Page MenuHomeClusterLabs Projects
Files F4624371

No OneTemporary
Actions

View Options

diff --git a/include/crm/common/actions.h b/include/crm/common/actions.h
index 32106d2d3b..f9f8932bd0 100644
--- a/include/crm/common/actions.h
+++ b/include/crm/common/actions.h
@@ -1,47 +1,48 @@
/*
* Copyright 2023 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
* This source code is licensed under the GNU Lesser General Public License
* version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY.
*/
#ifndef PCMK__CRM_COMMON_ACTIONS__H
#define PCMK__CRM_COMMON_ACTIONS__H
#ifdef __cplusplus
extern "C" {
#endif
/*!
* \file
* \brief APIs related to actions
* \ingroup core
*/
// Action names as strings
#define PCMK_ACTION_CANCEL "cancel"
#define PCMK_ACTION_DELETE "delete"
#define PCMK_ACTION_DEMOTE "demote"
#define PCMK_ACTION_DEMOTED "demoted"
#define PCMK_ACTION_META_DATA "meta-data"
#define PCMK_ACTION_MIGRATE_FROM "migrate_from"
#define PCMK_ACTION_MIGRATE_TO "migrate_to"
#define PCMK_ACTION_MONITOR "monitor"
+#define PCMK_ACTION_NOTIFIED "notified"
#define PCMK_ACTION_NOTIFY "notify"
#define PCMK_ACTION_PROMOTE "promote"
#define PCMK_ACTION_PROMOTED "promoted"
#define PCMK_ACTION_RELOAD "reload"
#define PCMK_ACTION_RELOAD_AGENT "reload-agent"
#define PCMK_ACTION_RUNNING "running"
#define PCMK_ACTION_START "start"
#define PCMK_ACTION_STOP "stop"
#define PCMK_ACTION_STOPPED "stopped"
#define PCMK_ACTION_VALIDATE_ALL "validate-all"
#ifdef __cplusplus
}
#endif
#endif // PCMK__CRM_COMMON_ACTIONS__H
diff --git a/include/crm/crm.h b/include/crm/crm.h
index 4f0d0cde41..7f5c34c0b3 100644
--- a/include/crm/crm.h
+++ b/include/crm/crm.h
@@ -1,196 +1,196 @@
/*
* Copyright 2004-2023 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
* This source code is licensed under the GNU Lesser General Public License
* version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY.
*/
#ifndef PCMK__CRM_CRM__H
# define PCMK__CRM_CRM__H
# include <crm_config.h>
# include <stdlib.h>
# include <glib.h>
# include <stdbool.h>
# include <string.h>
# include <libxml/tree.h>
#ifdef __cplusplus
extern "C" {
#endif
/**
* \file
* \brief A dumping ground
* \ingroup core
*/
#ifndef PCMK_ALLOW_DEPRECATED
/*!
* \brief Allow use of deprecated Pacemaker APIs
*
* By default, external code using Pacemaker headers is allowed to use
* deprecated Pacemaker APIs. If PCMK_ALLOW_DEPRECATED is defined to 0 before
* including any Pacemaker headers, deprecated APIs will be unusable. It is
* strongly recommended to leave this unchanged for production and release
* builds, to avoid breakage when users upgrade to new Pacemaker releases that
* deprecate more APIs. This should be defined to 0 only for development and
* testing builds when desiring to check for usage of currently deprecated APIs.
*/
#define PCMK_ALLOW_DEPRECATED 1
#endif
/*!
* The CRM feature set assists with compatibility in mixed-version clusters.
* The major version number increases when nodes with different versions
* would not work (rolling upgrades are not allowed). The minor version
* number increases when mixed-version clusters are allowed only during
* rolling upgrades (a node with the oldest feature set will be elected DC). The
* minor-minor version number is ignored, but allows resource agents to detect
* cluster support for various features.
*
* The feature set also affects the processing of old saved CIBs (such as for
* many scheduler regression tests).
*
* Particular feature points currently tested by Pacemaker code:
*
* >2.1: Operation updates include timing data
* >=3.0.5: XML v2 digests are created
* >=3.0.8: Peers do not need acks for cancellations
* >=3.0.9: DC will send its own shutdown request to all peers
* XML v2 patchsets are created by default
* >=3.0.13: Fail counts include operation name and interval
* >=3.2.0: DC supports PCMK_EXEC_INVALID and PCMK_EXEC_NOT_CONNECTED
*/
# define CRM_FEATURE_SET "3.18.0"
/* Pacemaker's CPG protocols use fixed-width binary fields for the sender and
* recipient of a CPG message. This imposes an arbitrary limit on cluster node
* names.
*/
//! \brief Maximum length of a Corosync cluster node name (in bytes)
#define MAX_NAME 256
# define CRM_META "CRM_meta"
extern char *crm_system_name;
/* *INDENT-OFF* */
// How we represent "infinite" scores
# define CRM_SCORE_INFINITY 1000000
# define CRM_INFINITY_S "INFINITY"
# define CRM_PLUS_INFINITY_S "+" CRM_INFINITY_S
# define CRM_MINUS_INFINITY_S "-" CRM_INFINITY_S
/* @COMPAT API < 2.0.0 Deprecated "infinity" aliases
*
* INFINITY might be defined elsewhere (e.g. math.h), so undefine it first.
* This, of course, complicates any attempt to use the other definition in any
* code that includes this header.
*/
# undef INFINITY
# define INFINITY_S "INFINITY"
# define MINUS_INFINITY_S "-INFINITY"
# define INFINITY 1000000
/* Sub-systems */
# define CRM_SYSTEM_DC "dc"
#define CRM_SYSTEM_DCIB "dcib" // Primary instance of CIB manager
# define CRM_SYSTEM_CIB "cib"
# define CRM_SYSTEM_CRMD "crmd"
# define CRM_SYSTEM_LRMD "lrmd"
# define CRM_SYSTEM_PENGINE "pengine"
# define CRM_SYSTEM_TENGINE "tengine"
# define CRM_SYSTEM_STONITHD "stonithd"
# define CRM_SYSTEM_MCP "pacemakerd"
// Names of internally generated node attributes
# define CRM_ATTR_UNAME "#uname"
# define CRM_ATTR_ID "#id"
# define CRM_ATTR_KIND "#kind"
# define CRM_ATTR_ROLE "#role"
# define CRM_ATTR_IS_DC "#is_dc"
# define CRM_ATTR_CLUSTER_NAME "#cluster-name"
# define CRM_ATTR_SITE_NAME "#site-name"
# define CRM_ATTR_UNFENCED "#node-unfenced"
# define CRM_ATTR_DIGESTS_ALL "#digests-all"
# define CRM_ATTR_DIGESTS_SECURE "#digests-secure"
# define CRM_ATTR_PROTOCOL "#attrd-protocol"
# define CRM_ATTR_FEATURE_SET "#feature-set"
/* Valid operations */
# define CRM_OP_NOOP "noop"
# define CRM_OP_JOIN_ANNOUNCE "join_announce"
# define CRM_OP_JOIN_OFFER "join_offer"
# define CRM_OP_JOIN_REQUEST "join_request"
# define CRM_OP_JOIN_ACKNAK "join_ack_nack"
# define CRM_OP_JOIN_CONFIRM "join_confirm"
# define CRM_OP_PING "ping"
# define CRM_OP_NODE_INFO "node-info"
# define CRM_OP_THROTTLE "throttle"
# define CRM_OP_VOTE "vote"
# define CRM_OP_NOVOTE "no-vote"
# define CRM_OP_HELLO "hello"
# define CRM_OP_PECALC "pe_calc"
# define CRM_OP_QUIT "quit"
# define CRM_OP_LOCAL_SHUTDOWN "start_shutdown"
# define CRM_OP_SHUTDOWN_REQ "req_shutdown"
# define CRM_OP_SHUTDOWN "do_shutdown"
# define CRM_OP_FENCE "stonith"
# define CRM_OP_REGISTER "register"
# define CRM_OP_IPC_FWD "ipc_fwd"
# define CRM_OP_INVOKE_LRM "lrm_invoke"
# define CRM_OP_LRM_REFRESH "lrm_refresh" //!< Deprecated since 1.1.10
# define CRM_OP_LRM_DELETE "lrm_delete"
# define CRM_OP_LRM_FAIL "lrm_fail"
# define CRM_OP_PROBED "probe_complete"
# define CRM_OP_REPROBE "probe_again"
# define CRM_OP_CLEAR_FAILCOUNT "clear_failcount"
# define CRM_OP_REMOTE_STATE "remote_state"
# define CRM_OP_RELAXED_SET "one-or-more"
# define CRM_OP_RELAXED_CLONE "clone-one-or-more"
# define CRM_OP_RM_NODE_CACHE "rm_node_cache"
# define CRM_OP_MAINTENANCE_NODES "maintenance_nodes"
/* Possible cluster membership states */
# define CRMD_JOINSTATE_DOWN "down"
# define CRMD_JOINSTATE_PENDING "pending"
# define CRMD_JOINSTATE_MEMBER "member"
# define CRMD_JOINSTATE_NACK "banned"
-# define CRMD_ACTION_NOTIFIED "notified"
+# define CRMD_ACTION_NOTIFIED PCMK_ACTION_NOTIFIED
# define CRMD_METADATA_CALL_TIMEOUT 30000
/* short names */
-# define RSC_NOTIFIED CRMD_ACTION_NOTIFIED
+# define RSC_NOTIFIED PCMK_ACTION_NOTIFIED
/* *INDENT-ON* */
# include <crm/common/actions.h>
# include <crm/common/cib.h>
# include <crm/common/logging.h>
# include <crm/common/util.h>
static inline const char *
crm_action_str(const char *task, guint interval_ms) {
if ((task != NULL) && (interval_ms == 0)
&& (strcasecmp(task, PCMK_ACTION_MONITOR) == 0)) {
return "probe";
}
return task;
}
#if !defined(PCMK_ALLOW_DEPRECATED) || (PCMK_ALLOW_DEPRECATED == 1)
#include <crm/crm_compat.h>
#endif
#ifdef __cplusplus
}
#endif
#endif
diff --git a/lib/pacemaker/pcmk_sched_instances.c b/lib/pacemaker/pcmk_sched_instances.c
index 3cf6b9fb11..19dbde9e6c 100644
--- a/lib/pacemaker/pcmk_sched_instances.c
+++ b/lib/pacemaker/pcmk_sched_instances.c
@@ -1,1668 +1,1668 @@
/*
* Copyright 2004-2023 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
* This source code is licensed under the GNU General Public License version 2
* or later (GPLv2+) WITHOUT ANY WARRANTY.
*/
/* This file is intended for code usable with both clone instances and bundle
* replica containers.
*/
#include <crm_internal.h>
#include <crm/msg_xml.h>
#include <pacemaker-internal.h>
#include "libpacemaker_private.h"
/*!
* \internal
* \brief Check whether a node is allowed to run an instance
*
* \param[in] instance Clone instance or bundle container to check
* \param[in] node Node to check
* \param[in] max_per_node Maximum number of instances allowed to run on a node
*
* \return true if \p node is allowed to run \p instance, otherwise false
*/
static bool
can_run_instance(const pe_resource_t *instance, const pe_node_t *node,
int max_per_node)
{
pe_node_t *allowed_node = NULL;
if (pcmk_is_set(instance->flags, pe_rsc_orphan)) {
pe_rsc_trace(instance, "%s cannot run on %s: orphaned",
instance->id, pe__node_name(node));
return false;
}
if (!pcmk__node_available(node, false, false)) {
pe_rsc_trace(instance,
"%s cannot run on %s: node cannot run resources",
instance->id, pe__node_name(node));
return false;
}
allowed_node = pcmk__top_allowed_node(instance, node);
if (allowed_node == NULL) {
crm_warn("%s cannot run on %s: node not allowed",
instance->id, pe__node_name(node));
return false;
}
if (allowed_node->weight < 0) {
pe_rsc_trace(instance, "%s cannot run on %s: parent score is %s there",
instance->id, pe__node_name(node),
pcmk_readable_score(allowed_node->weight));
return false;
}
if (allowed_node->count >= max_per_node) {
pe_rsc_trace(instance,
"%s cannot run on %s: node already has %d instance%s",
instance->id, pe__node_name(node), max_per_node,
pcmk__plural_s(max_per_node));
return false;
}
pe_rsc_trace(instance, "%s can run on %s (%d already running)",
instance->id, pe__node_name(node), allowed_node->count);
return true;
}
/*!
* \internal
* \brief Ban a clone instance or bundle replica from unavailable allowed nodes
*
* \param[in,out] instance Clone instance or bundle replica to ban
* \param[in] max_per_node Maximum instances allowed to run on a node
*/
static void
ban_unavailable_allowed_nodes(pe_resource_t *instance, int max_per_node)
{
if (instance->allowed_nodes != NULL) {
GHashTableIter iter;
pe_node_t *node = NULL;
g_hash_table_iter_init(&iter, instance->allowed_nodes);
while (g_hash_table_iter_next(&iter, NULL, (void **) &node)) {
if (!can_run_instance(instance, node, max_per_node)) {
pe_rsc_trace(instance, "Banning %s from unavailable node %s",
instance->id, pe__node_name(node));
node->weight = -INFINITY;
for (GList *child_iter = instance->children;
child_iter != NULL; child_iter = child_iter->next) {
pe_resource_t *child = (pe_resource_t *) child_iter->data;
pe_node_t *child_node = NULL;
child_node = g_hash_table_lookup(child->allowed_nodes,
node->details->id);
if (child_node != NULL) {
pe_rsc_trace(instance,
"Banning %s child %s "
"from unavailable node %s",
instance->id, child->id,
pe__node_name(node));
child_node->weight = -INFINITY;
}
}
}
}
}
}
/*!
* \internal
* \brief Create a hash table with a single node in it
*
* \param[in] node Node to copy into new table
*
* \return Newly created hash table containing a copy of \p node
* \note The caller is responsible for freeing the result with
* g_hash_table_destroy().
*/
static GHashTable *
new_node_table(pe_node_t *node)
{
GHashTable *table = pcmk__strkey_table(NULL, free);
node = pe__copy_node(node);
g_hash_table_insert(table, (gpointer) node->details->id, node);
return table;
}
/*!
* \internal
* \brief Apply a resource's parent's colocation scores to a node table
*
* \param[in] rsc Resource whose colocations should be applied
* \param[in,out] nodes Node table to apply colocations to
*/
static void
apply_parent_colocations(const pe_resource_t *rsc, GHashTable **nodes)
{
GList *colocations = pcmk__this_with_colocations(rsc);
for (const GList *iter = colocations; iter != NULL; iter = iter->next) {
const pcmk__colocation_t *colocation = iter->data;
pe_resource_t *other = colocation->primary;
float factor = colocation->score / (float) INFINITY;
other->cmds->add_colocated_node_scores(other, rsc, rsc->id, nodes,
colocation, factor,
pcmk__coloc_select_default);
}
g_list_free(colocations);
colocations = pcmk__with_this_colocations(rsc);
for (const GList *iter = colocations; iter != NULL; iter = iter->next) {
const pcmk__colocation_t *colocation = iter->data;
pe_resource_t *other = colocation->dependent;
float factor = colocation->score / (float) INFINITY;
if (!pcmk__colocation_has_influence(colocation, rsc)) {
continue;
}
other->cmds->add_colocated_node_scores(other, rsc, rsc->id, nodes,
colocation, factor,
pcmk__coloc_select_nonnegative);
}
g_list_free(colocations);
}
/*!
* \internal
* \brief Compare clone or bundle instances based on colocation scores
*
* Determine the relative order in which two clone or bundle instances should be
* assigned to nodes, considering the scores of colocation constraints directly
* or indirectly involving them.
*
* \param[in] instance1 First instance to compare
* \param[in] instance2 Second instance to compare
*
* \return A negative number if \p instance1 should be assigned first,
* a positive number if \p instance2 should be assigned first,
* or 0 if assignment order doesn't matter
*/
static int
cmp_instance_by_colocation(const pe_resource_t *instance1,
const pe_resource_t *instance2)
{
int rc = 0;
pe_node_t *node1 = NULL;
pe_node_t *node2 = NULL;
pe_node_t *current_node1 = pe__current_node(instance1);
pe_node_t *current_node2 = pe__current_node(instance2);
GHashTable *colocated_scores1 = NULL;
GHashTable *colocated_scores2 = NULL;
CRM_ASSERT((instance1 != NULL) && (instance1->parent != NULL)
&& (instance2 != NULL) && (instance2->parent != NULL)
&& (current_node1 != NULL) && (current_node2 != NULL));
// Create node tables initialized with each node
colocated_scores1 = new_node_table(current_node1);
colocated_scores2 = new_node_table(current_node2);
// Apply parental colocations
apply_parent_colocations(instance1, &colocated_scores1);
apply_parent_colocations(instance2, &colocated_scores2);
// Find original nodes again, with scores updated for colocations
node1 = g_hash_table_lookup(colocated_scores1, current_node1->details->id);
node2 = g_hash_table_lookup(colocated_scores2, current_node2->details->id);
// Compare nodes by updated scores
if (node1->weight < node2->weight) {
crm_trace("Assign %s (%d on %s) after %s (%d on %s)",
instance1->id, node1->weight, pe__node_name(node1),
instance2->id, node2->weight, pe__node_name(node2));
rc = 1;
} else if (node1->weight > node2->weight) {
crm_trace("Assign %s (%d on %s) before %s (%d on %s)",
instance1->id, node1->weight, pe__node_name(node1),
instance2->id, node2->weight, pe__node_name(node2));
rc = -1;
}
g_hash_table_destroy(colocated_scores1);
g_hash_table_destroy(colocated_scores2);
return rc;
}
/*!
* \internal
* \brief Check whether a resource or any of its children are failed
*
* \param[in] rsc Resource to check
*
* \return true if \p rsc or any of its children are failed, otherwise false
*/
static bool
did_fail(const pe_resource_t *rsc)
{
if (pcmk_is_set(rsc->flags, pe_rsc_failed)) {
return true;
}
for (GList *iter = rsc->children; iter != NULL; iter = iter->next) {
if (did_fail((const pe_resource_t *) iter->data)) {
return true;
}
}
return false;
}
/*!
* \internal
* \brief Check whether a node is allowed to run a resource
*
* \param[in] rsc Resource to check
* \param[in,out] node Node to check (will be set NULL if not allowed)
*
* \return true if *node is either NULL or allowed for \p rsc, otherwise false
*/
static bool
node_is_allowed(const pe_resource_t *rsc, pe_node_t **node)
{
if (*node != NULL) {
pe_node_t *allowed = g_hash_table_lookup(rsc->allowed_nodes,
(*node)->details->id);
if ((allowed == NULL) || (allowed->weight < 0)) {
pe_rsc_trace(rsc, "%s: current location (%s) is unavailable",
rsc->id, pe__node_name(*node));
*node = NULL;
return false;
}
}
return true;
}
/*!
* \internal
* \brief Compare two clone or bundle instances' instance numbers
*
* \param[in] a First instance to compare
* \param[in] b Second instance to compare
*
* \return A negative number if \p a's instance number is lower,
* a positive number if \p b's instance number is lower,
* or 0 if their instance numbers are the same
*/
gint
pcmk__cmp_instance_number(gconstpointer a, gconstpointer b)
{
const pe_resource_t *instance1 = (const pe_resource_t *) a;
const pe_resource_t *instance2 = (const pe_resource_t *) b;
char *div1 = NULL;
char *div2 = NULL;
CRM_ASSERT((instance1 != NULL) && (instance2 != NULL));
// Clone numbers are after a colon, bundle numbers after a dash
div1 = strrchr(instance1->id, ':');
if (div1 == NULL) {
div1 = strrchr(instance1->id, '-');
}
div2 = strrchr(instance2->id, ':');
if (div2 == NULL) {
div2 = strrchr(instance2->id, '-');
}
CRM_ASSERT((div1 != NULL) && (div2 != NULL));
return (gint) (strtol(div1 + 1, NULL, 10) - strtol(div2 + 1, NULL, 10));
}
/*!
* \internal
* \brief Compare clone or bundle instances according to assignment order
*
* Compare two clone or bundle instances according to the order they should be
* assigned to nodes, preferring (in order):
*
* - Active instance that is less multiply active
* - Instance that is not active on a disallowed node
* - Instance with higher configured priority
* - Active instance whose current node can run resources
* - Active instance whose parent is allowed on current node
* - Active instance whose current node has fewer other instances
* - Active instance
* - Instance that isn't failed
* - Instance whose colocations result in higher score on current node
* - Instance with lower ID in lexicographic order
*
* \param[in] a First instance to compare
* \param[in] b Second instance to compare
*
* \return A negative number if \p a should be assigned first,
* a positive number if \p b should be assigned first,
* or 0 if assignment order doesn't matter
*/
gint
pcmk__cmp_instance(gconstpointer a, gconstpointer b)
{
int rc = 0;
pe_node_t *node1 = NULL;
pe_node_t *node2 = NULL;
unsigned int nnodes1 = 0;
unsigned int nnodes2 = 0;
bool can1 = true;
bool can2 = true;
const pe_resource_t *instance1 = (const pe_resource_t *) a;
const pe_resource_t *instance2 = (const pe_resource_t *) b;
CRM_ASSERT((instance1 != NULL) && (instance2 != NULL));
node1 = instance1->fns->active_node(instance1, &nnodes1, NULL);
node2 = instance2->fns->active_node(instance2, &nnodes2, NULL);
/* If both instances are running and at least one is multiply
* active, prefer instance that's running on fewer nodes.
*/
if ((nnodes1 > 0) && (nnodes2 > 0)) {
if (nnodes1 < nnodes2) {
crm_trace("Assign %s (active on %d) before %s (active on %d): "
"less multiply active",
instance1->id, nnodes1, instance2->id, nnodes2);
return -1;
} else if (nnodes1 > nnodes2) {
crm_trace("Assign %s (active on %d) after %s (active on %d): "
"more multiply active",
instance1->id, nnodes1, instance2->id, nnodes2);
return 1;
}
}
/* An instance that is either inactive or active on an allowed node is
* preferred over an instance that is active on a no-longer-allowed node.
*/
can1 = node_is_allowed(instance1, &node1);
can2 = node_is_allowed(instance2, &node2);
if (can1 && !can2) {
crm_trace("Assign %s before %s: not active on a disallowed node",
instance1->id, instance2->id);
return -1;
} else if (!can1 && can2) {
crm_trace("Assign %s after %s: active on a disallowed node",
instance1->id, instance2->id);
return 1;
}
// Prefer instance with higher configured priority
if (instance1->priority > instance2->priority) {
crm_trace("Assign %s before %s: priority (%d > %d)",
instance1->id, instance2->id,
instance1->priority, instance2->priority);
return -1;
} else if (instance1->priority < instance2->priority) {
crm_trace("Assign %s after %s: priority (%d < %d)",
instance1->id, instance2->id,
instance1->priority, instance2->priority);
return 1;
}
// Prefer active instance
if ((node1 == NULL) && (node2 == NULL)) {
crm_trace("No assignment preference for %s vs. %s: inactive",
instance1->id, instance2->id);
return 0;
} else if (node1 == NULL) {
crm_trace("Assign %s after %s: active", instance1->id, instance2->id);
return 1;
} else if (node2 == NULL) {
crm_trace("Assign %s before %s: active", instance1->id, instance2->id);
return -1;
}
// Prefer instance whose current node can run resources
can1 = pcmk__node_available(node1, false, false);
can2 = pcmk__node_available(node2, false, false);
if (can1 && !can2) {
crm_trace("Assign %s before %s: current node can run resources",
instance1->id, instance2->id);
return -1;
} else if (!can1 && can2) {
crm_trace("Assign %s after %s: current node can't run resources",
instance1->id, instance2->id);
return 1;
}
// Prefer instance whose parent is allowed to run on instance's current node
node1 = pcmk__top_allowed_node(instance1, node1);
node2 = pcmk__top_allowed_node(instance2, node2);
if ((node1 == NULL) && (node2 == NULL)) {
crm_trace("No assignment preference for %s vs. %s: "
"parent not allowed on either instance's current node",
instance1->id, instance2->id);
return 0;
} else if (node1 == NULL) {
crm_trace("Assign %s after %s: parent not allowed on current node",
instance1->id, instance2->id);
return 1;
} else if (node2 == NULL) {
crm_trace("Assign %s before %s: parent allowed on current node",
instance1->id, instance2->id);
return -1;
}
// Prefer instance whose current node is running fewer other instances
if (node1->count < node2->count) {
crm_trace("Assign %s before %s: fewer active instances on current node",
instance1->id, instance2->id);
return -1;
} else if (node1->count > node2->count) {
crm_trace("Assign %s after %s: more active instances on current node",
instance1->id, instance2->id);
return 1;
}
// Prefer instance that isn't failed
can1 = did_fail(instance1);
can2 = did_fail(instance2);
if (!can1 && can2) {
crm_trace("Assign %s before %s: not failed",
instance1->id, instance2->id);
return -1;
} else if (can1 && !can2) {
crm_trace("Assign %s after %s: failed",
instance1->id, instance2->id);
return 1;
}
// Prefer instance with higher cumulative colocation score on current node
rc = cmp_instance_by_colocation(instance1, instance2);
if (rc != 0) {
return rc;
}
// Prefer instance with lower instance number
rc = pcmk__cmp_instance_number(instance1, instance2);
if (rc < 0) {
crm_trace("Assign %s before %s: instance number",
instance1->id, instance2->id);
} else if (rc > 0) {
crm_trace("Assign %s after %s: instance number",
instance1->id, instance2->id);
} else {
crm_trace("No assignment preference for %s vs. %s",
instance1->id, instance2->id);
}
return rc;
}
/*!
* \internal
* \brief Increment the parent's instance count after assigning an instance
*
* An instance's parent tracks how many instances have been assigned to each
* node via its pe_node_t:count member. After assigning an instance to a node,
* find the corresponding node in the parent's allowed table and increment it.
*
* \param[in,out] instance Instance whose parent to update
* \param[in] assigned_to Node to which the instance was assigned
*/
static void
increment_parent_count(pe_resource_t *instance, const pe_node_t *assigned_to)
{
pe_node_t *allowed = NULL;
if (assigned_to == NULL) {
return;
}
allowed = pcmk__top_allowed_node(instance, assigned_to);
if (allowed == NULL) {
/* The instance is allowed on the node, but its parent isn't. This
* shouldn't be possible if the resource is managed, and we won't be
* able to limit the number of instances assigned to the node.
*/
CRM_LOG_ASSERT(!pcmk_is_set(instance->flags, pe_rsc_managed));
} else {
allowed->count++;
}
}
/*!
* \internal
* \brief Assign an instance to a node
*
* \param[in,out] instance Clone instance or bundle replica container
* \param[in] prefer If not NULL, attempt early assignment to this
* node, if still the best choice; otherwise,
* perform final assignment
* \param[in] max_per_node Assign at most this many instances to one node
*
* \return Node to which \p instance is assigned
*/
static const pe_node_t *
assign_instance(pe_resource_t *instance, const pe_node_t *prefer,
int max_per_node)
{
pe_node_t *chosen = NULL;
pe_rsc_trace(instance, "Assigning %s (preferring %s)", instance->id,
((prefer == NULL)? "no node" : prefer->details->uname));
if (pcmk_is_set(instance->flags, pe_rsc_allocating)) {
pe_rsc_debug(instance,
"Assignment loop detected involving %s colocations",
instance->id);
return NULL;
}
ban_unavailable_allowed_nodes(instance, max_per_node);
// Failed early assignments are reversible (stop_if_fail=false)
chosen = instance->cmds->assign(instance, prefer, (prefer == NULL));
increment_parent_count(instance, chosen);
return chosen;
}
/*!
* \internal
* \brief Try to assign an instance to its current node early
*
* \param[in] rsc Clone or bundle being assigned (for logs only)
* \param[in] instance Clone instance or bundle replica container
* \param[in] current Instance's current node
* \param[in] max_per_node Maximum number of instances per node
* \param[in] available Number of instances still available for assignment
*
* \return \c true if \p instance was successfully assigned to its current node,
* or \c false otherwise
*/
static bool
assign_instance_early(const pe_resource_t *rsc, pe_resource_t *instance,
const pe_node_t *current, int max_per_node, int available)
{
const pe_node_t *chosen = NULL;
int reserved = 0;
pe_resource_t *parent = instance->parent;
GHashTable *allowed_orig = NULL;
GHashTable *allowed_orig_parent = parent->allowed_nodes;
const pe_node_t *allowed_node = g_hash_table_lookup(instance->allowed_nodes,
current->details->id);
pe_rsc_trace(instance, "Trying to assign %s to its current node %s",
instance->id, pe__node_name(current));
if (!pcmk__node_available(allowed_node, true, false)) {
pe_rsc_info(instance,
"Not assigning %s to current node %s: unavailable",
instance->id, pe__node_name(current));
return false;
}
/* On each iteration, if instance gets assigned to a node other than its
* current one, we reserve one instance for the chosen node, unassign
* instance, restore instance's original node tables, and try again. This
* way, instances are proportionally assigned to nodes based on preferences,
* but shuffling of specific instances is minimized. If a node will be
* assigned instances at all, it preferentially receives instances that are
* currently active there.
*
* parent->allowed_nodes tracks the number of instances assigned to each
* node. If a node already has max_per_node instances assigned,
* ban_unavailable_allowed_nodes() marks it as unavailable.
*
* In the end, we restore the original parent->allowed_nodes to undo the
* changes to counts during tentative assignments. If we successfully
* assigned instance to its current node, we increment that node's counter.
*/
// Back up the allowed node tables of instance and its children recursively
pcmk__copy_node_tables(instance, &allowed_orig);
// Update instances-per-node counts in a scratch table
parent->allowed_nodes = pcmk__copy_node_table(parent->allowed_nodes);
while (reserved < available) {
chosen = assign_instance(instance, current, max_per_node);
if (pe__same_node(chosen, current)) {
// Successfully assigned to current node
break;
}
// Assignment updates scores, so restore to original state
pe_rsc_debug(instance, "Rolling back node scores for %s", instance->id);
pcmk__restore_node_tables(instance, allowed_orig);
if (chosen == NULL) {
// Assignment failed, so give up
pe_rsc_info(instance,
"Not assigning %s to current node %s: unavailable",
instance->id, pe__node_name(current));
pe__set_resource_flags(instance, pe_rsc_provisional);
break;
}
// We prefer more strongly to assign an instance to the chosen node
pe_rsc_debug(instance,
"Not assigning %s to current node %s: %s is better",
instance->id, pe__node_name(current),
pe__node_name(chosen));
// Reserve one instance for the chosen node and try again
if (++reserved >= available) {
pe_rsc_info(instance,
"Not assigning %s to current node %s: "
"other assignments are more important",
instance->id, pe__node_name(current));
} else {
pe_rsc_debug(instance,
"Reserved an instance of %s for %s. Retrying "
"assignment of %s to %s",
rsc->id, pe__node_name(chosen), instance->id,
pe__node_name(current));
}
// Clear this assignment (frees chosen); leave instance counts in parent
pcmk__unassign_resource(instance);
chosen = NULL;
}
g_hash_table_destroy(allowed_orig);
// Restore original instances-per-node counts
g_hash_table_destroy(parent->allowed_nodes);
parent->allowed_nodes = allowed_orig_parent;
if (chosen == NULL) {
// Couldn't assign instance to current node
return false;
}
pe_rsc_trace(instance, "Assigned %s to current node %s",
instance->id, pe__node_name(current));
increment_parent_count(instance, chosen);
return true;
}
/*!
* \internal
* \brief Reset the node counts of a resource's allowed nodes to zero
*
* \param[in,out] rsc Resource to reset
*
* \return Number of nodes that are available to run resources
*/
static unsigned int
reset_allowed_node_counts(pe_resource_t *rsc)
{
unsigned int available_nodes = 0;
pe_node_t *node = NULL;
GHashTableIter iter;
g_hash_table_iter_init(&iter, rsc->allowed_nodes);
while (g_hash_table_iter_next(&iter, NULL, (gpointer *) &node)) {
node->count = 0;
if (pcmk__node_available(node, false, false)) {
available_nodes++;
}
}
return available_nodes;
}
/*!
* \internal
* \brief Check whether an instance has a preferred node
*
* \param[in] instance Clone instance or bundle replica container
* \param[in] optimal_per_node Optimal number of instances per node
*
* \return Instance's current node if still available, otherwise NULL
*/
static const pe_node_t *
preferred_node(const pe_resource_t *instance, int optimal_per_node)
{
const pe_node_t *node = NULL;
const pe_node_t *parent_node = NULL;
// Check whether instance is active, healthy, and not yet assigned
if ((instance->running_on == NULL)
|| !pcmk_is_set(instance->flags, pe_rsc_provisional)
|| pcmk_is_set(instance->flags, pe_rsc_failed)) {
return NULL;
}
// Check whether instance's current node can run resources
node = pe__current_node(instance);
if (!pcmk__node_available(node, true, false)) {
pe_rsc_trace(instance, "Not assigning %s to %s early (unavailable)",
instance->id, pe__node_name(node));
return NULL;
}
// Check whether node already has optimal number of instances assigned
parent_node = pcmk__top_allowed_node(instance, node);
if ((parent_node != NULL) && (parent_node->count >= optimal_per_node)) {
pe_rsc_trace(instance,
"Not assigning %s to %s early "
"(optimal instances already assigned)",
instance->id, pe__node_name(node));
return NULL;
}
return node;
}
/*!
* \internal
* \brief Assign collective instances to nodes
*
* \param[in,out] collective Clone or bundle resource being assigned
* \param[in,out] instances List of clone instances or bundle containers
* \param[in] max_total Maximum instances to assign in total
* \param[in] max_per_node Maximum instances to assign to any one node
*/
void
pcmk__assign_instances(pe_resource_t *collective, GList *instances,
int max_total, int max_per_node)
{
// Reuse node count to track number of assigned instances
unsigned int available_nodes = reset_allowed_node_counts(collective);
int optimal_per_node = 0;
int assigned = 0;
GList *iter = NULL;
pe_resource_t *instance = NULL;
const pe_node_t *current = NULL;
if (available_nodes > 0) {
optimal_per_node = max_total / available_nodes;
}
if (optimal_per_node < 1) {
optimal_per_node = 1;
}
pe_rsc_debug(collective,
"Assigning up to %d %s instance%s to up to %u node%s "
"(at most %d per host, %d optimal)",
max_total, collective->id, pcmk__plural_s(max_total),
available_nodes, pcmk__plural_s(available_nodes),
max_per_node, optimal_per_node);
// Assign as many instances as possible to their current location
for (iter = instances; (iter != NULL) && (assigned < max_total);
iter = iter->next) {
int available = max_total - assigned;
instance = iter->data;
if (!pcmk_is_set(instance->flags, pe_rsc_provisional)) {
continue; // Already assigned
}
current = preferred_node(instance, optimal_per_node);
if ((current != NULL)
&& assign_instance_early(collective, instance, current,
max_per_node, available)) {
assigned++;
}
}
pe_rsc_trace(collective, "Assigned %d of %d instance%s to current node",
assigned, max_total, pcmk__plural_s(max_total));
for (iter = instances; iter != NULL; iter = iter->next) {
instance = (pe_resource_t *) iter->data;
if (!pcmk_is_set(instance->flags, pe_rsc_provisional)) {
continue; // Already assigned
}
if (instance->running_on != NULL) {
current = pe__current_node(instance);
if (pcmk__top_allowed_node(instance, current) == NULL) {
const char *unmanaged = "";
if (!pcmk_is_set(instance->flags, pe_rsc_managed)) {
unmanaged = "Unmanaged resource ";
}
crm_notice("%s%s is running on %s which is no longer allowed",
unmanaged, instance->id, pe__node_name(current));
}
}
if (assigned >= max_total) {
pe_rsc_debug(collective,
"Not assigning %s because maximum %d instances "
"already assigned",
instance->id, max_total);
resource_location(instance, NULL, -INFINITY,
"collective_limit_reached", collective->cluster);
} else if (assign_instance(instance, NULL, max_per_node) != NULL) {
assigned++;
}
}
pe_rsc_debug(collective, "Assigned %d of %d possible instance%s of %s",
assigned, max_total, pcmk__plural_s(max_total),
collective->id);
}
enum instance_state {
instance_starting = (1 << 0),
instance_stopping = (1 << 1),
/* This indicates that some instance is restarting. It's not the same as
* instance_starting|instance_stopping, which would indicate that some
* instance is starting, and some instance (not necessarily the same one) is
* stopping.
*/
instance_restarting = (1 << 2),
instance_active = (1 << 3),
instance_all = instance_starting|instance_stopping
|instance_restarting|instance_active,
};
/*!
* \internal
* \brief Check whether an instance is active, starting, and/or stopping
*
* \param[in] instance Clone instance or bundle replica container
* \param[in,out] state Whether any instance is starting, stopping, etc.
*/
static void
check_instance_state(const pe_resource_t *instance, uint32_t *state)
{
const GList *iter = NULL;
uint32_t instance_state = 0; // State of just this instance
// No need to check further if all conditions have already been detected
if (pcmk_all_flags_set(*state, instance_all)) {
return;
}
// If instance is a collective (a cloned group), check its children instead
if (instance->variant > pe_native) {
for (iter = instance->children;
(iter != NULL) && !pcmk_all_flags_set(*state, instance_all);
iter = iter->next) {
check_instance_state((const pe_resource_t *) iter->data, state);
}
return;
}
// If we get here, instance is a primitive
if (instance->running_on != NULL) {
instance_state |= instance_active;
}
// Check each of the instance's actions for runnable start or stop
for (iter = instance->actions;
(iter != NULL) && !pcmk_all_flags_set(instance_state,
instance_starting
|instance_stopping);
iter = iter->next) {
const pe_action_t *action = (const pe_action_t *) iter->data;
const bool optional = pcmk_is_set(action->flags, pe_action_optional);
if (pcmk__str_eq(PCMK_ACTION_START, action->task, pcmk__str_none)) {
if (!optional && pcmk_is_set(action->flags, pe_action_runnable)) {
pe_rsc_trace(instance, "Instance is starting due to %s",
action->uuid);
instance_state |= instance_starting;
} else {
pe_rsc_trace(instance, "%s doesn't affect %s state (%s)",
action->uuid, instance->id,
(optional? "optional" : "unrunnable"));
}
} else if (pcmk__str_eq(PCMK_ACTION_STOP, action->task,
pcmk__str_none)) {
/* Only stop actions can be pseudo-actions for primitives. That
* indicates that the node they are on is being fenced, so the stop
* is implied rather than actually executed.
*/
if (!optional
&& pcmk_any_flags_set(action->flags,
pe_action_pseudo|pe_action_runnable)) {
pe_rsc_trace(instance, "Instance is stopping due to %s",
action->uuid);
instance_state |= instance_stopping;
} else {
pe_rsc_trace(instance, "%s doesn't affect %s state (%s)",
action->uuid, instance->id,
(optional? "optional" : "unrunnable"));
}
}
}
if (pcmk_all_flags_set(instance_state,
instance_starting|instance_stopping)) {
instance_state |= instance_restarting;
}
*state |= instance_state;
}
/*!
* \internal
* \brief Create actions for collective resource instances
*
* \param[in,out] collective Clone or bundle resource to create actions for
* \param[in,out] instances List of clone instances or bundle containers
*/
void
pcmk__create_instance_actions(pe_resource_t *collective, GList *instances)
{
uint32_t state = 0;
pe_action_t *stop = NULL;
pe_action_t *stopped = NULL;
pe_action_t *start = NULL;
pe_action_t *started = NULL;
pe_rsc_trace(collective, "Creating collective instance actions for %s",
collective->id);
// Create actions for each instance appropriate to its variant
for (GList *iter = instances; iter != NULL; iter = iter->next) {
pe_resource_t *instance = (pe_resource_t *) iter->data;
instance->cmds->create_actions(instance);
check_instance_state(instance, &state);
}
// Create pseudo-actions for rsc start and started
start = pe__new_rsc_pseudo_action(collective, PCMK_ACTION_START,
!pcmk_is_set(state, instance_starting),
true);
started = pe__new_rsc_pseudo_action(collective, PCMK_ACTION_RUNNING,
!pcmk_is_set(state, instance_starting),
false);
started->priority = INFINITY;
if (pcmk_any_flags_set(state, instance_active|instance_starting)) {
pe__set_action_flags(started, pe_action_runnable);
}
// Create pseudo-actions for rsc stop and stopped
stop = pe__new_rsc_pseudo_action(collective, PCMK_ACTION_STOP,
!pcmk_is_set(state, instance_stopping),
true);
stopped = pe__new_rsc_pseudo_action(collective, PCMK_ACTION_STOPPED,
!pcmk_is_set(state, instance_stopping),
true);
stopped->priority = INFINITY;
if (!pcmk_is_set(state, instance_restarting)) {
pe__set_action_flags(stop, pe_action_migrate_runnable);
}
if (collective->variant == pe_clone) {
pe__create_clone_notif_pseudo_ops(collective, start, started, stop,
stopped);
}
}
/*!
* \internal
* \brief Get a list of clone instances or bundle replica containers
*
* \param[in] rsc Clone or bundle resource
*
* \return Clone instances if \p rsc is a clone, or a newly created list of
* \p rsc's replica containers if \p rsc is a bundle
* \note The caller must call free_instance_list() on the result when the list
* is no longer needed.
*/
static inline GList *
get_instance_list(const pe_resource_t *rsc)
{
if (rsc->variant == pe_container) {
return pe__bundle_containers(rsc);
} else {
return rsc->children;
}
}
/*!
* \internal
* \brief Free any memory created by get_instance_list()
*
* \param[in] rsc Clone or bundle resource passed to get_instance_list()
* \param[in,out] list Return value of get_instance_list() for \p rsc
*/
static inline void
free_instance_list(const pe_resource_t *rsc, GList *list)
{
if (list != rsc->children) {
g_list_free(list);
}
}
/*!
* \internal
* \brief Check whether an instance is compatible with a role and node
*
* \param[in] instance Clone instance or bundle replica container
* \param[in] node Instance must match this node
* \param[in] role If not RSC_ROLE_UNKNOWN, instance must match this role
* \param[in] current If true, compare instance's original node and role,
* otherwise compare assigned next node and role
*
* \return true if \p instance is compatible with \p node and \p role,
* otherwise false
*/
bool
pcmk__instance_matches(const pe_resource_t *instance, const pe_node_t *node,
enum rsc_role_e role, bool current)
{
pe_node_t *instance_node = NULL;
CRM_CHECK((instance != NULL) && (node != NULL), return false);
if ((role != RSC_ROLE_UNKNOWN)
&& (role != instance->fns->state(instance, current))) {
pe_rsc_trace(instance,
"%s is not a compatible instance (role is not %s)",
instance->id, role2text(role));
return false;
}
if (!is_set_recursive(instance, pe_rsc_block, true)) {
// We only want instances that haven't failed
instance_node = instance->fns->location(instance, NULL, current);
}
if (instance_node == NULL) {
pe_rsc_trace(instance,
"%s is not a compatible instance (not assigned to a node)",
instance->id);
return false;
}
if (!pe__same_node(instance_node, node)) {
pe_rsc_trace(instance,
"%s is not a compatible instance (assigned to %s not %s)",
instance->id, pe__node_name(instance_node),
pe__node_name(node));
return false;
}
return true;
}
/*!
* \internal
* \brief Find an instance that matches a given resource by node and role
*
* \param[in] match_rsc Resource that instance must match (for logging only)
* \param[in] rsc Clone or bundle resource to check for matching instance
* \param[in] node Instance must match this node
* \param[in] role If not RSC_ROLE_UNKNOWN, instance must match this role
* \param[in] current If true, compare instance's original node and role,
* otherwise compare assigned next node and role
*
* \return \p rsc instance matching \p node and \p role if any, otherwise NULL
*/
static pe_resource_t *
find_compatible_instance_on_node(const pe_resource_t *match_rsc,
const pe_resource_t *rsc,
const pe_node_t *node, enum rsc_role_e role,
bool current)
{
GList *instances = NULL;
instances = get_instance_list(rsc);
for (GList *iter = instances; iter != NULL; iter = iter->next) {
pe_resource_t *instance = (pe_resource_t *) iter->data;
if (pcmk__instance_matches(instance, node, role, current)) {
pe_rsc_trace(match_rsc,
"Found %s %s instance %s compatible with %s on %s",
role == RSC_ROLE_UNKNOWN? "matching" : role2text(role),
rsc->id, instance->id, match_rsc->id,
pe__node_name(node));
free_instance_list(rsc, instances); // Only frees list, not contents
return instance;
}
}
free_instance_list(rsc, instances);
pe_rsc_trace(match_rsc, "No %s %s instance found compatible with %s on %s",
((role == RSC_ROLE_UNKNOWN)? "matching" : role2text(role)),
rsc->id, match_rsc->id, pe__node_name(node));
return NULL;
}
/*!
* \internal
* \brief Find a clone instance or bundle container compatible with a resource
*
* \param[in] match_rsc Resource that instance must match
* \param[in] rsc Clone or bundle resource to check for matching instance
* \param[in] role If not RSC_ROLE_UNKNOWN, instance must match this role
* \param[in] current If true, compare instance's original node and role,
* otherwise compare assigned next node and role
*
* \return Compatible (by \p role and \p match_rsc location) instance of \p rsc
* if any, otherwise NULL
*/
pe_resource_t *
pcmk__find_compatible_instance(const pe_resource_t *match_rsc,
const pe_resource_t *rsc, enum rsc_role_e role,
bool current)
{
pe_resource_t *instance = NULL;
GList *nodes = NULL;
const pe_node_t *node = match_rsc->fns->location(match_rsc, NULL, current);
// If match_rsc has a node, check only that node
if (node != NULL) {
return find_compatible_instance_on_node(match_rsc, rsc, node, role,
current);
}
// Otherwise check for an instance matching any of match_rsc's allowed nodes
nodes = pcmk__sort_nodes(g_hash_table_get_values(match_rsc->allowed_nodes),
NULL);
for (GList *iter = nodes; (iter != NULL) && (instance == NULL);
iter = iter->next) {
instance = find_compatible_instance_on_node(match_rsc, rsc,
(pe_node_t *) iter->data,
role, current);
}
if (instance == NULL) {
pe_rsc_debug(rsc, "No %s instance found compatible with %s",
rsc->id, match_rsc->id);
}
g_list_free(nodes);
return instance;
}
/*!
* \internal
* \brief Unassign an instance if mandatory ordering has no interleave match
*
* \param[in] first 'First' action in an ordering
* \param[in] then 'Then' action in an ordering
* \param[in,out] then_instance 'Then' instance that has no interleave match
* \param[in] type Group of enum pe_ordering flags to apply
* \param[in] current If true, "then" action is stopped or demoted
*
* \return true if \p then_instance was unassigned, otherwise false
*/
static bool
unassign_if_mandatory(const pe_action_t *first, const pe_action_t *then,
pe_resource_t *then_instance, uint32_t type, bool current)
{
// Allow "then" instance to go down even without an interleave match
if (current) {
pe_rsc_trace(then->rsc,
"%s has no instance to order before stopping "
"or demoting %s",
first->rsc->id, then_instance->id);
/* If the "first" action must be runnable, but there is no "first"
* instance, the "then" instance must not be allowed to come up.
*/
} else if (pcmk_any_flags_set(type, pe_order_runnable_left
|pe_order_implies_then)) {
pe_rsc_info(then->rsc,
"Inhibiting %s from being active "
"because there is no %s instance to interleave",
then_instance->id, first->rsc->id);
return pcmk__assign_resource(then_instance, NULL, true, true);
}
return false;
}
/*!
* \internal
* \brief Find first matching action for a clone instance or bundle container
*
* \param[in] action Action in an interleaved ordering
* \param[in] instance Clone instance or bundle container being interleaved
* \param[in] action_name Action to look for
* \param[in] node If not NULL, require action to be on this node
* \param[in] for_first If true, \p instance is the 'first' resource in the
* ordering, otherwise it is the 'then' resource
*
* \return First action for \p instance (or in some cases if \p instance is a
* bundle container, its containerized resource) that matches
* \p action_name and \p node if any, otherwise NULL
*/
static pe_action_t *
find_instance_action(const pe_action_t *action, const pe_resource_t *instance,
const char *action_name, const pe_node_t *node,
bool for_first)
{
const pe_resource_t *rsc = NULL;
pe_action_t *matching_action = NULL;
/* If instance is a bundle container, sometimes we should interleave the
* action for the container itself, and sometimes for the containerized
* resource.
*
* For example, given "start bundle A then bundle B", B likely requires the
* service inside A's container to be active, rather than just the
* container, so we should interleave the action for A's containerized
* resource. On the other hand, it's possible B's container itself requires
* something from A, so we should interleave the action for B's container.
*
* Essentially, for 'first', we should use the containerized resource for
* everything except stop, and for 'then', we should use the container for
* everything except promote and demote (which can only be performed on the
* containerized resource).
*/
if ((for_first && !pcmk__str_any_of(action->task, PCMK_ACTION_STOP,
PCMK_ACTION_STOPPED, NULL))
|| (!for_first && pcmk__str_any_of(action->task, PCMK_ACTION_PROMOTE,
PCMK_ACTION_PROMOTED,
PCMK_ACTION_DEMOTE,
PCMK_ACTION_DEMOTED, NULL))) {
rsc = pe__get_rsc_in_container(instance);
}
if (rsc == NULL) {
rsc = instance; // No containerized resource, use instance itself
} else {
node = NULL; // Containerized actions are on bundle-created guest
}
matching_action = find_first_action(rsc->actions, NULL, action_name, node);
if (matching_action != NULL) {
return matching_action;
}
if (pcmk_is_set(instance->flags, pe_rsc_orphan)
|| pcmk__str_any_of(action_name, PCMK_ACTION_STOP, PCMK_ACTION_DEMOTE,
NULL)) {
crm_trace("No %s action found for %s%s",
action_name,
pcmk_is_set(instance->flags, pe_rsc_orphan)? "orphan " : "",
instance->id);
} else {
crm_err("No %s action found for %s to interleave (bug?)",
action_name, instance->id);
}
return NULL;
}
/*!
* \internal
* \brief Get the original action name of a bundle or clone action
*
* Given an action for a bundle or clone, get the original action name,
* mapping notify to the action being notified, and if the instances are
* primitives, mapping completion actions to the action that was completed
* (for example, stopped to stop).
*
* \param[in] action Clone or bundle action to check
*
* \return Original action name for \p action
*/
static const char *
orig_action_name(const pe_action_t *action)
{
const pe_resource_t *instance = action->rsc->children->data; // Any instance
char *action_type = NULL;
const char *action_name = action->task;
enum action_tasks orig_task = no_action;
if (pcmk__strcase_any_of(action->task, PCMK_ACTION_NOTIFY,
- CRMD_ACTION_NOTIFIED, NULL)) {
+ PCMK_ACTION_NOTIFIED, NULL)) {
// action->uuid is RSC_(confirmed-){pre,post}_notify_ACTION_INTERVAL
CRM_CHECK(parse_op_key(action->uuid, NULL, &action_type, NULL),
return task2text(no_action));
action_name = strstr(action_type, "_notify_");
CRM_CHECK(action_name != NULL, return task2text(no_action));
action_name += strlen("_notify_");
}
orig_task = get_complex_task(instance, action_name);
free(action_type);
return task2text(orig_task);
}
/*!
* \internal
* \brief Update two interleaved actions according to an ordering between them
*
* Given information about an ordering of two interleaved actions, update the
* actions' flags (and runnable_before members if appropriate) as appropriate
* for the ordering. Effects may cascade to other orderings involving the
* actions as well.
*
* \param[in,out] first 'First' action in an ordering
* \param[in,out] then 'Then' action in an ordering
* \param[in] node If not NULL, limit scope of ordering to this node
* \param[in] filter Action flags to limit scope of certain updates (may
* include pe_action_optional to affect only mandatory
* actions, and pe_action_runnable to affect only
* runnable actions)
* \param[in] type Group of enum pe_ordering flags to apply
*
* \return Group of enum pcmk__updated flags indicating what was updated
*/
static uint32_t
update_interleaved_actions(pe_action_t *first, pe_action_t *then,
const pe_node_t *node, uint32_t filter,
uint32_t type)
{
GList *instances = NULL;
uint32_t changed = pcmk__updated_none;
const char *orig_first_task = orig_action_name(first);
// Stops and demotes must be interleaved with instance on current node
bool current = pcmk__ends_with(first->uuid, "_" PCMK_ACTION_STOPPED "_0")
|| pcmk__ends_with(first->uuid,
"_" PCMK_ACTION_DEMOTED "_0");
// Update the specified actions for each "then" instance individually
instances = get_instance_list(then->rsc);
for (GList *iter = instances; iter != NULL; iter = iter->next) {
pe_resource_t *first_instance = NULL;
pe_resource_t *then_instance = iter->data;
pe_action_t *first_action = NULL;
pe_action_t *then_action = NULL;
// Find a "first" instance to interleave with this "then" instance
first_instance = pcmk__find_compatible_instance(then_instance,
first->rsc,
RSC_ROLE_UNKNOWN,
current);
if (first_instance == NULL) { // No instance can be interleaved
if (unassign_if_mandatory(first, then, then_instance, type,
current)) {
pcmk__set_updated_flags(changed, first, pcmk__updated_then);
}
continue;
}
first_action = find_instance_action(first, first_instance,
orig_first_task, node, true);
if (first_action == NULL) {
continue;
}
then_action = find_instance_action(then, then_instance, then->task,
node, false);
if (then_action == NULL) {
continue;
}
if (order_actions(first_action, then_action, type)) {
pcmk__set_updated_flags(changed, first,
pcmk__updated_first|pcmk__updated_then);
}
changed |= then_instance->cmds->update_ordered_actions(
first_action, then_action, node,
first_instance->cmds->action_flags(first_action, node), filter,
type, then->rsc->cluster);
}
free_instance_list(then->rsc, instances);
return changed;
}
/*!
* \internal
* \brief Check whether two actions in an ordering can be interleaved
*
* \param[in] first 'First' action in the ordering
* \param[in] then 'Then' action in the ordering
*
* \return true if \p first and \p then can be interleaved, otherwise false
*/
static bool
can_interleave_actions(const pe_action_t *first, const pe_action_t *then)
{
bool interleave = false;
pe_resource_t *rsc = NULL;
if ((first->rsc == NULL) || (then->rsc == NULL)) {
crm_trace("Not interleaving %s with %s: not resource actions",
first->uuid, then->uuid);
return false;
}
if (first->rsc == then->rsc) {
crm_trace("Not interleaving %s with %s: same resource",
first->uuid, then->uuid);
return false;
}
if ((first->rsc->variant < pe_clone) || (then->rsc->variant < pe_clone)) {
crm_trace("Not interleaving %s with %s: not clones or bundles",
first->uuid, then->uuid);
return false;
}
if (pcmk__ends_with(then->uuid, "_stop_0")
|| pcmk__ends_with(then->uuid, "_demote_0")) {
rsc = first->rsc;
} else {
rsc = then->rsc;
}
interleave = crm_is_true(g_hash_table_lookup(rsc->meta,
XML_RSC_ATTR_INTERLEAVE));
pe_rsc_trace(rsc, "'%s then %s' will %sbe interleaved (based on %s)",
first->uuid, then->uuid, (interleave? "" : "not "), rsc->id);
return interleave;
}
/*!
* \internal
* \brief Update non-interleaved instance actions according to an ordering
*
* Given information about an ordering of two non-interleaved actions, update
* the actions' flags (and runnable_before members if appropriate) as
* appropriate for the ordering. Effects may cascade to other orderings
* involving the actions as well.
*
* \param[in,out] instance Clone instance or bundle container
* \param[in,out] first "First" action in ordering
* \param[in] then "Then" action in ordering (for \p instance's parent)
* \param[in] node If not NULL, limit scope of ordering to this node
* \param[in] flags Action flags for \p first for ordering purposes
* \param[in] filter Action flags to limit scope of certain updates (may
* include pe_action_optional to affect only mandatory
* actions, and pe_action_runnable to affect only
* runnable actions)
* \param[in] type Group of enum pe_ordering flags to apply
*
* \return Group of enum pcmk__updated flags indicating what was updated
*/
static uint32_t
update_noninterleaved_actions(pe_resource_t *instance, pe_action_t *first,
const pe_action_t *then, const pe_node_t *node,
uint32_t flags, uint32_t filter, uint32_t type)
{
pe_action_t *instance_action = NULL;
uint32_t instance_flags = 0;
uint32_t changed = pcmk__updated_none;
// Check whether instance has an equivalent of "then" action
instance_action = find_first_action(instance->actions, NULL, then->task,
node);
if (instance_action == NULL) {
return changed;
}
// Check whether action is runnable
instance_flags = instance->cmds->action_flags(instance_action, node);
if (!pcmk_is_set(instance_flags, pe_action_runnable)) {
return changed;
}
// If so, update actions for the instance
changed = instance->cmds->update_ordered_actions(first, instance_action,
node, flags, filter, type,
instance->cluster);
// Propagate any changes to later actions
if (pcmk_is_set(changed, pcmk__updated_then)) {
for (GList *after_iter = instance_action->actions_after;
after_iter != NULL; after_iter = after_iter->next) {
pe_action_wrapper_t *after = after_iter->data;
pcmk__update_action_for_orderings(after->action, instance->cluster);
}
}
return changed;
}
/*!
* \internal
* \brief Update two actions according to an ordering between them
*
* Given information about an ordering of two clone or bundle actions, update
* the actions' flags (and runnable_before members if appropriate) as
* appropriate for the ordering. Effects may cascade to other orderings
* involving the actions as well.
*
* \param[in,out] first 'First' action in an ordering
* \param[in,out] then 'Then' action in an ordering
* \param[in] node If not NULL, limit scope of ordering to this node
* (only used when interleaving instances)
* \param[in] flags Action flags for \p first for ordering purposes
* \param[in] filter Action flags to limit scope of certain updates (may
* include pe_action_optional to affect only mandatory
* actions, and pe_action_runnable to affect only
* runnable actions)
* \param[in] type Group of enum pe_ordering flags to apply
* \param[in,out] data_set Cluster working set
*
* \return Group of enum pcmk__updated flags indicating what was updated
*/
uint32_t
pcmk__instance_update_ordered_actions(pe_action_t *first, pe_action_t *then,
const pe_node_t *node, uint32_t flags,
uint32_t filter, uint32_t type,
pe_working_set_t *data_set)
{
CRM_ASSERT((first != NULL) && (then != NULL) && (data_set != NULL));
if (then->rsc == NULL) {
return pcmk__updated_none;
} else if (can_interleave_actions(first, then)) {
return update_interleaved_actions(first, then, node, filter, type);
} else {
uint32_t changed = pcmk__updated_none;
GList *instances = get_instance_list(then->rsc);
// Update actions for the clone or bundle resource itself
changed |= pcmk__update_ordered_actions(first, then, node, flags,
filter, type, data_set);
// Update the 'then' clone instances or bundle containers individually
for (GList *iter = instances; iter != NULL; iter = iter->next) {
pe_resource_t *instance = iter->data;
changed |= update_noninterleaved_actions(instance, first, then,
node, flags, filter, type);
}
free_instance_list(then->rsc, instances);
return changed;
}
}
#define pe__clear_action_summary_flags(flags, action, flag) do { \
flags = pcmk__clear_flags_as(__func__, __LINE__, LOG_TRACE, \
"Action summary", action->rsc->id, \
flags, flag, #flag); \
} while (0)
/*!
* \internal
* \brief Return action flags for a given clone or bundle action
*
* \param[in,out] action Action for a clone or bundle
* \param[in] instances Clone instances or bundle containers
* \param[in] node If not NULL, limit effects to this node
*
* \return Flags appropriate to \p action on \p node
*/
uint32_t
pcmk__collective_action_flags(pe_action_t *action, const GList *instances,
const pe_node_t *node)
{
bool any_runnable = false;
const char *action_name = orig_action_name(action);
// Set original assumptions (optional and runnable may be cleared below)
uint32_t flags = pe_action_optional|pe_action_runnable|pe_action_pseudo;
for (const GList *iter = instances; iter != NULL; iter = iter->next) {
const pe_resource_t *instance = iter->data;
const pe_node_t *instance_node = NULL;
pe_action_t *instance_action = NULL;
uint32_t instance_flags;
// Node is relevant only to primitive instances
if (instance->variant == pe_native) {
instance_node = node;
}
instance_action = find_first_action(instance->actions, NULL,
action_name, instance_node);
if (instance_action == NULL) {
pe_rsc_trace(action->rsc, "%s has no %s action on %s",
instance->id, action_name, pe__node_name(node));
continue;
}
pe_rsc_trace(action->rsc, "%s has %s for %s on %s",
instance->id, instance_action->uuid, action_name,
pe__node_name(node));
instance_flags = instance->cmds->action_flags(instance_action, node);
// If any instance action is mandatory, so is the collective action
if (pcmk_is_set(flags, pe_action_optional)
&& !pcmk_is_set(instance_flags, pe_action_optional)) {
pe_rsc_trace(instance, "%s is mandatory because %s is",
action->uuid, instance_action->uuid);
pe__clear_action_summary_flags(flags, action, pe_action_optional);
pe__clear_action_flags(action, pe_action_optional);
}
// If any instance action is runnable, so is the collective action
if (pcmk_is_set(instance_flags, pe_action_runnable)) {
any_runnable = true;
}
}
if (!any_runnable) {
pe_rsc_trace(action->rsc,
"%s is not runnable because no instance can run %s",
action->uuid, action_name);
pe__clear_action_summary_flags(flags, action, pe_action_runnable);
if (node == NULL) {
pe__clear_action_flags(action, pe_action_runnable);
}
}
return flags;
}
diff --git a/lib/pacemaker/pcmk_simulate.c b/lib/pacemaker/pcmk_simulate.c
index 90d8533f1d..61ba443f07 100644
--- a/lib/pacemaker/pcmk_simulate.c
+++ b/lib/pacemaker/pcmk_simulate.c
@@ -1,1003 +1,1003 @@
/*
* Copyright 2021-2023 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
* This source code is licensed under the GNU Lesser General Public License
* version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY.
*/
#include <crm_internal.h>
#include <crm/cib/internal.h>
#include <crm/common/output.h>
#include <crm/common/results.h>
#include <crm/pengine/pe_types.h>
#include <pacemaker-internal.h>
#include <pacemaker.h>
#include <stdint.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <unistd.h>
#include "libpacemaker_private.h"
static pcmk__output_t *out = NULL;
static cib_t *fake_cib = NULL;
static GList *fake_resource_list = NULL;
static const GList *fake_op_fail_list = NULL;
static void set_effective_date(pe_working_set_t *data_set, bool print_original,
const char *use_date);
/*!
* \internal
* \brief Create an action name for use in a dot graph
*
* \param[in] action Action to create name for
* \param[in] verbose If true, add action ID to name
*
* \return Newly allocated string with action name
* \note It is the caller's responsibility to free the result.
*/
static char *
create_action_name(const pe_action_t *action, bool verbose)
{
char *action_name = NULL;
const char *prefix = "";
const char *action_host = NULL;
const char *clone_name = NULL;
const char *task = action->task;
if (action->node != NULL) {
action_host = action->node->details->uname;
} else if (!pcmk_is_set(action->flags, pe_action_pseudo)) {
action_host = "<none>";
}
if (pcmk__str_eq(action->task, PCMK_ACTION_CANCEL, pcmk__str_none)) {
prefix = "Cancel ";
task = action->cancel_task;
}
if (action->rsc != NULL) {
clone_name = action->rsc->clone_name;
}
if (clone_name != NULL) {
char *key = NULL;
guint interval_ms = 0;
if (pcmk__guint_from_hash(action->meta,
XML_LRM_ATTR_INTERVAL_MS, 0,
&interval_ms) != pcmk_rc_ok) {
interval_ms = 0;
}
- if (pcmk__strcase_any_of(action->task, PCMK_ACTION_NOTIFY, RSC_NOTIFIED,
- NULL)) {
+ if (pcmk__strcase_any_of(action->task, PCMK_ACTION_NOTIFY,
+ PCMK_ACTION_NOTIFIED, NULL)) {
const char *n_type = g_hash_table_lookup(action->meta,
"notify_key_type");
const char *n_task = g_hash_table_lookup(action->meta,
"notify_key_operation");
CRM_ASSERT(n_type != NULL);
CRM_ASSERT(n_task != NULL);
key = pcmk__notify_key(clone_name, n_type, n_task);
} else {
key = pcmk__op_key(clone_name, task, interval_ms);
}
if (action_host != NULL) {
action_name = crm_strdup_printf("%s%s %s",
prefix, key, action_host);
} else {
action_name = crm_strdup_printf("%s%s", prefix, key);
}
free(key);
} else if (pcmk__str_eq(action->task, CRM_OP_FENCE, pcmk__str_none)) {
const char *op = g_hash_table_lookup(action->meta, "stonith_action");
action_name = crm_strdup_printf("%s%s '%s' %s",
prefix, action->task, op, action_host);
} else if (action->rsc && action_host) {
action_name = crm_strdup_printf("%s%s %s",
prefix, action->uuid, action_host);
} else if (action_host) {
action_name = crm_strdup_printf("%s%s %s",
prefix, action->task, action_host);
} else {
action_name = crm_strdup_printf("%s", action->uuid);
}
if (verbose) {
char *with_id = crm_strdup_printf("%s (%d)", action_name, action->id);
free(action_name);
action_name = with_id;
}
return action_name;
}
/*!
* \internal
* \brief Display the status of a cluster
*
* \param[in,out] data_set Cluster working set
* \param[in] show_opts How to modify display (as pcmk_show_opt_e flags)
* \param[in] section_opts Sections to display (as pcmk_section_e flags)
* \param[in] title What to use as list title
* \param[in] print_spacer Whether to display a spacer first
*/
static void
print_cluster_status(pe_working_set_t *data_set, uint32_t show_opts,
uint32_t section_opts, const char *title,
bool print_spacer)
{
pcmk__output_t *out = data_set->priv;
GList *all = NULL;
crm_exit_t stonith_rc = 0;
enum pcmk_pacemakerd_state state = pcmk_pacemakerd_state_invalid;
section_opts |= pcmk_section_nodes | pcmk_section_resources;
show_opts |= pcmk_show_inactive_rscs | pcmk_show_failed_detail;
all = g_list_prepend(all, (gpointer) "*");
PCMK__OUTPUT_SPACER_IF(out, print_spacer);
out->begin_list(out, NULL, NULL, "%s", title);
out->message(out, "cluster-status",
data_set, state, stonith_rc, NULL,
false, section_opts, show_opts, NULL, all, all);
out->end_list(out);
g_list_free(all);
}
/*!
* \internal
* \brief Display a summary of all actions scheduled in a transition
*
* \param[in,out] data_set Cluster working set (fully scheduled)
* \param[in] print_spacer Whether to display a spacer first
*/
static void
print_transition_summary(pe_working_set_t *data_set, bool print_spacer)
{
pcmk__output_t *out = data_set->priv;
PCMK__OUTPUT_SPACER_IF(out, print_spacer);
out->begin_list(out, NULL, NULL, "Transition Summary");
pcmk__output_actions(data_set);
out->end_list(out);
}
/*!
* \internal
* \brief Reset a cluster working set's input, output, date, and flags
*
* \param[in,out] data_set Cluster working set
* \param[in] input What to set as cluster input
* \param[in] out What to set as cluster output object
* \param[in] use_date What to set as cluster's current timestamp
* \param[in] flags Cluster flags to add (pe_flag_*)
*/
static void
reset(pe_working_set_t *data_set, xmlNodePtr input, pcmk__output_t *out,
const char *use_date, unsigned int flags)
{
data_set->input = input;
data_set->priv = out;
set_effective_date(data_set, true, use_date);
if (pcmk_is_set(flags, pcmk_sim_sanitized)) {
pe__set_working_set_flags(data_set, pe_flag_sanitized);
}
if (pcmk_is_set(flags, pcmk_sim_show_scores)) {
pe__set_working_set_flags(data_set, pe_flag_show_scores);
}
if (pcmk_is_set(flags, pcmk_sim_show_utilization)) {
pe__set_working_set_flags(data_set, pe_flag_show_utilization);
}
}
/*!
* \brief Write out a file in dot(1) format describing the actions that will
* be taken by the scheduler in response to an input CIB file.
*
* \param[in,out] data_set Working set for the cluster
* \param[in] dot_file The filename to write
* \param[in] all_actions Write all actions, even those that are optional
* or are on unmanaged resources
* \param[in] verbose Add extra information, such as action IDs, to the
* output
*
* \return Standard Pacemaker return code
*/
static int
write_sim_dotfile(pe_working_set_t *data_set, const char *dot_file,
bool all_actions, bool verbose)
{
GList *iter = NULL;
FILE *dot_strm = fopen(dot_file, "w");
if (dot_strm == NULL) {
return errno;
}
fprintf(dot_strm, " digraph \"g\" {\n");
for (iter = data_set->actions; iter != NULL; iter = iter->next) {
pe_action_t *action = (pe_action_t *) iter->data;
const char *style = "dashed";
const char *font = "black";
const char *color = "black";
char *action_name = create_action_name(action, verbose);
if (pcmk_is_set(action->flags, pe_action_pseudo)) {
font = "orange";
}
if (pcmk_is_set(action->flags, pe_action_dumped)) {
style = "bold";
color = "green";
} else if ((action->rsc != NULL)
&& !pcmk_is_set(action->rsc->flags, pe_rsc_managed)) {
color = "red";
font = "purple";
if (!all_actions) {
goto do_not_write;
}
} else if (pcmk_is_set(action->flags, pe_action_optional)) {
color = "blue";
if (!all_actions) {
goto do_not_write;
}
} else {
color = "red";
CRM_LOG_ASSERT(!pcmk_is_set(action->flags, pe_action_runnable));
}
pe__set_action_flags(action, pe_action_dumped);
fprintf(dot_strm, "\"%s\" [ style=%s color=\"%s\" fontcolor=\"%s\"]\n",
action_name, style, color, font);
do_not_write:
free(action_name);
}
for (iter = data_set->actions; iter != NULL; iter = iter->next) {
pe_action_t *action = (pe_action_t *) iter->data;
for (GList *before_iter = action->actions_before;
before_iter != NULL; before_iter = before_iter->next) {
pe_action_wrapper_t *before = before_iter->data;
char *before_name = NULL;
char *after_name = NULL;
const char *style = "dashed";
bool optional = true;
if (before->state == pe_link_dumped) {
optional = false;
style = "bold";
} else if (before->type == pe_order_none) {
continue;
} else if (pcmk_is_set(before->action->flags, pe_action_dumped)
&& pcmk_is_set(action->flags, pe_action_dumped)
&& before->type != pe_order_load) {
optional = false;
}
if (all_actions || !optional) {
before_name = create_action_name(before->action, verbose);
after_name = create_action_name(action, verbose);
fprintf(dot_strm, "\"%s\" -> \"%s\" [ style = %s]\n",
before_name, after_name, style);
free(before_name);
free(after_name);
}
}
}
fprintf(dot_strm, "}\n");
fflush(dot_strm);
fclose(dot_strm);
return pcmk_rc_ok;
}
/*!
* \brief Profile the configuration updates and scheduler actions in a single
* CIB file, printing the profiling timings.
*
* \note \p data_set->priv must have been set to a valid \p pcmk__output_t
* object before this function is called.
*
* \param[in] xml_file The CIB file to profile
* \param[in] repeat Number of times to run
* \param[in,out] data_set Working set for the cluster
* \param[in] use_date The date to set the cluster's time to (may be NULL)
*/
static void
profile_file(const char *xml_file, long long repeat, pe_working_set_t *data_set,
const char *use_date)
{
pcmk__output_t *out = data_set->priv;
xmlNode *cib_object = NULL;
clock_t start = 0;
clock_t end;
unsigned long long data_set_flags = pe_flag_no_compat;
CRM_ASSERT(out != NULL);
cib_object = filename2xml(xml_file);
start = clock();
if (pcmk_find_cib_element(cib_object, XML_CIB_TAG_STATUS) == NULL) {
create_xml_node(cib_object, XML_CIB_TAG_STATUS);
}
if (cli_config_update(&cib_object, NULL, FALSE) == FALSE) {
free_xml(cib_object);
return;
}
if (validate_xml(cib_object, NULL, FALSE) != TRUE) {
free_xml(cib_object);
return;
}
if (pcmk_is_set(data_set->flags, pe_flag_show_scores)) {
data_set_flags |= pe_flag_show_scores;
}
if (pcmk_is_set(data_set->flags, pe_flag_show_utilization)) {
data_set_flags |= pe_flag_show_utilization;
}
for (int i = 0; i < repeat; ++i) {
xmlNode *input = (repeat == 1)? cib_object : copy_xml(cib_object);
data_set->input = input;
set_effective_date(data_set, false, use_date);
pcmk__schedule_actions(input, data_set_flags, data_set);
pe_reset_working_set(data_set);
}
end = clock();
out->message(out, "profile", xml_file, start, end);
}
void
pcmk__profile_dir(const char *dir, long long repeat, pe_working_set_t *data_set,
const char *use_date)
{
pcmk__output_t *out = data_set->priv;
struct dirent **namelist;
int file_num = scandir(dir, &namelist, 0, alphasort);
CRM_ASSERT(out != NULL);
if (file_num > 0) {
struct stat prop;
char buffer[FILENAME_MAX];
out->begin_list(out, NULL, NULL, "Timings");
while (file_num--) {
if ('.' == namelist[file_num]->d_name[0]) {
free(namelist[file_num]);
continue;
} else if (!pcmk__ends_with_ext(namelist[file_num]->d_name,
".xml")) {
free(namelist[file_num]);
continue;
}
snprintf(buffer, sizeof(buffer), "%s/%s",
dir, namelist[file_num]->d_name);
if (stat(buffer, &prop) == 0 && S_ISREG(prop.st_mode)) {
profile_file(buffer, repeat, data_set, use_date);
}
free(namelist[file_num]);
}
free(namelist);
out->end_list(out);
}
}
/*!
* \brief Set the date of the cluster, either to the value given by
* \p use_date, or to the "execution-date" value in the CIB.
*
* \note \p data_set->priv must have been set to a valid \p pcmk__output_t
* object before this function is called.
*
* \param[in,out] data_set Working set for the cluster
* \param[in] print_original If \p true, the "execution-date" should
* also be printed
* \param[in] use_date The date to set the cluster's time to
* (may be NULL)
*/
static void
set_effective_date(pe_working_set_t *data_set, bool print_original,
const char *use_date)
{
pcmk__output_t *out = data_set->priv;
time_t original_date = 0;
CRM_ASSERT(out != NULL);
crm_element_value_epoch(data_set->input, "execution-date", &original_date);
if (use_date) {
data_set->now = crm_time_new(use_date);
out->info(out, "Setting effective cluster time: %s", use_date);
crm_time_log(LOG_NOTICE, "Pretending 'now' is", data_set->now,
crm_time_log_date | crm_time_log_timeofday);
} else if (original_date != 0) {
data_set->now = pcmk__copy_timet(original_date);
if (print_original) {
char *when = crm_time_as_string(data_set->now,
crm_time_log_date|crm_time_log_timeofday);
out->info(out, "Using the original execution date of: %s", when);
free(when);
}
}
}
/*!
* \internal
* \brief Simulate successfully executing a pseudo-action in a graph
*
* \param[in,out] graph Graph to update with pseudo-action result
* \param[in,out] action Pseudo-action to simulate executing
*
* \return Standard Pacemaker return code
*/
static int
simulate_pseudo_action(pcmk__graph_t *graph, pcmk__graph_action_t *action)
{
const char *node = crm_element_value(action->xml, XML_LRM_ATTR_TARGET);
const char *task = crm_element_value(action->xml, XML_LRM_ATTR_TASK_KEY);
pcmk__set_graph_action_flags(action, pcmk__graph_action_confirmed);
out->message(out, "inject-pseudo-action", node, task);
pcmk__update_graph(graph, action);
return pcmk_rc_ok;
}
/*!
* \internal
* \brief Simulate executing a resource action in a graph
*
* \param[in,out] graph Graph to update with resource action result
* \param[in,out] action Resource action to simulate executing
*
* \return Standard Pacemaker return code
*/
static int
simulate_resource_action(pcmk__graph_t *graph, pcmk__graph_action_t *action)
{
int rc;
lrmd_event_data_t *op = NULL;
int target_outcome = PCMK_OCF_OK;
const char *rtype = NULL;
const char *rclass = NULL;
const char *resource = NULL;
const char *rprovider = NULL;
const char *resource_config_name = NULL;
const char *operation = crm_element_value(action->xml, "operation");
const char *target_rc_s = crm_meta_value(action->params,
XML_ATTR_TE_TARGET_RC);
xmlNode *cib_node = NULL;
xmlNode *cib_resource = NULL;
xmlNode *action_rsc = first_named_child(action->xml, XML_CIB_TAG_RESOURCE);
char *node = crm_element_value_copy(action->xml, XML_LRM_ATTR_TARGET);
char *uuid = NULL;
const char *router_node = crm_element_value(action->xml,
XML_LRM_ATTR_ROUTER_NODE);
// Certain actions don't need to be displayed or history entries
if (pcmk__str_eq(operation, CRM_OP_REPROBE, pcmk__str_none)) {
crm_debug("No history injection for %s op on %s", operation, node);
goto done; // Confirm action and update graph
}
if (action_rsc == NULL) { // Shouldn't be possible
crm_log_xml_err(action->xml, "Bad");
free(node);
return EPROTO;
}
/* A resource might be known by different names in the configuration and in
* the action (for example, a clone instance). Grab the configuration name
* (which is preferred when writing history), and if necessary, the instance
* name.
*/
resource_config_name = crm_element_value(action_rsc, XML_ATTR_ID);
if (resource_config_name == NULL) { // Shouldn't be possible
crm_log_xml_err(action->xml, "No ID");
free(node);
return EPROTO;
}
resource = resource_config_name;
if (pe_find_resource(fake_resource_list, resource) == NULL) {
const char *longname = crm_element_value(action_rsc, XML_ATTR_ID_LONG);
if ((longname != NULL)
&& (pe_find_resource(fake_resource_list, longname) != NULL)) {
resource = longname;
}
}
// Certain actions need to be displayed but don't need history entries
if (pcmk__strcase_any_of(operation, PCMK_ACTION_DELETE,
PCMK_ACTION_META_DATA, NULL)) {
out->message(out, "inject-rsc-action", resource, operation, node,
(guint) 0);
goto done; // Confirm action and update graph
}
rclass = crm_element_value(action_rsc, XML_AGENT_ATTR_CLASS);
rtype = crm_element_value(action_rsc, XML_ATTR_TYPE);
rprovider = crm_element_value(action_rsc, XML_AGENT_ATTR_PROVIDER);
pcmk__scan_min_int(target_rc_s, &target_outcome, 0);
CRM_ASSERT(fake_cib->cmds->query(fake_cib, NULL, NULL,
cib_sync_call|cib_scope_local) == pcmk_ok);
// Ensure the action node is in the CIB
uuid = crm_element_value_copy(action->xml, XML_LRM_ATTR_TARGET_UUID);
cib_node = pcmk__inject_node(fake_cib, node,
((router_node == NULL)? uuid: node));
free(uuid);
CRM_ASSERT(cib_node != NULL);
// Add a history entry for the action
cib_resource = pcmk__inject_resource_history(out, cib_node, resource,
resource_config_name,
rclass, rtype, rprovider);
if (cib_resource == NULL) {
crm_err("Could not simulate action %d history for resource %s",
action->id, resource);
free(node);
free_xml(cib_node);
return EINVAL;
}
// Simulate and display an executor event for the action result
op = pcmk__event_from_graph_action(cib_resource, action, PCMK_EXEC_DONE,
target_outcome, "User-injected result");
out->message(out, "inject-rsc-action", resource, op->op_type, node,
op->interval_ms);
// Check whether action is in a list of desired simulated failures
for (const GList *iter = fake_op_fail_list;
iter != NULL; iter = iter->next) {
const char *spec = (const char *) iter->data;
char *key = NULL;
const char *match_name = NULL;
// Allow user to specify anonymous clone with or without instance number
key = crm_strdup_printf(PCMK__OP_FMT "@%s=", resource, op->op_type,
op->interval_ms, node);
if (strncasecmp(key, spec, strlen(key)) == 0) {
match_name = resource;
}
free(key);
// If not found, try the resource's name in the configuration
if ((match_name == NULL)
&& (strcmp(resource, resource_config_name) != 0)) {
key = crm_strdup_printf(PCMK__OP_FMT "@%s=", resource_config_name,
op->op_type, op->interval_ms, node);
if (strncasecmp(key, spec, strlen(key)) == 0) {
match_name = resource_config_name;
}
free(key);
}
if (match_name == NULL) {
continue; // This failed action entry doesn't match
}
// ${match_name}_${task}_${interval_in_ms}@${node}=${rc}
rc = sscanf(spec, "%*[^=]=%d", (int *) &op->rc);
if (rc != 1) {
out->err(out, "Invalid failed operation '%s' "
"(result code must be integer)", spec);
continue; // Keep checking other list entries
}
out->info(out, "Pretending action %d failed with rc=%d",
action->id, op->rc);
pcmk__set_graph_action_flags(action, pcmk__graph_action_failed);
graph->abort_priority = INFINITY;
pcmk__inject_failcount(out, cib_node, match_name, op->op_type,
op->interval_ms, op->rc);
break;
}
pcmk__inject_action_result(cib_resource, op, target_outcome);
lrmd_free_event(op);
rc = fake_cib->cmds->modify(fake_cib, XML_CIB_TAG_STATUS, cib_node,
cib_sync_call|cib_scope_local);
CRM_ASSERT(rc == pcmk_ok);
done:
free(node);
free_xml(cib_node);
pcmk__set_graph_action_flags(action, pcmk__graph_action_confirmed);
pcmk__update_graph(graph, action);
return pcmk_rc_ok;
}
/*!
* \internal
* \brief Simulate successfully executing a cluster action
*
* \param[in,out] graph Graph to update with action result
* \param[in,out] action Cluster action to simulate
*
* \return Standard Pacemaker return code
*/
static int
simulate_cluster_action(pcmk__graph_t *graph, pcmk__graph_action_t *action)
{
const char *node = crm_element_value(action->xml, XML_LRM_ATTR_TARGET);
const char *task = crm_element_value(action->xml, XML_LRM_ATTR_TASK);
xmlNode *rsc = first_named_child(action->xml, XML_CIB_TAG_RESOURCE);
pcmk__set_graph_action_flags(action, pcmk__graph_action_confirmed);
out->message(out, "inject-cluster-action", node, task, rsc);
pcmk__update_graph(graph, action);
return pcmk_rc_ok;
}
/*!
* \internal
* \brief Simulate successfully executing a fencing action
*
* \param[in,out] graph Graph to update with action result
* \param[in,out] action Fencing action to simulate
*
* \return Standard Pacemaker return code
*/
static int
simulate_fencing_action(pcmk__graph_t *graph, pcmk__graph_action_t *action)
{
const char *op = crm_meta_value(action->params, "stonith_action");
char *target = crm_element_value_copy(action->xml, XML_LRM_ATTR_TARGET);
out->message(out, "inject-fencing-action", target, op);
if (!pcmk__str_eq(op, "on", pcmk__str_casei)) {
int rc = pcmk_ok;
GString *xpath = g_string_sized_new(512);
// Set node state to offline
xmlNode *cib_node = pcmk__inject_node_state_change(fake_cib, target,
false);
CRM_ASSERT(cib_node != NULL);
crm_xml_add(cib_node, XML_ATTR_ORIGIN, __func__);
rc = fake_cib->cmds->replace(fake_cib, XML_CIB_TAG_STATUS, cib_node,
cib_sync_call|cib_scope_local);
CRM_ASSERT(rc == pcmk_ok);
// Simulate controller clearing node's resource history and attributes
pcmk__g_strcat(xpath,
"//" XML_CIB_TAG_STATE
"[@" XML_ATTR_UNAME "='", target, "']/" XML_CIB_TAG_LRM,
NULL);
fake_cib->cmds->remove(fake_cib, (const char *) xpath->str, NULL,
cib_xpath|cib_sync_call|cib_scope_local);
g_string_truncate(xpath, 0);
pcmk__g_strcat(xpath,
"//" XML_CIB_TAG_STATE
"[@" XML_ATTR_UNAME "='", target, "']"
"/" XML_TAG_TRANSIENT_NODEATTRS, NULL);
fake_cib->cmds->remove(fake_cib, (const char *) xpath->str, NULL,
cib_xpath|cib_sync_call|cib_scope_local);
free_xml(cib_node);
g_string_free(xpath, TRUE);
}
pcmk__set_graph_action_flags(action, pcmk__graph_action_confirmed);
pcmk__update_graph(graph, action);
free(target);
return pcmk_rc_ok;
}
enum pcmk__graph_status
pcmk__simulate_transition(pe_working_set_t *data_set, cib_t *cib,
const GList *op_fail_list)
{
pcmk__graph_t *transition = NULL;
enum pcmk__graph_status graph_rc;
pcmk__graph_functions_t simulation_fns = {
simulate_pseudo_action,
simulate_resource_action,
simulate_cluster_action,
simulate_fencing_action,
};
out = data_set->priv;
fake_cib = cib;
fake_op_fail_list = op_fail_list;
if (!out->is_quiet(out)) {
out->begin_list(out, NULL, NULL, "Executing Cluster Transition");
}
pcmk__set_graph_functions(&simulation_fns);
transition = pcmk__unpack_graph(data_set->graph, crm_system_name);
pcmk__log_graph(LOG_DEBUG, transition);
fake_resource_list = data_set->resources;
do {
graph_rc = pcmk__execute_graph(transition);
} while (graph_rc == pcmk__graph_active);
fake_resource_list = NULL;
if (graph_rc != pcmk__graph_complete) {
out->err(out, "Transition failed: %s",
pcmk__graph_status2text(graph_rc));
pcmk__log_graph(LOG_ERR, transition);
out->err(out, "An invalid transition was produced");
}
pcmk__free_graph(transition);
if (!out->is_quiet(out)) {
// If not quiet, we'll need the resulting CIB for later display
xmlNode *cib_object = NULL;
int rc = fake_cib->cmds->query(fake_cib, NULL, &cib_object,
cib_sync_call|cib_scope_local);
CRM_ASSERT(rc == pcmk_ok);
pe_reset_working_set(data_set);
data_set->input = cib_object;
out->end_list(out);
}
return graph_rc;
}
int
pcmk__simulate(pe_working_set_t *data_set, pcmk__output_t *out,
const pcmk_injections_t *injections, unsigned int flags,
uint32_t section_opts, const char *use_date,
const char *input_file, const char *graph_file,
const char *dot_file)
{
int printed = pcmk_rc_no_output;
int rc = pcmk_rc_ok;
xmlNodePtr input = NULL;
cib_t *cib = NULL;
rc = cib__signon_query(out, &cib, &input);
if (rc != pcmk_rc_ok) {
goto simulate_done;
}
reset(data_set, input, out, use_date, flags);
cluster_status(data_set);
if ((cib->variant == cib_native)
&& pcmk_is_set(section_opts, pcmk_section_times)) {
if (pcmk__our_nodename == NULL) {
// Currently used only in the times section
pcmk__query_node_name(out, 0, &pcmk__our_nodename, 0);
}
data_set->localhost = pcmk__our_nodename;
}
if (!out->is_quiet(out)) {
const bool show_pending = pcmk_is_set(flags, pcmk_sim_show_pending);
if (pcmk_is_set(data_set->flags, pe_flag_maintenance_mode)) {
printed = out->message(out, "maint-mode", data_set->flags);
}
if (data_set->disabled_resources || data_set->blocked_resources) {
PCMK__OUTPUT_SPACER_IF(out, printed == pcmk_rc_ok);
printed = out->info(out,
"%d of %d resource instances DISABLED and "
"%d BLOCKED from further action due to failure",
data_set->disabled_resources,
data_set->ninstances,
data_set->blocked_resources);
}
/* Most formatted output headers use caps for each word, but this one
* only has the first word capitalized for compatibility with pcs.
*/
print_cluster_status(data_set, (show_pending? pcmk_show_pending : 0),
section_opts, "Current cluster status",
(printed == pcmk_rc_ok));
printed = pcmk_rc_ok;
}
// If the user requested any injections, handle them
if ((injections->node_down != NULL)
|| (injections->node_fail != NULL)
|| (injections->node_up != NULL)
|| (injections->op_inject != NULL)
|| (injections->ticket_activate != NULL)
|| (injections->ticket_grant != NULL)
|| (injections->ticket_revoke != NULL)
|| (injections->ticket_standby != NULL)
|| (injections->watchdog != NULL)) {
PCMK__OUTPUT_SPACER_IF(out, printed == pcmk_rc_ok);
pcmk__inject_scheduler_input(data_set, cib, injections);
printed = pcmk_rc_ok;
rc = cib->cmds->query(cib, NULL, &input, cib_sync_call);
if (rc != pcmk_rc_ok) {
rc = pcmk_legacy2rc(rc);
goto simulate_done;
}
cleanup_calculations(data_set);
reset(data_set, input, out, use_date, flags);
cluster_status(data_set);
}
if (input_file != NULL) {
rc = write_xml_file(input, input_file, FALSE);
if (rc < 0) {
rc = pcmk_legacy2rc(rc);
goto simulate_done;
}
}
if (pcmk_any_flags_set(flags, pcmk_sim_process | pcmk_sim_simulate)) {
pcmk__output_t *logger_out = NULL;
unsigned long long data_set_flags = pe_flag_no_compat;
if (pcmk_is_set(data_set->flags, pe_flag_show_scores)) {
data_set_flags |= pe_flag_show_scores;
}
if (pcmk_is_set(data_set->flags, pe_flag_show_utilization)) {
data_set_flags |= pe_flag_show_utilization;
}
if (pcmk_all_flags_set(data_set->flags,
pe_flag_show_scores|pe_flag_show_utilization)) {
PCMK__OUTPUT_SPACER_IF(out, printed == pcmk_rc_ok);
out->begin_list(out, NULL, NULL,
"Assignment Scores and Utilization Information");
printed = pcmk_rc_ok;
} else if (pcmk_is_set(data_set->flags, pe_flag_show_scores)) {
PCMK__OUTPUT_SPACER_IF(out, printed == pcmk_rc_ok);
out->begin_list(out, NULL, NULL, "Assignment Scores");
printed = pcmk_rc_ok;
} else if (pcmk_is_set(data_set->flags, pe_flag_show_utilization)) {
PCMK__OUTPUT_SPACER_IF(out, printed == pcmk_rc_ok);
out->begin_list(out, NULL, NULL, "Utilization Information");
printed = pcmk_rc_ok;
} else {
rc = pcmk__log_output_new(&logger_out);
if (rc != pcmk_rc_ok) {
goto simulate_done;
}
pe__register_messages(logger_out);
pcmk__register_lib_messages(logger_out);
data_set->priv = logger_out;
}
pcmk__schedule_actions(input, data_set_flags, data_set);
if (logger_out == NULL) {
out->end_list(out);
} else {
logger_out->finish(logger_out, CRM_EX_OK, true, NULL);
pcmk__output_free(logger_out);
data_set->priv = out;
}
input = NULL; /* Don't try and free it twice */
if (graph_file != NULL) {
rc = write_xml_file(data_set->graph, graph_file, FALSE);
if (rc < 0) {
rc = pcmk_rc_graph_error;
goto simulate_done;
}
}
if (dot_file != NULL) {
rc = write_sim_dotfile(data_set, dot_file,
pcmk_is_set(flags, pcmk_sim_all_actions),
pcmk_is_set(flags, pcmk_sim_verbose));
if (rc != pcmk_rc_ok) {
rc = pcmk_rc_dot_error;
goto simulate_done;
}
}
if (!out->is_quiet(out)) {
print_transition_summary(data_set, printed == pcmk_rc_ok);
}
}
rc = pcmk_rc_ok;
if (!pcmk_is_set(flags, pcmk_sim_simulate)) {
goto simulate_done;
}
PCMK__OUTPUT_SPACER_IF(out, printed == pcmk_rc_ok);
if (pcmk__simulate_transition(data_set, cib, injections->op_fail)
!= pcmk__graph_complete) {
rc = pcmk_rc_invalid_transition;
}
if (out->is_quiet(out)) {
goto simulate_done;
}
set_effective_date(data_set, true, use_date);
if (pcmk_is_set(flags, pcmk_sim_show_scores)) {
pe__set_working_set_flags(data_set, pe_flag_show_scores);
}
if (pcmk_is_set(flags, pcmk_sim_show_utilization)) {
pe__set_working_set_flags(data_set, pe_flag_show_utilization);
}
cluster_status(data_set);
print_cluster_status(data_set, 0, section_opts, "Revised Cluster Status",
true);
simulate_done:
cib__clean_up_connection(&cib);
return rc;
}
int
pcmk_simulate(xmlNodePtr *xml, pe_working_set_t *data_set,
const pcmk_injections_t *injections, unsigned int flags,
unsigned int section_opts, const char *use_date,
const char *input_file, const char *graph_file,
const char *dot_file)
{
pcmk__output_t *out = NULL;
int rc = pcmk_rc_ok;
rc = pcmk__xml_output_new(&out, xml);
if (rc != pcmk_rc_ok) {
return rc;
}
pe__register_messages(out);
pcmk__register_lib_messages(out);
rc = pcmk__simulate(data_set, out, injections, flags, section_opts,
use_date, input_file, graph_file, dot_file);
pcmk__xml_output_finish(out, xml);
return rc;
}
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index 4b142a6857..e7191dbece 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -1,627 +1,627 @@
/*
* Copyright 2004-2023 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
* This source code is licensed under the GNU Lesser General Public License
* version 2.1 or later (LGPLv2.1+) WITHOUT ANY WARRANTY.
*/
#include <crm_internal.h>
#include <crm/crm.h>
#include <crm/msg_xml.h>
#include <crm/common/xml.h>
#include <crm/common/util.h>
#include <glib.h>
#include <crm/pengine/internal.h>
gboolean was_processing_error = FALSE;
gboolean was_processing_warning = FALSE;
static bool
check_placement_strategy(const char *value)
{
return pcmk__strcase_any_of(value, "default", "utilization", "minimal",
"balanced", NULL);
}
static pcmk__cluster_option_t pe_opts[] = {
/* name, old name, type, allowed values,
* default value, validator,
* short description,
* long description
*/
{
"no-quorum-policy", NULL, "select", "stop, freeze, ignore, demote, suicide",
"stop", pcmk__valid_quorum,
N_("What to do when the cluster does not have quorum"),
NULL
},
{
"symmetric-cluster", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
N_("Whether resources can run on any node by default"),
NULL
},
{
"maintenance-mode", NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
N_("Whether the cluster should refrain from monitoring, starting, "
"and stopping resources"),
NULL
},
{
"start-failure-is-fatal", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
N_("Whether a start failure should prevent a resource from being "
"recovered on the same node"),
N_("When true, the cluster will immediately ban a resource from a node "
"if it fails to start there. When false, the cluster will instead "
"check the resource's fail count against its migration-threshold.")
},
{
"enable-startup-probes", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
N_("Whether the cluster should check for active resources during start-up"),
NULL
},
{
XML_CONFIG_ATTR_SHUTDOWN_LOCK, NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
N_("Whether to lock resources to a cleanly shut down node"),
N_("When true, resources active on a node when it is cleanly shut down "
"are kept \"locked\" to that node (not allowed to run elsewhere) "
"until they start again on that node after it rejoins (or for at "
"most shutdown-lock-limit, if set). Stonith resources and "
"Pacemaker Remote connections are never locked. Clone and bundle "
"instances and the promoted role of promotable clones are "
"currently never locked, though support could be added in a future "
"release.")
},
{
XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT, NULL, "time", NULL,
"0", pcmk__valid_interval_spec,
N_("Do not lock resources to a cleanly shut down node longer than "
"this"),
N_("If shutdown-lock is true and this is set to a nonzero time "
"duration, shutdown locks will expire after this much time has "
"passed since the shutdown was initiated, even if the node has not "
"rejoined.")
},
// Fencing-related options
{
"stonith-enabled", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
N_("*** Advanced Use Only *** "
"Whether nodes may be fenced as part of recovery"),
N_("If false, unresponsive nodes are immediately assumed to be harmless, "
"and resources that were active on them may be recovered "
"elsewhere. This can result in a \"split-brain\" situation, "
"potentially leading to data loss and/or service unavailability.")
},
{
"stonith-action", NULL, "select", "reboot, off, poweroff",
"reboot", pcmk__is_fencing_action,
N_("Action to send to fence device when a node needs to be fenced "
"(\"poweroff\" is a deprecated alias for \"off\")"),
NULL
},
{
"stonith-timeout", NULL, "time", NULL,
"60s", pcmk__valid_interval_spec,
N_("*** Advanced Use Only *** Unused by Pacemaker"),
N_("This value is not used by Pacemaker, but is kept for backward "
"compatibility, and certain legacy fence agents might use it.")
},
{
XML_ATTR_HAVE_WATCHDOG, NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
N_("Whether watchdog integration is enabled"),
N_("This is set automatically by the cluster according to whether SBD "
"is detected to be in use. User-configured values are ignored. "
"The value `true` is meaningful if diskless SBD is used and "
"`stonith-watchdog-timeout` is nonzero. In that case, if fencing "
"is required, watchdog-based self-fencing will be performed via "
"SBD without requiring a fencing resource explicitly configured.")
},
{
"concurrent-fencing", NULL, "boolean", NULL,
PCMK__CONCURRENT_FENCING_DEFAULT, pcmk__valid_boolean,
N_("Allow performing fencing operations in parallel"),
NULL
},
{
"startup-fencing", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
N_("*** Advanced Use Only *** Whether to fence unseen nodes at start-up"),
N_("Setting this to false may lead to a \"split-brain\" situation,"
"potentially leading to data loss and/or service unavailability.")
},
{
XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY, NULL, "time", NULL,
"0", pcmk__valid_interval_spec,
N_("Apply fencing delay targeting the lost nodes with the highest total resource priority"),
N_("Apply specified delay for the fencings that are targeting the lost "
"nodes with the highest total resource priority in case we don't "
"have the majority of the nodes in our cluster partition, so that "
"the more significant nodes potentially win any fencing match, "
"which is especially meaningful under split-brain of 2-node "
"cluster. A promoted resource instance takes the base priority + 1 "
"on calculation if the base priority is not 0. Any static/random "
"delays that are introduced by `pcmk_delay_base/max` configured "
"for the corresponding fencing resources will be added to this "
"delay. This delay should be significantly greater than, safely "
"twice, the maximum `pcmk_delay_base/max`. By default, priority "
"fencing delay is disabled.")
},
{
XML_CONFIG_ATTR_NODE_PENDING_TIMEOUT, NULL, "time", NULL,
"10min", pcmk__valid_interval_spec,
N_("How long to wait for a node that has joined the cluster to join "
"the process group"),
N_("A node that has joined the cluster can be pending on joining the "
"process group. We wait up to this much time for it. If it times "
"out, fencing targeting the node will be issued if enabled.")
},
{
"cluster-delay", NULL, "time", NULL,
"60s", pcmk__valid_interval_spec,
N_("Maximum time for node-to-node communication"),
N_("The node elected Designated Controller (DC) will consider an action "
"failed if it does not get a response from the node executing the "
"action within this time (after considering the action's own "
"timeout). The \"correct\" value will depend on the speed and "
"load of your network and cluster nodes.")
},
{
"batch-limit", NULL, "integer", NULL,
"0", pcmk__valid_number,
N_("Maximum number of jobs that the cluster may execute in parallel "
"across all nodes"),
N_("The \"correct\" value will depend on the speed and load of your "
"network and cluster nodes. If set to 0, the cluster will "
"impose a dynamically calculated limit when any node has a "
"high load.")
},
{
"migration-limit", NULL, "integer", NULL,
"-1", pcmk__valid_number,
N_("The number of live migration actions that the cluster is allowed "
"to execute in parallel on a node (-1 means no limit)")
},
/* Orphans and stopping */
{
"stop-all-resources", NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
N_("Whether the cluster should stop all active resources"),
NULL
},
{
"stop-orphan-resources", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
N_("Whether to stop resources that were removed from the configuration"),
NULL
},
{
"stop-orphan-actions", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
N_("Whether to cancel recurring actions removed from the configuration"),
NULL
},
{
"remove-after-stop", NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
N_("*** Deprecated *** Whether to remove stopped resources from "
"the executor"),
N_("Values other than default are poorly tested and potentially dangerous."
" This option will be removed in a future release.")
},
/* Storing inputs */
{
"pe-error-series-max", NULL, "integer", NULL,
"-1", pcmk__valid_number,
N_("The number of scheduler inputs resulting in errors to save"),
N_("Zero to disable, -1 to store unlimited.")
},
{
"pe-warn-series-max", NULL, "integer", NULL,
"5000", pcmk__valid_number,
N_("The number of scheduler inputs resulting in warnings to save"),
N_("Zero to disable, -1 to store unlimited.")
},
{
"pe-input-series-max", NULL, "integer", NULL,
"4000", pcmk__valid_number,
N_("The number of scheduler inputs without errors or warnings to save"),
N_("Zero to disable, -1 to store unlimited.")
},
/* Node health */
{
PCMK__OPT_NODE_HEALTH_STRATEGY, NULL, "select",
PCMK__VALUE_NONE ", " PCMK__VALUE_MIGRATE_ON_RED ", "
PCMK__VALUE_ONLY_GREEN ", " PCMK__VALUE_PROGRESSIVE ", "
PCMK__VALUE_CUSTOM,
PCMK__VALUE_NONE, pcmk__validate_health_strategy,
N_("How cluster should react to node health attributes"),
N_("Requires external entities to create node attributes (named with "
"the prefix \"#health\") with values \"red\", "
"\"yellow\", or \"green\".")
},
{
PCMK__OPT_NODE_HEALTH_BASE, NULL, "integer", NULL,
"0", pcmk__valid_number,
N_("Base health score assigned to a node"),
N_("Only used when \"node-health-strategy\" is set to \"progressive\".")
},
{
PCMK__OPT_NODE_HEALTH_GREEN, NULL, "integer", NULL,
"0", pcmk__valid_number,
N_("The score to use for a node health attribute whose value is \"green\""),
N_("Only used when \"node-health-strategy\" is set to \"custom\" or \"progressive\".")
},
{
PCMK__OPT_NODE_HEALTH_YELLOW, NULL, "integer", NULL,
"0", pcmk__valid_number,
N_("The score to use for a node health attribute whose value is \"yellow\""),
N_("Only used when \"node-health-strategy\" is set to \"custom\" or \"progressive\".")
},
{
PCMK__OPT_NODE_HEALTH_RED, NULL, "integer", NULL,
"-INFINITY", pcmk__valid_number,
N_("The score to use for a node health attribute whose value is \"red\""),
N_("Only used when \"node-health-strategy\" is set to \"custom\" or \"progressive\".")
},
/*Placement Strategy*/
{
"placement-strategy", NULL, "select",
"default, utilization, minimal, balanced",
"default", check_placement_strategy,
N_("How the cluster should allocate resources to nodes"),
NULL
},
};
void
pe_metadata(pcmk__output_t *out)
{
const char *desc_short = "Pacemaker scheduler options";
const char *desc_long = "Cluster options used by Pacemaker's scheduler";
gchar *s = pcmk__format_option_metadata("pacemaker-schedulerd", desc_short,
desc_long, pe_opts,
PCMK__NELEM(pe_opts));
out->output_xml(out, "metadata", s);
g_free(s);
}
void
verify_pe_options(GHashTable * options)
{
pcmk__validate_cluster_options(options, pe_opts, PCMK__NELEM(pe_opts));
}
const char *
pe_pref(GHashTable * options, const char *name)
{
return pcmk__cluster_option(options, pe_opts, PCMK__NELEM(pe_opts), name);
}
const char *
fail2text(enum action_fail_response fail)
{
const char *result = "<unknown>";
switch (fail) {
case action_fail_ignore:
result = "ignore";
break;
case action_fail_demote:
result = "demote";
break;
case action_fail_block:
result = "block";
break;
case action_fail_recover:
result = "recover";
break;
case action_fail_migrate:
result = "migrate";
break;
case action_fail_stop:
result = "stop";
break;
case action_fail_fence:
result = "fence";
break;
case action_fail_standby:
result = "standby";
break;
case action_fail_restart_container:
result = "restart-container";
break;
case action_fail_reset_remote:
result = "reset-remote";
break;
}
return result;
}
enum action_tasks
text2task(const char *task)
{
if (pcmk__str_eq(task, PCMK_ACTION_STOP, pcmk__str_casei)) {
return stop_rsc;
} else if (pcmk__str_eq(task, PCMK_ACTION_STOPPED, pcmk__str_casei)) {
return stopped_rsc;
} else if (pcmk__str_eq(task, PCMK_ACTION_START, pcmk__str_casei)) {
return start_rsc;
} else if (pcmk__str_eq(task, PCMK_ACTION_RUNNING, pcmk__str_casei)) {
return started_rsc;
} else if (pcmk__str_eq(task, CRM_OP_SHUTDOWN, pcmk__str_casei)) {
return shutdown_crm;
} else if (pcmk__str_eq(task, CRM_OP_FENCE, pcmk__str_casei)) {
return stonith_node;
} else if (pcmk__str_eq(task, PCMK_ACTION_MONITOR, pcmk__str_casei)) {
return monitor_rsc;
} else if (pcmk__str_eq(task, PCMK_ACTION_NOTIFY, pcmk__str_casei)) {
return action_notify;
- } else if (pcmk__str_eq(task, CRMD_ACTION_NOTIFIED, pcmk__str_casei)) {
+ } else if (pcmk__str_eq(task, PCMK_ACTION_NOTIFIED, pcmk__str_casei)) {
return action_notified;
} else if (pcmk__str_eq(task, PCMK_ACTION_PROMOTE, pcmk__str_casei)) {
return action_promote;
} else if (pcmk__str_eq(task, PCMK_ACTION_DEMOTE, pcmk__str_casei)) {
return action_demote;
} else if (pcmk__str_eq(task, PCMK_ACTION_PROMOTED, pcmk__str_casei)) {
return action_promoted;
} else if (pcmk__str_eq(task, PCMK_ACTION_DEMOTED, pcmk__str_casei)) {
return action_demoted;
}
#if SUPPORT_TRACING
if (pcmk__str_eq(task, PCMK_ACTION_CANCEL, pcmk__str_casei)) {
return no_action;
} else if (pcmk__str_eq(task, PCMK_ACTION_DELETE, pcmk__str_casei)) {
return no_action;
} else if (pcmk__str_eq(task, PCMK_ACTION_MONITOR, pcmk__str_casei)) {
return no_action;
} else if (pcmk__str_eq(task, PCMK_ACTION_MIGRATE_TO, pcmk__str_casei)) {
return no_action;
} else if (pcmk__str_eq(task, PCMK_ACTION_MIGRATE_FROM, pcmk__str_casei)) {
return no_action;
}
crm_trace("Unsupported action: %s", task);
#endif
return no_action;
}
const char *
task2text(enum action_tasks task)
{
const char *result = "<unknown>";
switch (task) {
case no_action:
result = "no_action";
break;
case stop_rsc:
result = PCMK_ACTION_STOP;
break;
case stopped_rsc:
result = PCMK_ACTION_STOPPED;
break;
case start_rsc:
result = PCMK_ACTION_START;
break;
case started_rsc:
result = PCMK_ACTION_RUNNING;
break;
case shutdown_crm:
result = CRM_OP_SHUTDOWN;
break;
case stonith_node:
result = CRM_OP_FENCE;
break;
case monitor_rsc:
result = PCMK_ACTION_MONITOR;
break;
case action_notify:
result = PCMK_ACTION_NOTIFY;
break;
case action_notified:
- result = CRMD_ACTION_NOTIFIED;
+ result = PCMK_ACTION_NOTIFIED;
break;
case action_promote:
result = PCMK_ACTION_PROMOTE;
break;
case action_promoted:
result = PCMK_ACTION_PROMOTED;
break;
case action_demote:
result = PCMK_ACTION_DEMOTE;
break;
case action_demoted:
result = PCMK_ACTION_DEMOTED;
break;
}
return result;
}
const char *
role2text(enum rsc_role_e role)
{
switch (role) {
case RSC_ROLE_UNKNOWN:
return RSC_ROLE_UNKNOWN_S;
case RSC_ROLE_STOPPED:
return RSC_ROLE_STOPPED_S;
case RSC_ROLE_STARTED:
return RSC_ROLE_STARTED_S;
case RSC_ROLE_UNPROMOTED:
#ifdef PCMK__COMPAT_2_0
return RSC_ROLE_UNPROMOTED_LEGACY_S;
#else
return RSC_ROLE_UNPROMOTED_S;
#endif
case RSC_ROLE_PROMOTED:
#ifdef PCMK__COMPAT_2_0
return RSC_ROLE_PROMOTED_LEGACY_S;
#else
return RSC_ROLE_PROMOTED_S;
#endif
}
CRM_CHECK(role >= RSC_ROLE_UNKNOWN, return RSC_ROLE_UNKNOWN_S);
CRM_CHECK(role < RSC_ROLE_MAX, return RSC_ROLE_UNKNOWN_S);
// coverity[dead_error_line]
return RSC_ROLE_UNKNOWN_S;
}
enum rsc_role_e
text2role(const char *role)
{
CRM_ASSERT(role != NULL);
if (pcmk__str_eq(role, RSC_ROLE_STOPPED_S, pcmk__str_casei)) {
return RSC_ROLE_STOPPED;
} else if (pcmk__str_eq(role, RSC_ROLE_STARTED_S, pcmk__str_casei)) {
return RSC_ROLE_STARTED;
} else if (pcmk__strcase_any_of(role, RSC_ROLE_UNPROMOTED_S,
RSC_ROLE_UNPROMOTED_LEGACY_S, NULL)) {
return RSC_ROLE_UNPROMOTED;
} else if (pcmk__strcase_any_of(role, RSC_ROLE_PROMOTED_S,
RSC_ROLE_PROMOTED_LEGACY_S, NULL)) {
return RSC_ROLE_PROMOTED;
} else if (pcmk__str_eq(role, RSC_ROLE_UNKNOWN_S, pcmk__str_casei)) {
return RSC_ROLE_UNKNOWN;
}
crm_err("Unknown role: %s", role);
return RSC_ROLE_UNKNOWN;
}
void
add_hash_param(GHashTable * hash, const char *name, const char *value)
{
CRM_CHECK(hash != NULL, return);
crm_trace("Adding name='%s' value='%s' to hash table",
pcmk__s(name, "<null>"), pcmk__s(value, "<null>"));
if (name == NULL || value == NULL) {
return;
} else if (pcmk__str_eq(value, "#default", pcmk__str_casei)) {
return;
} else if (g_hash_table_lookup(hash, name) == NULL) {
g_hash_table_insert(hash, strdup(name), strdup(value));
}
}
/*!
* \internal
* \brief Look up an attribute value on the appropriate node
*
* If \p node is a guest node and either the \c XML_RSC_ATTR_TARGET meta
* attribute is set to "host" for \p rsc or \p force_host is \c true, query the
* attribute on the node's host. Otherwise, query the attribute on \p node
* itself.
*
* \param[in] node Node to query attribute value on by default
* \param[in] name Name of attribute to query
* \param[in] rsc Resource on whose behalf we're querying
* \param[in] node_type Type of resource location lookup
* \param[in] force_host Force a lookup on the guest node's host, regardless of
* the \c XML_RSC_ATTR_TARGET value
*
* \return Value of the attribute on \p node or on the host of \p node
*
* \note If \p force_host is \c true, \p node \e must be a guest node.
*/
const char *
pe__node_attribute_calculated(const pe_node_t *node, const char *name,
const pe_resource_t *rsc,
enum pe__rsc_node node_type,
bool force_host)
{
// @TODO: Use pe__is_guest_node() after merging libpe_{rules,status}
bool is_guest = (node != NULL) && (node->details->type == node_remote)
&& (node->details->remote_rsc != NULL)
&& (node->details->remote_rsc->container != NULL);
const char *source = NULL;
const char *node_type_s = NULL;
const char *reason = NULL;
const pe_resource_t *container = NULL;
const pe_node_t *host = NULL;
CRM_ASSERT((node != NULL) && (name != NULL) && (rsc != NULL)
&& (!force_host || is_guest));
/* Ignore XML_RSC_ATTR_TARGET if node is not a guest node. This represents a
* user configuration error.
*/
source = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET);
if (!force_host
&& (!is_guest || !pcmk__str_eq(source, "host", pcmk__str_casei))) {
return g_hash_table_lookup(node->details->attrs, name);
}
container = node->details->remote_rsc->container;
switch (node_type) {
case pe__rsc_node_assigned:
node_type_s = "assigned";
host = container->allocated_to;
if (host == NULL) {
reason = "not assigned";
}
break;
case pe__rsc_node_current:
node_type_s = "current";
if (container->running_on != NULL) {
host = container->running_on->data;
}
if (host == NULL) {
reason = "inactive";
}
break;
default:
// Add support for other enum pe__rsc_node values if needed
CRM_ASSERT(false);
break;
}
if (host != NULL) {
const char *value = g_hash_table_lookup(host->details->attrs, name);
pe_rsc_trace(rsc,
"%s: Value lookup for %s on %s container host %s %s%s",
rsc->id, name, node_type_s, pe__node_name(host),
((value != NULL)? "succeeded: " : "failed"),
pcmk__s(value, ""));
return value;
}
pe_rsc_trace(rsc,
"%s: Not looking for %s on %s container host: %s is %s",
rsc->id, name, node_type_s, container->id, reason);
return NULL;
}
const char *
pe_node_attribute_raw(const pe_node_t *node, const char *name)
{
if(node == NULL) {
return NULL;
}
return g_hash_table_lookup(node->details->attrs, name);
}
diff --git a/lib/pengine/pe_notif.c b/lib/pengine/pe_notif.c
index be01d11c57..8fc4fcb2b1 100644
--- a/lib/pengine/pe_notif.c
+++ b/lib/pengine/pe_notif.c
@@ -1,1001 +1,1002 @@
/*
* Copyright 2004-2023 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
* This source code is licensed under the GNU General Public License version 2
* or later (GPLv2+) WITHOUT ANY WARRANTY.
*/
#include <crm_internal.h>
#include <crm/msg_xml.h>
#include <crm/pengine/internal.h>
#include <pacemaker-internal.h>
#include "pe_status_private.h"
typedef struct notify_entry_s {
const pe_resource_t *rsc;
const pe_node_t *node;
} notify_entry_t;
/*!
* \internal
* \brief Compare two notification entries
*
* Compare two notification entries, where the one with the alphabetically first
* resource name (or if equal, node name) sorts as first, with NULL sorting as
* less than non-NULL.
*
* \param[in] a First notification entry to compare
* \param[in] b Second notification entry to compare
*
* \return -1 if \p a sorts before \p b, 0 if they are equal, otherwise 1
*/
static gint
compare_notify_entries(gconstpointer a, gconstpointer b)
{
int tmp;
const notify_entry_t *entry_a = a;
const notify_entry_t *entry_b = b;
// NULL a or b is not actually possible
if ((entry_a == NULL) && (entry_b == NULL)) {
return 0;
}
if (entry_a == NULL) {
return 1;
}
if (entry_b == NULL) {
return -1;
}
// NULL resources sort first
if ((entry_a->rsc == NULL) && (entry_b->rsc == NULL)) {
return 0;
}
if (entry_a->rsc == NULL) {
return 1;
}
if (entry_b->rsc == NULL) {
return -1;
}
// Compare resource names
tmp = strcmp(entry_a->rsc->id, entry_b->rsc->id);
if (tmp != 0) {
return tmp;
}
// Otherwise NULL nodes sort first
if ((entry_a->node == NULL) && (entry_b->node == NULL)) {
return 0;
}
if (entry_a->node == NULL) {
return 1;
}
if (entry_b->node == NULL) {
return -1;
}
// Finally, compare node names
return strcmp(entry_a->node->details->id, entry_b->node->details->id);
}
/*!
* \internal
* \brief Duplicate a notification entry
*
* \param[in] entry Entry to duplicate
*
* \return Newly allocated duplicate of \p entry
* \note It is the caller's responsibility to free the return value.
*/
static notify_entry_t *
dup_notify_entry(const notify_entry_t *entry)
{
notify_entry_t *dup = calloc(1, sizeof(notify_entry_t));
CRM_ASSERT(dup != NULL);
dup->rsc = entry->rsc;
dup->node = entry->node;
return dup;
}
/*!
* \internal
* \brief Given a list of nodes, create strings with node names
*
* \param[in] list List of nodes (as pe_node_t *)
* \param[out] all_node_names If not NULL, will be set to space-separated list
* of the names of all nodes in \p list
* \param[out] host_node_names Same as \p all_node_names, except active
* guest nodes will list the name of their host
*
* \note The caller is responsible for freeing the output argument values using
* \p g_string_free().
*/
static void
get_node_names(const GList *list, GString **all_node_names,
GString **host_node_names)
{
if (all_node_names != NULL) {
*all_node_names = NULL;
}
if (host_node_names != NULL) {
*host_node_names = NULL;
}
for (const GList *iter = list; iter != NULL; iter = iter->next) {
const pe_node_t *node = (const pe_node_t *) iter->data;
if (node->details->uname == NULL) {
continue;
}
// Always add to list of all node names
if (all_node_names != NULL) {
pcmk__add_word(all_node_names, 1024, node->details->uname);
}
// Add to host node name list if appropriate
if (host_node_names != NULL) {
if (pe__is_guest_node(node)
&& (node->details->remote_rsc->container->running_on != NULL)) {
node = pe__current_node(node->details->remote_rsc->container);
if (node->details->uname == NULL) {
continue;
}
}
pcmk__add_word(host_node_names, 1024, node->details->uname);
}
}
if ((all_node_names != NULL) && (*all_node_names == NULL)) {
*all_node_names = g_string_new(" ");
}
if ((host_node_names != NULL) && (*host_node_names == NULL)) {
*host_node_names = g_string_new(" ");
}
}
/*!
* \internal
* \brief Create strings of instance and node names from notification entries
*
* \param[in,out] list List of notification entries (will be sorted here)
* \param[out] rsc_names If not NULL, will be set to space-separated list
* of clone instances from \p list
* \param[out] node_names If not NULL, will be set to space-separated list
* of node names from \p list
*
* \return (Possibly new) head of sorted \p list
* \note The caller is responsible for freeing the output argument values using
* \p g_list_free_full() and \p g_string_free().
*/
static GList *
notify_entries_to_strings(GList *list, GString **rsc_names,
GString **node_names)
{
const char *last_rsc_id = NULL;
// Initialize output lists to NULL
if (rsc_names != NULL) {
*rsc_names = NULL;
}
if (node_names != NULL) {
*node_names = NULL;
}
// Sort input list for user-friendliness (and ease of filtering duplicates)
list = g_list_sort(list, compare_notify_entries);
for (GList *gIter = list; gIter != NULL; gIter = gIter->next) {
notify_entry_t *entry = (notify_entry_t *) gIter->data;
// Entry must have a resource (with ID)
CRM_LOG_ASSERT((entry != NULL) && (entry->rsc != NULL)
&& (entry->rsc->id != NULL));
if ((entry == NULL) || (entry->rsc == NULL)
|| (entry->rsc->id == NULL)) {
continue;
}
// Entry must have a node unless listing inactive resources
CRM_LOG_ASSERT((node_names == NULL) || (entry->node != NULL));
if ((node_names != NULL) && (entry->node == NULL)) {
continue;
}
// Don't add duplicates of a particular clone instance
if (pcmk__str_eq(entry->rsc->id, last_rsc_id, pcmk__str_none)) {
continue;
}
last_rsc_id = entry->rsc->id;
if (rsc_names != NULL) {
pcmk__add_word(rsc_names, 1024, entry->rsc->id);
}
if ((node_names != NULL) && (entry->node->details->uname != NULL)) {
pcmk__add_word(node_names, 1024, entry->node->details->uname);
}
}
// If there are no entries, return "empty" lists
if ((rsc_names != NULL) && (*rsc_names == NULL)) {
*rsc_names = g_string_new(" ");
}
if ((node_names != NULL) && (*node_names == NULL)) {
*node_names = g_string_new(" ");
}
return list;
}
/*!
* \internal
* \brief Copy a meta-attribute into a notify action
*
* \param[in] key Name of meta-attribute to copy
* \param[in] value Value of meta-attribute to copy
* \param[in,out] user_data Notify action to copy into
*/
static void
copy_meta_to_notify(gpointer key, gpointer value, gpointer user_data)
{
pe_action_t *notify = (pe_action_t *) user_data;
/* Any existing meta-attributes (for example, the action timeout) are for
* the notify action itself, so don't override those.
*/
if (g_hash_table_lookup(notify->meta, (const char *) key) != NULL) {
return;
}
g_hash_table_insert(notify->meta, strdup((const char *) key),
strdup((const char *) value));
}
static void
add_notify_data_to_action_meta(const notify_data_t *n_data, pe_action_t *action)
{
for (const GSList *item = n_data->keys; item; item = item->next) {
const pcmk_nvpair_t *nvpair = (const pcmk_nvpair_t *) item->data;
add_hash_param(action->meta, nvpair->name, nvpair->value);
}
}
/*!
* \internal
* \brief Create a new notify pseudo-action for a clone resource
*
* \param[in,out] rsc Clone resource that notification is for
* \param[in] action Action to use in notify action key
- * \param[in] notif_action PCMK_ACTION_NOTIFY or RSC_NOTIFIED
+ * \param[in] notif_action PCMK_ACTION_NOTIFY or PCMK_ACTION_NOTIFIED
* \param[in] notif_type "pre", "post", "confirmed-pre", "confirmed-post"
*
* \return Newly created notify pseudo-action
*/
static pe_action_t *
new_notify_pseudo_action(pe_resource_t *rsc, const pe_action_t *action,
const char *notif_action, const char *notif_type)
{
pe_action_t *notify = NULL;
notify = custom_action(rsc,
pcmk__notify_key(rsc->id, notif_type, action->task),
notif_action, NULL,
pcmk_is_set(action->flags, pe_action_optional),
TRUE, rsc->cluster);
pe__set_action_flags(notify, pe_action_pseudo);
add_hash_param(notify->meta, "notify_key_type", notif_type);
add_hash_param(notify->meta, "notify_key_operation", action->task);
return notify;
}
/*!
* \internal
* \brief Create a new notify action for a clone instance
*
* \param[in,out] rsc Clone instance that notification is for
* \param[in] node Node that notification is for
* \param[in,out] op Action that notification is for
* \param[in,out] notify_done Parent pseudo-action for notifications complete
* \param[in] n_data Notification values to add to action meta-data
*
* \return Newly created notify action
*/
static pe_action_t *
new_notify_action(pe_resource_t *rsc, const pe_node_t *node, pe_action_t *op,
pe_action_t *notify_done, const notify_data_t *n_data)
{
char *key = NULL;
pe_action_t *notify_action = NULL;
const char *value = NULL;
const char *task = NULL;
const char *skip_reason = NULL;
CRM_CHECK((rsc != NULL) && (node != NULL), return NULL);
// Ensure we have all the info we need
if (op == NULL) {
skip_reason = "no action";
} else if (notify_done == NULL) {
skip_reason = "no parent notification";
} else if (!node->details->online) {
skip_reason = "node offline";
} else if (!pcmk_is_set(op->flags, pe_action_runnable)) {
skip_reason = "original action not runnable";
}
if (skip_reason != NULL) {
pe_rsc_trace(rsc, "Skipping notify action for %s on %s: %s",
rsc->id, pe__node_name(node), skip_reason);
return NULL;
}
value = g_hash_table_lookup(op->meta, "notify_type"); // "pre" or "post"
task = g_hash_table_lookup(op->meta, "notify_operation"); // original action
pe_rsc_trace(rsc, "Creating notify action for %s on %s (%s-%s)",
rsc->id, pe__node_name(node), value, task);
// Create the notify action
key = pcmk__notify_key(rsc->id, value, task);
notify_action = custom_action(rsc, key, op->task, node,
pcmk_is_set(op->flags, pe_action_optional),
TRUE, rsc->cluster);
// Add meta-data to notify action
g_hash_table_foreach(op->meta, copy_meta_to_notify, notify_action);
add_notify_data_to_action_meta(n_data, notify_action);
// Order notify after original action and before parent notification
order_actions(op, notify_action, pe_order_optional);
order_actions(notify_action, notify_done, pe_order_optional);
return notify_action;
}
/*!
* \internal
* \brief Create a new "post-" notify action for a clone instance
*
* \param[in,out] rsc Clone instance that notification is for
* \param[in] node Node that notification is for
* \param[in,out] n_data Notification values to add to action meta-data
*/
static void
new_post_notify_action(pe_resource_t *rsc, const pe_node_t *node,
notify_data_t *n_data)
{
pe_action_t *notify = NULL;
CRM_ASSERT(n_data != NULL);
// Create the "post-" notify action for specified instance
notify = new_notify_action(rsc, node, n_data->post, n_data->post_done,
n_data);
if (notify != NULL) {
notify->priority = INFINITY;
}
// Order recurring monitors after all "post-" notifications complete
if (n_data->post_done == NULL) {
return;
}
for (GList *iter = rsc->actions; iter != NULL; iter = iter->next) {
pe_action_t *mon = (pe_action_t *) iter->data;
const char *interval_ms_s = NULL;
interval_ms_s = g_hash_table_lookup(mon->meta,
XML_LRM_ATTR_INTERVAL_MS);
if (pcmk__str_eq(interval_ms_s, "0", pcmk__str_null_matches)
|| pcmk__str_eq(mon->task, PCMK_ACTION_CANCEL, pcmk__str_none)) {
continue; // Not a recurring monitor
}
order_actions(n_data->post_done, mon, pe_order_optional);
}
}
/*!
* \internal
* \brief Create and order notification pseudo-actions for a clone action
*
* In addition to the actual notify actions needed for each clone instance,
* clone notifications also require pseudo-actions to provide ordering points
* in the notification process. This creates the notification data, along with
* appropriate pseudo-actions and their orderings.
*
* For example, the ordering sequence for starting a clone is:
*
* "pre-" notify pseudo-action for clone
* -> "pre-" notify actions for each clone instance
* -> "pre-" notifications complete pseudo-action for clone
* -> start actions for each clone instance
* -> "started" pseudo-action for clone
* -> "post-" notify pseudo-action for clone
* -> "post-" notify actions for each clone instance
* -> "post-" notifications complete pseudo-action for clone
*
* \param[in,out] rsc Clone that notifications are for
* \param[in] task Name of action that notifications are for
* \param[in,out] action If not NULL, create a "pre-" pseudo-action ordered
* before a "pre-" complete pseudo-action, ordered
* before this action
* \param[in,out] complete If not NULL, create a "post-" pseudo-action ordered
* after this action, and a "post-" complete
* pseudo-action ordered after that
*
* \return Newly created notification data
*/
notify_data_t *
pe__action_notif_pseudo_ops(pe_resource_t *rsc, const char *task,
pe_action_t *action, pe_action_t *complete)
{
notify_data_t *n_data = NULL;
if (!pcmk_is_set(rsc->flags, pe_rsc_notify)) {
return NULL;
}
n_data = calloc(1, sizeof(notify_data_t));
CRM_ASSERT(n_data != NULL);
n_data->action = task;
if (action != NULL) { // Need "pre-" pseudo-actions
// Create "pre-" notify pseudo-action for clone
n_data->pre = new_notify_pseudo_action(rsc, action, PCMK_ACTION_NOTIFY,
"pre");
pe__set_action_flags(n_data->pre, pe_action_runnable);
add_hash_param(n_data->pre->meta, "notify_type", "pre");
add_hash_param(n_data->pre->meta, "notify_operation", n_data->action);
// Create "pre-" notifications complete pseudo-action for clone
- n_data->pre_done = new_notify_pseudo_action(rsc, action, RSC_NOTIFIED,
+ n_data->pre_done = new_notify_pseudo_action(rsc, action,
+ PCMK_ACTION_NOTIFIED,
"confirmed-pre");
pe__set_action_flags(n_data->pre_done, pe_action_runnable);
add_hash_param(n_data->pre_done->meta, "notify_type", "pre");
add_hash_param(n_data->pre_done->meta,
"notify_operation", n_data->action);
// Order "pre-" -> "pre-" complete -> original action
order_actions(n_data->pre, n_data->pre_done, pe_order_optional);
order_actions(n_data->pre_done, action, pe_order_optional);
}
if (complete != NULL) { // Need "post-" pseudo-actions
// Create "post-" notify pseudo-action for clone
n_data->post = new_notify_pseudo_action(rsc, complete,
PCMK_ACTION_NOTIFY, "post");
n_data->post->priority = INFINITY;
if (pcmk_is_set(complete->flags, pe_action_runnable)) {
pe__set_action_flags(n_data->post, pe_action_runnable);
} else {
pe__clear_action_flags(n_data->post, pe_action_runnable);
}
add_hash_param(n_data->post->meta, "notify_type", "post");
add_hash_param(n_data->post->meta, "notify_operation", n_data->action);
// Create "post-" notifications complete pseudo-action for clone
n_data->post_done = new_notify_pseudo_action(rsc, complete,
- RSC_NOTIFIED,
+ PCMK_ACTION_NOTIFIED,
"confirmed-post");
n_data->post_done->priority = INFINITY;
if (pcmk_is_set(complete->flags, pe_action_runnable)) {
pe__set_action_flags(n_data->post_done, pe_action_runnable);
} else {
pe__clear_action_flags(n_data->post_done, pe_action_runnable);
}
add_hash_param(n_data->post_done->meta, "notify_type", "post");
add_hash_param(n_data->post_done->meta,
"notify_operation", n_data->action);
// Order original action complete -> "post-" -> "post-" complete
order_actions(complete, n_data->post, pe_order_implies_then);
order_actions(n_data->post, n_data->post_done, pe_order_implies_then);
}
// If we created both, order "pre-" complete -> "post-"
if ((action != NULL) && (complete != NULL)) {
order_actions(n_data->pre_done, n_data->post, pe_order_optional);
}
return n_data;
}
/*!
* \internal
* \brief Create a new notification entry
*
* \param[in] rsc Resource for notification
* \param[in] node Node for notification
*
* \return Newly allocated notification entry
* \note The caller is responsible for freeing the return value.
*/
static notify_entry_t *
new_notify_entry(const pe_resource_t *rsc, const pe_node_t *node)
{
notify_entry_t *entry = calloc(1, sizeof(notify_entry_t));
CRM_ASSERT(entry != NULL);
entry->rsc = rsc;
entry->node = node;
return entry;
}
/*!
* \internal
* \brief Add notification data for resource state and optionally actions
*
* \param[in] rsc Clone or clone instance being notified
* \param[in] activity Whether to add notification entries for actions
* \param[in,out] n_data Notification data for clone
*/
static void
collect_resource_data(const pe_resource_t *rsc, bool activity,
notify_data_t *n_data)
{
const GList *iter = NULL;
notify_entry_t *entry = NULL;
const pe_node_t *node = NULL;
if (n_data == NULL) {
return;
}
if (n_data->allowed_nodes == NULL) {
n_data->allowed_nodes = rsc->allowed_nodes;
}
// If this is a clone, call recursively for each instance
if (rsc->children != NULL) {
for (iter = rsc->children; iter != NULL; iter = iter->next) {
const pe_resource_t *child = (const pe_resource_t *) iter->data;
collect_resource_data(child, activity, n_data);
}
return;
}
// This is a notification for a single clone instance
if (rsc->running_on != NULL) {
node = rsc->running_on->data; // First is sufficient
}
entry = new_notify_entry(rsc, node);
// Add notification indicating the resource state
switch (rsc->role) {
case RSC_ROLE_STOPPED:
n_data->inactive = g_list_prepend(n_data->inactive, entry);
break;
case RSC_ROLE_STARTED:
n_data->active = g_list_prepend(n_data->active, entry);
break;
case RSC_ROLE_UNPROMOTED:
n_data->unpromoted = g_list_prepend(n_data->unpromoted, entry);
n_data->active = g_list_prepend(n_data->active,
dup_notify_entry(entry));
break;
case RSC_ROLE_PROMOTED:
n_data->promoted = g_list_prepend(n_data->promoted, entry);
n_data->active = g_list_prepend(n_data->active,
dup_notify_entry(entry));
break;
default:
crm_err("Resource %s role on %s (%s) is not supported for "
"notifications (bug?)",
rsc->id, pe__node_name(node), role2text(rsc->role));
free(entry);
break;
}
if (!activity) {
return;
}
// Add notification entries for each of the resource's actions
for (iter = rsc->actions; iter != NULL; iter = iter->next) {
const pe_action_t *op = (const pe_action_t *) iter->data;
if (!pcmk_is_set(op->flags, pe_action_optional) && (op->node != NULL)) {
enum action_tasks task = text2task(op->task);
if ((task == stop_rsc) && op->node->details->unclean) {
// Create anyway (additional noise if node can't be fenced)
} else if (!pcmk_is_set(op->flags, pe_action_runnable)) {
continue;
}
entry = new_notify_entry(rsc, op->node);
switch (task) {
case start_rsc:
n_data->start = g_list_prepend(n_data->start, entry);
break;
case stop_rsc:
n_data->stop = g_list_prepend(n_data->stop, entry);
break;
case action_promote:
n_data->promote = g_list_prepend(n_data->promote, entry);
break;
case action_demote:
n_data->demote = g_list_prepend(n_data->demote, entry);
break;
default:
free(entry);
break;
}
}
}
}
// For (char *) value
#define add_notify_env(n_data, key, value) do { \
n_data->keys = pcmk_prepend_nvpair(n_data->keys, key, value); \
} while (0)
// For (GString *) value
#define add_notify_env_gs(n_data, key, value) do { \
n_data->keys = pcmk_prepend_nvpair(n_data->keys, key, \
(const char *) value->str); \
} while (0)
// For (GString *) value
#define add_notify_env_free_gs(n_data, key, value) do { \
n_data->keys = pcmk_prepend_nvpair(n_data->keys, key, \
(const char *) value->str); \
g_string_free(value, TRUE); value = NULL; \
} while (0)
/*!
* \internal
* \brief Create notification name/value pairs from structured data
*
* \param[in] rsc Resource that notification is for
* \param[in,out] n_data Notification data
*/
static void
add_notif_keys(const pe_resource_t *rsc, notify_data_t *n_data)
{
bool required = false; // Whether to make notify actions required
GString *rsc_list = NULL;
GString *node_list = NULL;
GString *metal_list = NULL;
const char *source = NULL;
GList *nodes = NULL;
n_data->stop = notify_entries_to_strings(n_data->stop,
&rsc_list, &node_list);
if ((strcmp(" ", (const char *) rsc_list->str) != 0)
&& pcmk__str_eq(n_data->action, PCMK_ACTION_STOP, pcmk__str_none)) {
required = true;
}
add_notify_env_free_gs(n_data, "notify_stop_resource", rsc_list);
add_notify_env_free_gs(n_data, "notify_stop_uname", node_list);
if ((n_data->start != NULL)
&& pcmk__str_eq(n_data->action, PCMK_ACTION_START, pcmk__str_none)) {
required = true;
}
n_data->start = notify_entries_to_strings(n_data->start,
&rsc_list, &node_list);
add_notify_env_free_gs(n_data, "notify_start_resource", rsc_list);
add_notify_env_free_gs(n_data, "notify_start_uname", node_list);
if ((n_data->demote != NULL)
&& pcmk__str_eq(n_data->action, PCMK_ACTION_DEMOTE, pcmk__str_none)) {
required = true;
}
n_data->demote = notify_entries_to_strings(n_data->demote,
&rsc_list, &node_list);
add_notify_env_free_gs(n_data, "notify_demote_resource", rsc_list);
add_notify_env_free_gs(n_data, "notify_demote_uname", node_list);
if ((n_data->promote != NULL)
&& pcmk__str_eq(n_data->action, PCMK_ACTION_PROMOTE, pcmk__str_none)) {
required = true;
}
n_data->promote = notify_entries_to_strings(n_data->promote,
&rsc_list, &node_list);
add_notify_env_free_gs(n_data, "notify_promote_resource", rsc_list);
add_notify_env_free_gs(n_data, "notify_promote_uname", node_list);
n_data->active = notify_entries_to_strings(n_data->active,
&rsc_list, &node_list);
add_notify_env_free_gs(n_data, "notify_active_resource", rsc_list);
add_notify_env_free_gs(n_data, "notify_active_uname", node_list);
n_data->unpromoted = notify_entries_to_strings(n_data->unpromoted,
&rsc_list, &node_list);
add_notify_env_gs(n_data, "notify_unpromoted_resource", rsc_list);
add_notify_env_gs(n_data, "notify_unpromoted_uname", node_list);
// Deprecated: kept for backward compatibility with older resource agents
add_notify_env_free_gs(n_data, "notify_slave_resource", rsc_list);
add_notify_env_free_gs(n_data, "notify_slave_uname", node_list);
n_data->promoted = notify_entries_to_strings(n_data->promoted,
&rsc_list, &node_list);
add_notify_env_gs(n_data, "notify_promoted_resource", rsc_list);
add_notify_env_gs(n_data, "notify_promoted_uname", node_list);
// Deprecated: kept for backward compatibility with older resource agents
add_notify_env_free_gs(n_data, "notify_master_resource", rsc_list);
add_notify_env_free_gs(n_data, "notify_master_uname", node_list);
n_data->inactive = notify_entries_to_strings(n_data->inactive,
&rsc_list, NULL);
add_notify_env_free_gs(n_data, "notify_inactive_resource", rsc_list);
nodes = g_hash_table_get_values(n_data->allowed_nodes);
if (!pcmk__is_daemon) {
/* For display purposes, sort the node list, for consistent
* regression test output (while avoiding the performance hit
* for the live cluster).
*/
nodes = g_list_sort(nodes, pe__cmp_node_name);
}
get_node_names(nodes, &node_list, NULL);
add_notify_env_free_gs(n_data, "notify_available_uname", node_list);
g_list_free(nodes);
source = g_hash_table_lookup(rsc->meta, XML_RSC_ATTR_TARGET);
if (pcmk__str_eq("host", source, pcmk__str_none)) {
get_node_names(rsc->cluster->nodes, &node_list, &metal_list);
add_notify_env_free_gs(n_data, "notify_all_hosts", metal_list);
} else {
get_node_names(rsc->cluster->nodes, &node_list, NULL);
}
add_notify_env_free_gs(n_data, "notify_all_uname", node_list);
if (required && (n_data->pre != NULL)) {
pe__clear_action_flags(n_data->pre, pe_action_optional);
pe__clear_action_flags(n_data->pre_done, pe_action_optional);
}
if (required && (n_data->post != NULL)) {
pe__clear_action_flags(n_data->post, pe_action_optional);
pe__clear_action_flags(n_data->post_done, pe_action_optional);
}
}
/*
* \internal
* \brief Find any remote connection start relevant to an action
*
* \param[in] action Action to check
*
* \return If action is behind a remote connection, connection's start
*/
static pe_action_t *
find_remote_start(pe_action_t *action)
{
if ((action != NULL) && (action->node != NULL)) {
pe_resource_t *remote_rsc = action->node->details->remote_rsc;
if (remote_rsc != NULL) {
return find_first_action(remote_rsc->actions, NULL,
PCMK_ACTION_START,
NULL);
}
}
return NULL;
}
/*!
* \internal
* \brief Create notify actions, and add notify data to original actions
*
* \param[in,out] rsc Clone or clone instance that notification is for
* \param[in,out] n_data Clone notification data for some action
*/
static void
create_notify_actions(pe_resource_t *rsc, notify_data_t *n_data)
{
GList *iter = NULL;
pe_action_t *stop = NULL;
pe_action_t *start = NULL;
enum action_tasks task = text2task(n_data->action);
// If this is a clone, call recursively for each instance
if (rsc->children != NULL) {
g_list_foreach(rsc->children, (GFunc) create_notify_actions, n_data);
return;
}
// Add notification meta-attributes to original actions
for (iter = rsc->actions; iter != NULL; iter = iter->next) {
pe_action_t *op = (pe_action_t *) iter->data;
if (!pcmk_is_set(op->flags, pe_action_optional) && (op->node != NULL)) {
switch (text2task(op->task)) {
case start_rsc:
case stop_rsc:
case action_promote:
case action_demote:
add_notify_data_to_action_meta(n_data, op);
break;
default:
break;
}
}
}
// Skip notify action itself if original action was not needed
switch (task) {
case start_rsc:
if (n_data->start == NULL) {
pe_rsc_trace(rsc, "No notify action needed for %s %s",
rsc->id, n_data->action);
return;
}
break;
case action_promote:
if (n_data->promote == NULL) {
pe_rsc_trace(rsc, "No notify action needed for %s %s",
rsc->id, n_data->action);
return;
}
break;
case action_demote:
if (n_data->demote == NULL) {
pe_rsc_trace(rsc, "No notify action needed for %s %s",
rsc->id, n_data->action);
return;
}
break;
default:
// We cannot do same for stop because it might be implied by fencing
break;
}
pe_rsc_trace(rsc, "Creating notify actions for %s %s",
rsc->id, n_data->action);
// Create notify actions for stop or demote
if ((rsc->role != RSC_ROLE_STOPPED)
&& ((task == stop_rsc) || (task == action_demote))) {
stop = find_first_action(rsc->actions, NULL, PCMK_ACTION_STOP, NULL);
for (iter = rsc->running_on; iter != NULL; iter = iter->next) {
pe_node_t *current_node = (pe_node_t *) iter->data;
/* If a stop is a pseudo-action implied by fencing, don't try to
* notify the node getting fenced.
*/
if ((stop != NULL) && pcmk_is_set(stop->flags, pe_action_pseudo)
&& (current_node->details->unclean
|| current_node->details->remote_requires_reset)) {
continue;
}
new_notify_action(rsc, current_node, n_data->pre,
n_data->pre_done, n_data);
if ((task == action_demote) || (stop == NULL)
|| pcmk_is_set(stop->flags, pe_action_optional)) {
new_post_notify_action(rsc, current_node, n_data);
}
}
}
// Create notify actions for start or promote
if ((rsc->next_role != RSC_ROLE_STOPPED)
&& ((task == start_rsc) || (task == action_promote))) {
start = find_first_action(rsc->actions, NULL, PCMK_ACTION_START, NULL);
if (start != NULL) {
pe_action_t *remote_start = find_remote_start(start);
if ((remote_start != NULL)
&& !pcmk_is_set(remote_start->flags, pe_action_runnable)) {
/* Start and promote actions for a clone instance behind
* a Pacemaker Remote connection happen after the
* connection starts. If the connection start is blocked, do
* not schedule notifications for these actions.
*/
return;
}
}
if (rsc->allocated_to == NULL) {
pe_proc_err("Next role '%s' but %s is not allocated",
role2text(rsc->next_role), rsc->id);
return;
}
if ((task != start_rsc) || (start == NULL)
|| pcmk_is_set(start->flags, pe_action_optional)) {
new_notify_action(rsc, rsc->allocated_to, n_data->pre,
n_data->pre_done, n_data);
}
new_post_notify_action(rsc, rsc->allocated_to, n_data);
}
}
/*!
* \internal
* \brief Create notification data and actions for one clone action
*
* \param[in,out] rsc Clone resource that notification is for
* \param[in,out] n_data Clone notification data for some action
*/
void
pe__create_action_notifications(pe_resource_t *rsc, notify_data_t *n_data)
{
if ((rsc == NULL) || (n_data == NULL)) {
return;
}
collect_resource_data(rsc, true, n_data);
add_notif_keys(rsc, n_data);
create_notify_actions(rsc, n_data);
}
/*!
* \internal
* \brief Free notification data for one action
*
* \param[in,out] n_data Notification data to free
*/
void
pe__free_action_notification_data(notify_data_t *n_data)
{
if (n_data == NULL) {
return;
}
g_list_free_full(n_data->stop, free);
g_list_free_full(n_data->start, free);
g_list_free_full(n_data->demote, free);
g_list_free_full(n_data->promote, free);
g_list_free_full(n_data->promoted, free);
g_list_free_full(n_data->unpromoted, free);
g_list_free_full(n_data->active, free);
g_list_free_full(n_data->inactive, free);
pcmk_free_nvpairs(n_data->keys);
free(n_data);
}
/*!
* \internal
* \brief Order clone "notifications complete" pseudo-action after fencing
*
* If a stop action is implied by fencing, the usual notification pseudo-actions
* will not be sufficient to order things properly, or even create all needed
* notifications if the clone is also stopping on another node, and another
* clone is ordered after it. This function creates new notification
* pseudo-actions relative to the fencing to ensure everything works properly.
*
* \param[in] stop Stop action implied by fencing
* \param[in,out] rsc Clone resource that notification is for
* \param[in,out] stonith_op Fencing action that implies \p stop
*/
void
pe__order_notifs_after_fencing(const pe_action_t *stop, pe_resource_t *rsc,
pe_action_t *stonith_op)
{
notify_data_t *n_data;
crm_info("Ordering notifications for implied %s after fencing", stop->uuid);
n_data = pe__action_notif_pseudo_ops(rsc, PCMK_ACTION_STOP, NULL,
stonith_op);
if (n_data != NULL) {
collect_resource_data(rsc, false, n_data);
add_notify_env(n_data, "notify_stop_resource", rsc->id);
add_notify_env(n_data, "notify_stop_uname", stop->node->details->uname);
create_notify_actions(uber_parent(rsc), n_data);
pe__free_action_notification_data(n_data);
}
}

File Metadata

Mime Type
text/x-diff
Expires
Tue, Jul 8, 6:17 PM (17 h, 21 m)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
2002561
Default Alt Text
(166 KB)

Event Timeline