t * pending_node
Definition: pe_types.h:383
const char * fail2text(enum action_fail_response fail)
Definition: common.c:314
Agent does not implement requested action.
Definition: results.h:311
GHashTable * pcmk__strkey_table(GDestroyNotify key_destroy_func, GDestroyNotify value_destroy_func)
Definition: strings.c:611
int pcmk__score_red
Definition: utils.c:56
#define XML_LRM_TAG_RESOURCES
Definition: msg_xml.h:263
#define crm_err(fmt, args...)
Definition: logging.h:357
#define XML_CIB_TAG_TICKET_STATE
Definition: msg_xml.h:427
#define set_config_flag(data_set, option, flag)
Definition: unpack.c:34
#define CRM_ASSERT(expr)
Definition: results.h:42
xmlXPathObjectPtr xpath_search(xmlNode *xml_top, const char *path)
Definition: xpath.c:139
Success.
Definition: results.h:162
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:364
GHashTable * node_hash
Definition: common.h:185
char guint crm_parse_interval_spec(const char *input)
Parse milliseconds from a Pacemaker interval specification.
Definition: utils.c:314
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
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:392
void pe__set_next_role(pe_resource_t *rsc, enum rsc_role_e role, const char *why)
Definition: complex.c:1116
#define XML_ATTR_HAVE_WATCHDOG
Definition: msg_xml.h:119
#define XML_NODE_ATTR_RSC_DISCOVERY
Definition: msg_xml.h:380
gboolean unpack_nodes(xmlNode *xml_nodes, pe_working_set_t *data_set)
Definition: unpack.c:529
int pcmk__score_yellow
Definition: utils.c:58
#define CRMD_ACTION_METADATA
Definition: crm.h:191
#define XML_LRM_ATTR_INTERVAL_MS
Definition: msg_xml.h:295
#define XML_LRM_ATTR_CALLID
Definition: msg_xml.h:309
#define CRMD_ACTION_MIGRATE
Definition: crm.h:173
#define XML_NVPAIR_ATTR_VALUE
Definition: msg_xml.h:378
GHashTable * utilization
Definition: pe_types.h:235
void native_add_running(pe_resource_t *rsc, pe_node_t *node, pe_working_set_t *data_set, gboolean failed)
Definition: native.c:90
gboolean shutdown
Definition: pe_types.h:219
char uname[MAX_NAME]
Definition: cpg.c:50
#define pe__clear_resource_flags(resource, flags_to_clear)
Definition: internal.h:53
#define crm_str(x)
Definition: logging.h:383
#define XML_LRM_ATTR_OPSTATUS
Definition: msg_xml.h:307
#define pe__set_working_set_flags(working_set, flags_to_set)
Definition: internal.h:35
#define CRMD_JOINSTATE_PENDING
Definition: crm.h:164
rsc_role_e
Possible roles that a resource can be in.
Definition: common.h:92
GList * running_on
Definition: pe_types.h:366
#define pe_rsc_block
Definition: pe_types.h:250
CRM_TRACE_INIT_DATA(pe_status)
Agent or dependency not available locally.
Definition: results.h:315
gboolean maintenance
Definition: pe_types.h:222
#define XML_LRM_ATTR_RC
Definition: msg_xml.h:308
#define pcmk_ok
Definition: results.h:68
GHashTable * digest_cache
cache of calculated resource digests
Definition: pe_types.h:236
bool pe__is_universal_clone(pe_resource_t *rsc, pe_working_set_t *data_set)
Definition: clone.c:1015
#define pe_rsc_failure_ignored
Definition: pe_types.h:275
#define XML_NODE_JOIN_STATE
Definition: msg_xml.h:275
Action is in progress.
Definition: results.h:307
void destroy_ticket(gpointer data)
Definition: utils.c:1964
#define XML_CIB_TAG_STATUS
Definition: msg_xml.h:179
#define XML_CIB_TAG_OBJ_REF
Definition: msg_xml.h:431
gboolean common_unpack(xmlNode *xml_obj, pe_resource_t **rsc, pe_resource_t *parent, pe_working_set_t *data_set)
Definition: complex.c:493
#define XML_NODE_IN_CLUSTER
Definition: msg_xml.h:277
gboolean crm_is_true(const char *s)
Definition: strings.c:416
#define pe_flag_have_stonith_resource
Definition: pe_types.h:99
#define CRM_ATTR_SITE_NAME
Definition: crm.h:120
#define XML_CIB_TAG_GROUP
Definition: msg_xml.h:215
#define pe_flag_enable_unfencing
Definition: pe_types.h:100
GList * find_operations(const char *rsc, const char *node, gboolean active_filter, pe_working_set_t *data_set)
Definition: unpack.c:4048
#define XML_LRM_TAG_RSC_OP
Definition: msg_xml.h:265
#define pe_rsc_trace(rsc, fmt, args...)
Definition: internal.h:20
uint32_t pe_wo
Definition: unpack.c:67
#define ID(x)
Definition: msg_xml.h:456
unsigned long long flags
Definition: pe_types.h:146
#define pe_err(fmt...)
Definition: internal.h:22
Action cannot be attempted (e.g. shutdown)
Definition: results.h:317
#define ONLINESTATUS
Definition: util.h:38
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: utils.c:1643
gboolean standby_onfail
Definition: pe_types.h:215
pe_action_t * pe__clear_resource_history(pe_resource_t *rsc, pe_node_t *node, pe_working_set_t *data_set)
Definition: utils.c:2456
char * name
Definition: pcmk_fence.c:31
#define XML_LRM_ATTR_MIGRATE_SOURCE
Definition: msg_xml.h:322
pe_resource_t *(* find_rsc)(pe_resource_t *parent, const char *search, const pe_node_t *node, int flags)
Definition: pe_types.h:45
void freeXpathObject(xmlXPathObjectPtr xpathObj)
Definition: xpath.c:39
#define CRM_ATTR_ID
Definition: crm.h:115
gint sort_node_uname(gconstpointer a, gconstpointer b)
Definition: utils.c:218
gboolean unclean
Definition: pe_types.h:217
#define pe__clear_working_set_flags(working_set, flags_to_clear)
Definition: internal.h:41
#define XPATH_ENABLE_UNFENCING
Definition: unpack.c:176
#define pe_flag_start_failure_fatal
Definition: pe_types.h:107
enum node_type type
Definition: pe_types.h:210
#define XML_CIB_TAG_TICKETS
Definition: msg_xml.h:426
crm_time_t * now
Definition: pe_types.h:138
Execution failed, may be retried.
Definition: results.h:312
#define crm_info(fmt, args...)
Definition: logging.h:360
#define pe_rsc_managed
Definition: pe_types.h:249
#define pe_rsc_orphan
Definition: pe_types.h:248
GHashTable * template_rsc_sets
Definition: pe_types.h:178
gboolean online
Definition: pe_types.h:213
char * clone_strip(const char *last_rsc_id)
Definition: unpack.c:1628
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
GList * stop_needed
Definition: pe_types.h:186
pe_resource_t * parent
Definition: pe_types.h:329
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: utils.c:2135
void pe_free_action(pe_action_t *action)
Definition: utils.c:1456
#define pe_flag_shutdown_lock
Definition: pe_types.h:113
#define pe_rsc_info(rsc, fmt, args...)
Definition: internal.h:18
char * id
Definition: pe_types.h:322
GHashTable * allowed_nodes
Definition: pe_types.h:368
pe_action_t * custom_action(pe_resource_t *rsc, char *key, const char *task, pe_node_t *on_node, gboolean optional, gboolean foo, pe_working_set_t *data_set)
Create or update an action object.
Definition: utils.c:731
#define pe_flag_startup_probes
Definition: pe_types.h:115
#define CRMD_ACTION_STATUS
Definition: crm.h:190
xmlNode * crm_next_same_xml(const xmlNode *sibling)
Get next instance of same XML tag.
Definition: xml.c:2815
#define pe_flag_stop_rsc_orphans
Definition: pe_types.h:103