rpms
/
pacemaker
7
0
Fork 0
Code Issues Pull Requests Releases Activity
pacemaker/SOURCES/007-option-metadata.patch

3690 lines
156 KiB
Diff
Raw Blame History

From 4401064f409921caed9444d18a74713250213c44 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Tue, 2 Jan 2024 19:52:41 -0800
Subject: [PATCH 01/24] Test: cts-cli: Update for added spaces
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
cts/cli/regression.daemons.exp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/cts/cli/regression.daemons.exp b/cts/cli/regression.daemons.exp
index b34fba8..1cd049f 100644
--- a/cts/cli/regression.daemons.exp
+++ b/cts/cli/regression.daemons.exp
@@ -47,7 +47,7 @@
<content type="time" default=""/>
</parameter>
<parameter name="cluster-recheck-interval">
- <longdesc lang="en">Pacemaker is primarily event-driven, and looks ahead to know when to recheck cluster state for failure timeouts and most time-based rules. However, it will also recheck the cluster after this amount of inactivity, to evaluate rules with date specifications and serve as a fail-safe for certain types of scheduler bugs. Allowed values: Zero disables polling, while positive values are an interval in seconds(unless other units are specified, for example "5min")</longdesc>
+ <longdesc lang="en">Pacemaker is primarily event-driven, and looks ahead to know when to recheck cluster state for failure timeouts and most time-based rules. However, it will also recheck the cluster after this amount of inactivity, to evaluate rules with date specifications and serve as a fail-safe for certain types of scheduler bugs. Allowed values: Zero disables polling, while positive values are an interval in seconds (unless other units are specified, for example "5min")</longdesc>
<shortdesc lang="en">Polling interval to recheck cluster state and evaluate rules with date specifications</shortdesc>
<content type="time" default=""/>
</parameter>
@@ -345,7 +345,7 @@
<content type="boolean" default=""/>
</parameter>
<parameter name="startup-fencing">
- <longdesc lang="en">Setting this to false may lead to a &quot;split-brain&quot; situation,potentially leading to data loss and/or service unavailability.</longdesc>
+ <longdesc lang="en">Setting this to false may lead to a &quot;split-brain&quot; situation, potentially leading to data loss and/or service unavailability.</longdesc>
<shortdesc lang="en">*** Advanced Use Only *** Whether to fence unseen nodes at start-up</shortdesc>
<content type="boolean" default=""/>
</parameter>
--
2.31.1
From e3597b061afa62a1f869a3e238ad78d3f7222029 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Tue, 2 Jan 2024 20:28:21 -0800
Subject: [PATCH 02/24] Test: cts-cli: Update daemon tests to note that
stonith-timeout is used
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
cts/cli/regression.daemons.exp | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/cts/cli/regression.daemons.exp b/cts/cli/regression.daemons.exp
index 1cd049f..6a24089 100644
--- a/cts/cli/regression.daemons.exp
+++ b/cts/cli/regression.daemons.exp
@@ -330,8 +330,8 @@
</content>
</parameter>
<parameter name="stonith-timeout">
- <longdesc lang="en">This value is not used by Pacemaker, but is kept for backward compatibility, and certain legacy fence agents might use it.</longdesc>
- <shortdesc lang="en">*** Advanced Use Only *** Unused by Pacemaker</shortdesc>
+ <longdesc lang="en">How long to wait for on, off, and reboot fence actions to complete by default</longdesc>
+ <shortdesc lang="en">How long to wait for on, off, and reboot fence actions to complete by default</shortdesc>
<content type="time" default=""/>
</parameter>
<parameter name="have-watchdog">
--
2.31.1
From cb3431c1058dd3cfbcc1cc490db268a95d1731e1 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Sat, 6 Jan 2024 19:20:26 -0800
Subject: [PATCH 03/24] Refactor: libcrmcommon: New
pcmk__valid_placement_strategy()
We'll soon need this function outside the scheduler.
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
include/crm/common/options_internal.h | 1 +
lib/common/options.c | 16 ++++++++++++++++
lib/pengine/common.c | 9 +--------
3 files changed, 18 insertions(+), 8 deletions(-)
diff --git a/include/crm/common/options_internal.h b/include/crm/common/options_internal.h
index a9316ca..1ea27ee 100644
--- a/include/crm/common/options_internal.h
+++ b/include/crm/common/options_internal.h
@@ -67,6 +67,7 @@ bool pcmk__valid_number(const char *value);
bool pcmk__valid_positive_number(const char *value);
bool pcmk__valid_quorum(const char *value);
bool pcmk__valid_script(const char *value);
+bool pcmk__valid_placement_strategy(const char *value);
bool pcmk__valid_percentage(const char *value);
// from watchdog.c
diff --git a/lib/common/options.c b/lib/common/options.c
index 2d86ebc..1db41a2 100644
--- a/lib/common/options.c
+++ b/lib/common/options.c
@@ -246,6 +246,22 @@ pcmk__valid_script(const char *value)
return true;
}
+
+/*!
+ * \internal
+ * \brief Check whether a string represents a valid placement strategy
+ *
+ * \param[in] value String to validate
+ *
+ * \return \c true if \p value is a valid placement strategy, or \c false
+ * otherwise
+ */
+bool
+pcmk__valid_placement_strategy(const char *value)
+{
+ return pcmk__strcase_any_of(value, "default", "utilization", "minimal",
+ "balanced", NULL);
+}
bool
pcmk__valid_percentage(const char *value)
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index 0fdd5a1..6878f4d 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -21,13 +21,6 @@
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,
@@ -285,7 +278,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"placement-strategy", NULL, "select",
"default, utilization, minimal, balanced",
- "default", check_placement_strategy,
+ "default", pcmk__valid_placement_strategy,
N_("How the cluster should allocate resources to nodes"),
NULL
},
--
2.31.1
From 4c877cefcde40da8a2cd776956ade62919a2c926 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Tue, 9 Jan 2024 22:13:19 -0800
Subject: [PATCH 04/24] Refactor: controller: Remove stonith-watchdog-timeout
validator function
...from options array. Instead, call it from the controller after
validating the options array.
We'll soon be moving the options array to libcrmcommon. There, we don't
have access to controld_verify_stonith_watchdog_timeout() or to the
controller's stonith API connection and node name. New comments and the
following task have more details: https://projects.clusterlabs.org/T749.
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/controld/controld_control.c | 25 ++++++++++++++++++++++++-
lib/common/watchdog.c | 4 ++++
2 files changed, 28 insertions(+), 1 deletion(-)
diff --git a/daemons/controld/controld_control.c b/daemons/controld/controld_control.c
index 644d686..83b802e 100644
--- a/daemons/controld/controld_control.c
+++ b/daemons/controld/controld_control.c
@@ -621,8 +621,18 @@ static pcmk__cluster_option_t controller_options[] = {
"the order in which ping updates arrive.")
},
{
+ /* @COMPAT Currently unparsable values default to -1 (auto-calculate),
+ * while missing values default to 0 (disable). All values are accepted
+ * (unless the controller finds that the value conflicts with the
+ * SBD_WATCHDOG_TIMEOUT).
+ *
+ * At a compatibility break: properly validate as a timeout, let
+ * either negative values or a particular string like "auto" mean auto-
+ * calculate, and use 0 as the single default for when the option either
+ * is unset or fails to validate.
+ */
"stonith-watchdog-timeout", NULL, "time", NULL,
- "0", controld_verify_stonith_watchdog_timeout,
+ "0", NULL,
N_("How long before nodes can be assumed to be safely down when "
"watchdog-based self-fencing via SBD is in use"),
N_("If this is set to a positive value, lost nodes are assumed to "
@@ -747,6 +757,19 @@ config_query_callback(xmlNode * msg, int call_id, int rc, xmlNode * output, void
pcmk__validate_cluster_options(config_hash, controller_options,
PCMK__NELEM(controller_options));
+ /* Validate the watchdog timeout in the context of the local node
+ * environment. If invalid, the controller will exit with a fatal error.
+ *
+ * We do this via a wrapper in the controller, so that we call
+ * pcmk__valid_stonith_watchdog_timeout() only if watchdog fencing is
+ * enabled for the local node. Otherwise, we may exit unnecessarily.
+ *
+ * A validator function in libcrmcommon can't act as such a wrapper, because
+ * it doesn't have a stonith API connection or the local node name.
+ */
+ value = g_hash_table_lookup(config_hash, "stonith-watchdog-timeout");
+ controld_verify_stonith_watchdog_timeout(value);
+
value = g_hash_table_lookup(config_hash, "no-quorum-policy");
if (pcmk__str_eq(value, "suicide", pcmk__str_casei) && pcmk__locate_sbd()) {
controld_set_global_flags(controld_no_quorum_suicide);
diff --git a/lib/common/watchdog.c b/lib/common/watchdog.c
index e569214..7d126af 100644
--- a/lib/common/watchdog.c
+++ b/lib/common/watchdog.c
@@ -278,6 +278,10 @@ pcmk__valid_sbd_timeout(const char *value)
{
long st_timeout = value? crm_get_msec(value) : 0;
+ /* @COMPAT At a compatibility break, accept either negative values or a
+ * specific string like "auto" (but not both) to mean "auto-calculate the
+ * timeout." Reject other values that aren't parsable as timeouts.
+ */
if (st_timeout < 0) {
st_timeout = pcmk__auto_watchdog_timeout();
crm_debug("Using calculated value %ld for stonith-watchdog-timeout (%s)",
--
2.31.1
From 28d96fc802bca24ed3e52b8ce5946f4b2b971b7d Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Sat, 6 Jan 2024 17:52:18 -0800
Subject: [PATCH 05/24] Refactor: libcrmcommon: New enum pcmk__opt_context
At first this will be used only for backward compatibility with metadata
commands for pacemaker-based, pacemaker-controld, and
pacemaker-schedulerd. It may be extended later for meta-attribute
contexts or similar.
The idea is that we'll consolidate all cluster options into a single
table and use these enum values as filters for getting daemon metadata.
We won't need a pcmk__opt_context_fenced, because its metadata consists
of stonith instance attributes, not cluster options. Those instance
attributes will be stored in a separate array from the cluster options.
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
include/crm/common/options_internal.h | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/include/crm/common/options_internal.h b/include/crm/common/options_internal.h
index 1ea27ee..0c6c9e8 100644
--- a/include/crm/common/options_internal.h
+++ b/include/crm/common/options_internal.h
@@ -34,6 +34,19 @@ bool pcmk__env_option_enabled(const char *daemon, const char *option);
* Cluster option handling
*/
+/*!
+ * \internal
+ * \enum pcmk__opt_context
+ * \brief Context flags for options
+ */
+enum pcmk__opt_context {
+ // @COMPAT Used only for daemon metadata
+ pcmk__opt_context_none = 0, //!< No additional context
+ pcmk__opt_context_based = (1 << 1), //!< CIB manager metadata
+ pcmk__opt_context_controld = (1 << 2), //!< Controller metadata
+ pcmk__opt_context_schedulerd = (1 << 3), //!< Scheduler metadata
+};
+
typedef struct pcmk__cluster_option_s {
const char *name;
const char *alt_name;
--
2.31.1
From 2f8537331e2948b9186555ffbd9c9f2c121587d1 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Sat, 6 Jan 2024 17:58:25 -0800
Subject: [PATCH 06/24] Refactor: libcrmcommon: New
pcmk__cluster_option_t:context member
Arguably makes more sense adjacent to the type member, but this
placement keeps the diffs smaller when updating existing options arrays.
We will use this soon to ensure that each option occurs in exactly one
daemon's metadata. Several options (for example,
PCMK_OPT_NO_QUORUM_POLICY) currently appear in the metadata of both the
controller and the scheduler, causing issues for external tools that
parse the output.
Where an option currently appears in the metadata of both the controller
and the scheduler, it will soon appear only in the scheduler's metadata.
We assign context flags accordingly.
Note that the fencer doesn't have a context flag. The options in its
metadata are actually stonith device instance attributes, not cluster
options. They will continue to reside in a separate table from the
cluster options, so there's no ambiguity about which daemon they "belong
to."
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/controld/controld_control.c | 20 +++++++++++++++++
daemons/fenced/pacemaker-fenced.c | 26 ++++++++++++++++++++++
include/crm/common/options_internal.h | 3 +++
lib/cib/cib_utils.c | 3 +++
lib/pengine/common.c | 32 +++++++++++++++++++++++++++
5 files changed, 84 insertions(+)
diff --git a/daemons/controld/controld_control.c b/daemons/controld/controld_control.c
index 83b802e..4d7cb14 100644
--- a/daemons/controld/controld_control.c
+++ b/daemons/controld/controld_control.c
@@ -518,22 +518,26 @@ do_recover(long long action,
static pcmk__cluster_option_t controller_options[] = {
/* name, old name, type, allowed values,
* default value, validator,
+ * context,
* short description,
* long description
*/
{
"dc-version", NULL, "string", NULL, PCMK__VALUE_NONE, NULL,
+ pcmk__opt_context_controld,
N_("Pacemaker version on cluster node elected Designated Controller (DC)"),
N_("Includes a hash which identifies the exact changeset the code was "
"built from. Used for diagnostic purposes.")
},
{
"cluster-infrastructure", NULL, "string", NULL, "corosync", NULL,
+ pcmk__opt_context_controld,
N_("The messaging stack on which Pacemaker is currently running"),
N_("Used for informational and diagnostic purposes.")
},
{
"cluster-name", NULL, "string", NULL, NULL, NULL,
+ pcmk__opt_context_controld,
N_("An arbitrary name for the cluster"),
N_("This optional value is mostly for users' convenience as desired "
"in administration, but may also be used in Pacemaker "
@@ -543,6 +547,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
XML_CONFIG_ATTR_DC_DEADTIME, NULL, "time",
NULL, "20s", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
N_("How long to wait for a response from other nodes during start-up"),
N_("The optimal value will depend on the speed and load of your network "
"and the type of switches used.")
@@ -552,6 +557,7 @@ static pcmk__cluster_option_t controller_options[] = {
N_("Zero disables polling, while positive values are an interval in seconds"
"(unless other units are specified, for example \"5min\")"),
"15min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
N_("Polling interval to recheck cluster state and evaluate rules "
"with date specifications"),
N_("Pacemaker is primarily event-driven, and looks ahead to know when to "
@@ -563,6 +569,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
"load-threshold", NULL, "percentage", NULL,
"80%", pcmk__valid_percentage,
+ pcmk__opt_context_controld,
N_("Maximum amount of system load that should be used by cluster nodes"),
N_("The cluster will slow down its recovery process when the amount of "
"system resources used (currently CPU) approaches this limit"),
@@ -570,10 +577,12 @@ static pcmk__cluster_option_t controller_options[] = {
{
"node-action-limit", NULL, "integer", NULL,
"0", pcmk__valid_number,
+ pcmk__opt_context_controld,
N_("Maximum number of jobs that can be scheduled per node "
"(defaults to 2x cores)")
},
{ XML_CONFIG_ATTR_FENCE_REACTION, NULL, "string", NULL, "stop", NULL,
+ pcmk__opt_context_controld,
N_("How a cluster node should react if notified of its own fencing"),
N_("A cluster node may receive notification of its own fencing if fencing "
"is misconfigured, or if fabric fencing is in use that doesn't cut "
@@ -584,6 +593,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
XML_CONFIG_ATTR_ELECTION_FAIL, NULL, "time", NULL,
"2min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
"*** Advanced Use Only ***",
N_("Declare an election failed if it is not decided within this much "
"time. If you need to adjust this value, it probably indicates "
@@ -592,6 +602,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
XML_CONFIG_ATTR_FORCE_QUIT, NULL, "time", NULL,
"20min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
"*** Advanced Use Only ***",
N_("Exit immediately if shutdown does not complete within this much "
"time. If you need to adjust this value, it probably indicates "
@@ -600,6 +611,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
"join-integration-timeout", "crmd-integration-timeout", "time", NULL,
"3min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
"*** Advanced Use Only ***",
N_("If you need to adjust this value, it probably indicates "
"the presence of a bug.")
@@ -607,6 +619,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
"join-finalization-timeout", "crmd-finalization-timeout", "time", NULL,
"30min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
"*** Advanced Use Only ***",
N_("If you need to adjust this value, it probably indicates "
"the presence of a bug.")
@@ -614,6 +627,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
"transition-delay", "crmd-transition-delay", "time", NULL,
"0s", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
N_("*** Advanced Use Only *** Enabling this option will slow down "
"cluster recovery under all conditions"),
N_("Delay cluster recovery for this much time to allow for additional "
@@ -633,6 +647,7 @@ static pcmk__cluster_option_t controller_options[] = {
*/
"stonith-watchdog-timeout", NULL, "time", NULL,
"0", NULL,
+ pcmk__opt_context_controld,
N_("How long before nodes can be assumed to be safely down when "
"watchdog-based self-fencing via SBD is in use"),
N_("If this is set to a positive value, lost nodes are assumed to "
@@ -654,6 +669,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
"stonith-max-attempts", NULL, "integer", NULL,
"10", pcmk__valid_positive_number,
+ pcmk__opt_context_controld,
N_("How many times fencing can fail before it will no longer be "
"immediately re-attempted on a target")
},
@@ -662,11 +678,13 @@ static pcmk__cluster_option_t controller_options[] = {
{
"no-quorum-policy", NULL, "select",
"stop, freeze, ignore, demote, suicide", "stop", pcmk__valid_quorum,
+ pcmk__opt_context_controld,
N_("What to do when the cluster does not have quorum"), NULL
},
{
XML_CONFIG_ATTR_SHUTDOWN_LOCK, NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
+ pcmk__opt_context_controld,
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) "
@@ -680,6 +698,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT, NULL, "time", NULL,
"0", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
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 "
@@ -690,6 +709,7 @@ static pcmk__cluster_option_t controller_options[] = {
{
XML_CONFIG_ATTR_NODE_PENDING_TIMEOUT, NULL, "time", NULL,
"0", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
N_("How long to wait for a node that has joined the cluster to join "
"the controller process group"),
N_("Fence nodes that do not join the controller process group within "
diff --git a/daemons/fenced/pacemaker-fenced.c b/daemons/fenced/pacemaker-fenced.c
index 7c69fb8..b2f4742 100644
--- a/daemons/fenced/pacemaker-fenced.c
+++ b/daemons/fenced/pacemaker-fenced.c
@@ -528,11 +528,13 @@ st_peer_update_callback(enum crm_status_type type, crm_node_t * node, const void
static pcmk__cluster_option_t fencer_options[] = {
/* name, old name, type, allowed values,
* default value, validator,
+ * context,
* short description,
* long description
*/
{
PCMK_STONITH_HOST_ARGUMENT, NULL, "string", NULL, "port", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: An alternate parameter to supply instead of 'port'"),
N_("some devices do not support the "
"standard 'port' parameter or may provide additional ones. Use "
@@ -543,17 +545,20 @@ static pcmk__cluster_option_t fencer_options[] = {
},
{
PCMK_STONITH_HOST_MAP,NULL, "string", NULL, "", NULL,
+ pcmk__opt_context_none,
N_("A mapping of host names to ports numbers for devices that do not support host names."),
N_("Eg. node1:1;node2:2,3 would tell the cluster to use port 1 for node1 and ports 2 and 3 for node2")
},
{
PCMK_STONITH_HOST_LIST,NULL, "string", NULL, "", NULL,
+ pcmk__opt_context_none,
N_("Eg. node1,node2,node3"),
N_("A list of machines controlled by "
"this device (Optional unless pcmk_host_list=static-list)")
},
{
PCMK_STONITH_HOST_CHECK,NULL, "string", NULL, "dynamic-list", NULL,
+ pcmk__opt_context_none,
N_("How to determine which machines are controlled by the device."),
N_("Allowed values: dynamic-list "
"(query the device via the 'list' command), static-list "
@@ -564,6 +569,7 @@ static pcmk__cluster_option_t fencer_options[] = {
},
{
PCMK_STONITH_DELAY_MAX,NULL, "time", NULL, "0s", NULL,
+ pcmk__opt_context_none,
N_("Enable a base delay for fencing actions and specify base delay value."),
N_("Enable a delay of no more than the "
"time specified before executing fencing actions. Pacemaker "
@@ -573,6 +579,7 @@ static pcmk__cluster_option_t fencer_options[] = {
},
{
PCMK_STONITH_DELAY_BASE,NULL, "string", NULL, "0s", NULL,
+ pcmk__opt_context_none,
N_("Enable a base delay for "
"fencing actions and specify base delay value."),
N_("This enables a static delay for "
@@ -587,6 +594,7 @@ static pcmk__cluster_option_t fencer_options[] = {
},
{
PCMK_STONITH_ACTION_LIMIT,NULL, "integer", NULL, "1", NULL,
+ pcmk__opt_context_none,
N_("The maximum number of actions can be performed in parallel on this device"),
N_("Cluster property concurrent-fencing=true needs to be configured first."
"Then use this to specify the maximum number of actions can be performed in parallel on this device. -1 is unlimited.")
@@ -594,18 +602,21 @@ static pcmk__cluster_option_t fencer_options[] = {
{
"pcmk_reboot_action", NULL, "string", NULL,
PCMK_ACTION_REBOOT, NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: An alternate command to run instead of 'reboot'"),
N_("Some devices do not support the standard commands or may provide additional ones.\n"
"Use this to specify an alternate, device-specific, command that implements the \'reboot\' action.")
},
{
"pcmk_reboot_timeout",NULL, "time", NULL, "60s", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: Specify an alternate timeout to use for reboot actions instead of stonith-timeout"),
N_("Some devices need much more/less time to complete than normal."
"Use this to specify an alternate, device-specific, timeout for \'reboot\' actions.")
},
{
"pcmk_reboot_retries",NULL, "integer", NULL, "2", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: The maximum number of times to retry the 'reboot' command within the timeout period"),
N_("Some devices do not support multiple connections."
" Operations may 'fail' if the device is busy with another task so Pacemaker will automatically retry the operation, if there is time remaining."
@@ -614,18 +625,21 @@ static pcmk__cluster_option_t fencer_options[] = {
{
"pcmk_off_action", NULL, "string", NULL,
PCMK_ACTION_OFF, NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: An alternate command to run instead of \'off\'"),
N_("Some devices do not support the standard commands or may provide additional ones."
"Use this to specify an alternate, device-specific, command that implements the \'off\' action.")
},
{
"pcmk_off_timeout",NULL, "time", NULL, "60s", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: Specify an alternate timeout to use for off actions instead of stonith-timeout"),
N_("Some devices need much more/less time to complete than normal."
"Use this to specify an alternate, device-specific, timeout for \'off\' actions.")
},
{
"pcmk_off_retries",NULL, "integer", NULL, "2", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: The maximum number of times to retry the 'off' command within the timeout period"),
N_("Some devices do not support multiple connections."
" Operations may 'fail' if the device is busy with another task so Pacemaker will automatically retry the operation, if there is time remaining."
@@ -634,18 +648,21 @@ static pcmk__cluster_option_t fencer_options[] = {
{
"pcmk_on_action", NULL, "string", NULL,
PCMK_ACTION_ON, NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: An alternate command to run instead of 'on'"),
N_("Some devices do not support the standard commands or may provide additional ones."
"Use this to specify an alternate, device-specific, command that implements the \'on\' action.")
},
{
"pcmk_on_timeout",NULL, "time", NULL, "60s", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: Specify an alternate timeout to use for on actions instead of stonith-timeout"),
N_("Some devices need much more/less time to complete than normal."
"Use this to specify an alternate, device-specific, timeout for \'on\' actions.")
},
{
"pcmk_on_retries",NULL, "integer", NULL, "2", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: The maximum number of times to retry the 'on' command within the timeout period"),
N_("Some devices do not support multiple connections."
" Operations may 'fail' if the device is busy with another task so Pacemaker will automatically retry the operation, if there is time remaining."
@@ -654,18 +671,21 @@ static pcmk__cluster_option_t fencer_options[] = {
{
"pcmk_list_action",NULL, "string", NULL,
PCMK_ACTION_LIST, NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: An alternate command to run instead of \'list\'"),
N_("Some devices do not support the standard commands or may provide additional ones."
"Use this to specify an alternate, device-specific, command that implements the \'list\' action.")
},
{
"pcmk_list_timeout",NULL, "time", NULL, "60s", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: Specify an alternate timeout to use for list actions instead of stonith-timeout"),
N_("Some devices need much more/less time to complete than normal."
"Use this to specify an alternate, device-specific, timeout for \'list\' actions.")
},
{
"pcmk_list_retries",NULL, "integer", NULL, "2", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: The maximum number of times to retry the \'list\' command within the timeout period"),
N_("Some devices do not support multiple connections."
" Operations may 'fail' if the device is busy with another task so Pacemaker will automatically retry the operation, if there is time remaining."
@@ -674,18 +694,21 @@ static pcmk__cluster_option_t fencer_options[] = {
{
"pcmk_monitor_action", NULL, "string", NULL,
PCMK_ACTION_MONITOR, NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: An alternate command to run instead of \'monitor\'"),
N_("Some devices do not support the standard commands or may provide additional ones."
"Use this to specify an alternate, device-specific, command that implements the \'monitor\' action.")
},
{
"pcmk_monitor_timeout",NULL, "time", NULL, "60s", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: Specify an alternate timeout to use for monitor actions instead of stonith-timeout"),
N_("Some devices need much more/less time to complete than normal.\n"
"Use this to specify an alternate, device-specific, timeout for \'monitor\' actions.")
},
{
"pcmk_monitor_retries",NULL, "integer", NULL, "2", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: The maximum number of times to retry the \'monitor\' command within the timeout period"),
N_("Some devices do not support multiple connections."
" Operations may 'fail' if the device is busy with another task so Pacemaker will automatically retry the operation, if there is time remaining."
@@ -694,18 +717,21 @@ static pcmk__cluster_option_t fencer_options[] = {
{
"pcmk_status_action", NULL, "string", NULL,
PCMK_ACTION_STATUS, NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: An alternate command to run instead of \'status\'"),
N_("Some devices do not support the standard commands or may provide additional ones."
"Use this to specify an alternate, device-specific, command that implements the \'status\' action.")
},
{
"pcmk_status_timeout",NULL, "time", NULL, "60s", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: Specify an alternate timeout to use for status actions instead of stonith-timeout"),
N_("Some devices need much more/less time to complete than normal."
"Use this to specify an alternate, device-specific, timeout for \'status\' actions.")
},
{
"pcmk_status_retries",NULL, "integer", NULL, "2", NULL,
+ pcmk__opt_context_none,
N_("Advanced use only: The maximum number of times to retry the \'status\' command within the timeout period"),
N_("Some devices do not support multiple connections."
" Operations may 'fail' if the device is busy with another task so Pacemaker will automatically retry the operation, if there is time remaining."
diff --git a/include/crm/common/options_internal.h b/include/crm/common/options_internal.h
index 0c6c9e8..b2525ef 100644
--- a/include/crm/common/options_internal.h
+++ b/include/crm/common/options_internal.h
@@ -56,6 +56,9 @@ typedef struct pcmk__cluster_option_s {
bool (*is_valid)(const char *);
+ // @COMPAT context is used only for daemon meta-data
+ enum pcmk__opt_context context;
+
const char *description_short;
const char *description_long;
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index 9c3f9f1..9e4060b 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -869,18 +869,21 @@ cib_native_notify(gpointer data, gpointer user_data)
static pcmk__cluster_option_t cib_opts[] = {
/* name, legacy name, type, allowed values,
* default value, validator,
+ * context,
* short description,
* long description
*/
{
"enable-acl", NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
+ pcmk__opt_context_based,
N_("Enable Access Control Lists (ACLs) for the CIB"),
NULL
},
{
"cluster-ipc-limit", NULL, "integer", NULL,
"500", pcmk__valid_positive_number,
+ pcmk__opt_context_based,
N_("Maximum IPC message backlog before disconnecting a cluster daemon"),
N_("Raise this if log has \"Evicting client\" messages for cluster daemon"
" PIDs (a good value is the number of resources in the cluster"
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index 6878f4d..383c4af 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -24,24 +24,28 @@ gboolean was_processing_warning = FALSE;
static pcmk__cluster_option_t pe_opts[] = {
/* name, old name, type, allowed values,
* default value, validator,
+ * context,
* short description,
* long description
*/
{
"no-quorum-policy", NULL, "select", "stop, freeze, ignore, demote, suicide",
"stop", pcmk__valid_quorum,
+ pcmk__opt_context_schedulerd,
N_("What to do when the cluster does not have quorum"),
NULL
},
{
"symmetric-cluster", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
N_("Whether resources can run on any node by default"),
NULL
},
{
"maintenance-mode", NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
N_("Whether the cluster should refrain from monitoring, starting, "
"and stopping resources"),
NULL
@@ -49,6 +53,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"start-failure-is-fatal", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
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 "
@@ -58,12 +63,14 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"enable-startup-probes", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
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,
+ pcmk__opt_context_schedulerd,
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) "
@@ -77,6 +84,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT, NULL, "time", NULL,
"0", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
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 "
@@ -89,6 +97,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"stonith-enabled", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
N_("*** Advanced Use Only *** "
"Whether nodes may be fenced as part of recovery"),
N_("If false, unresponsive nodes are immediately assumed to be harmless, "
@@ -99,6 +108,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"stonith-action", NULL, "select", "reboot, off, poweroff",
PCMK_ACTION_REBOOT, pcmk__is_fencing_action,
+ pcmk__opt_context_schedulerd,
N_("Action to send to fence device when a node needs to be fenced "
"(\"poweroff\" is a deprecated alias for \"off\")"),
NULL
@@ -106,6 +116,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"stonith-timeout", NULL, "time", NULL,
"60s", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
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.")
@@ -113,6 +124,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
XML_ATTR_HAVE_WATCHDOG, NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
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. "
@@ -124,12 +136,14 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"concurrent-fencing", NULL, "boolean", NULL,
PCMK__CONCURRENT_FENCING_DEFAULT, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
N_("Allow performing fencing operations in parallel"),
NULL
},
{
"startup-fencing", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
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.")
@@ -137,6 +151,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY, NULL, "time", NULL,
"0", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
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 "
@@ -154,6 +169,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
XML_CONFIG_ATTR_NODE_PENDING_TIMEOUT, NULL, "time", NULL,
"0", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
N_("How long to wait for a node that has joined the cluster to join "
"the controller process group"),
N_("Fence nodes that do not join the controller process group within "
@@ -165,6 +181,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"cluster-delay", NULL, "time", NULL,
"60s", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
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 "
@@ -175,6 +192,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"batch-limit", NULL, "integer", NULL,
"0", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
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 "
@@ -185,6 +203,7 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"migration-limit", NULL, "integer", NULL,
"-1", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
N_("The number of live migration actions that the cluster is allowed "
"to execute in parallel on a node (-1 means no limit)")
},
@@ -193,24 +212,28 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"stop-all-resources", NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
N_("Whether the cluster should stop all active resources"),
NULL
},
{
"stop-orphan-resources", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
N_("Whether to stop resources that were removed from the configuration"),
NULL
},
{
"stop-orphan-actions", NULL, "boolean", NULL,
"true", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
N_("Whether to cancel recurring actions removed from the configuration"),
NULL
},
{
"remove-after-stop", NULL, "boolean", NULL,
"false", pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
N_("*** Deprecated *** Whether to remove stopped resources from "
"the executor"),
N_("Values other than default are poorly tested and potentially dangerous."
@@ -221,18 +244,21 @@ static pcmk__cluster_option_t pe_opts[] = {
{
"pe-error-series-max", NULL, "integer", NULL,
"-1", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
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,
+ pcmk__opt_context_schedulerd,
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,
+ pcmk__opt_context_schedulerd,
N_("The number of scheduler inputs without errors or warnings to save"),
N_("Zero to disable, -1 to store unlimited.")
},
@@ -244,6 +270,7 @@ static pcmk__cluster_option_t pe_opts[] = {
PCMK__VALUE_ONLY_GREEN ", " PCMK__VALUE_PROGRESSIVE ", "
PCMK__VALUE_CUSTOM,
PCMK__VALUE_NONE, pcmk__validate_health_strategy,
+ pcmk__opt_context_schedulerd,
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\", "
@@ -252,24 +279,28 @@ static pcmk__cluster_option_t pe_opts[] = {
{
PCMK__OPT_NODE_HEALTH_BASE, NULL, "integer", NULL,
"0", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
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,
+ pcmk__opt_context_schedulerd,
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,
+ pcmk__opt_context_schedulerd,
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,
+ pcmk__opt_context_schedulerd,
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\".")
},
@@ -279,6 +310,7 @@ static pcmk__cluster_option_t pe_opts[] = {
"placement-strategy", NULL, "select",
"default, utilization, minimal, balanced",
"default", pcmk__valid_placement_strategy,
+ pcmk__opt_context_schedulerd,
N_("How the cluster should allocate resources to nodes"),
NULL
},
--
2.31.1
From cc7c3c87d333854d0f28abe461dd58d5c94b0888 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Sat, 6 Jan 2024 18:47:25 -0800
Subject: [PATCH 07/24] Refactor: libcrmcommon: Consolidate cluster option
metadata
This isn't plugged in yet. It's also currently defined out, to avoid an
unused variable warning from the compiler.
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
lib/common/options.c | 499 +++++++++++++++++++++++++++++++++++++++++++
1 file changed, 499 insertions(+)
diff --git a/lib/common/options.c b/lib/common/options.c
index 1db41a2..ff73dcc 100644
--- a/lib/common/options.c
+++ b/lib/common/options.c
@@ -20,6 +20,7 @@
#include <sys/stat.h>
#include <crm/crm.h>
+#include <crm/msg_xml.h>
void
pcmk__cli_help(char cmd)
@@ -38,6 +39,504 @@ pcmk__cli_help(char cmd)
}
+/*
+ * Option metadata
+ */
+
+#if 0
+static pcmk__cluster_option_t cluster_options[] = {
+ /* name, old name, type, allowed values,
+ * default value, validator,
+ * context,
+ * short description,
+ * long description
+ */
+ {
+ "dc-version", NULL, "string", NULL,
+ PCMK__VALUE_NONE, NULL,
+ pcmk__opt_context_controld,
+ N_("Pacemaker version on cluster node elected Designated Controller "
+ "(DC)"),
+ N_("Includes a hash which identifies the exact changeset the code was "
+ "built from. Used for diagnostic purposes."),
+ },
+ {
+ "cluster-infrastructure", NULL, "string", NULL,
+ "corosync", NULL,
+ pcmk__opt_context_controld,
+ N_("The messaging stack on which Pacemaker is currently running"),
+ N_("Used for informational and diagnostic purposes."),
+ },
+ {
+ "cluster-name", NULL, "string", NULL,
+ NULL, NULL,
+ pcmk__opt_context_controld,
+ N_("An arbitrary name for the cluster"),
+ N_("This optional value is mostly for users' convenience as desired "
+ "in administration, but may also be used in Pacemaker "
+ "configuration rules via the #cluster-name node attribute, and "
+ "by higher-level tools and resource agents."),
+ },
+ {
+ "dc-deadtime", NULL, "time", NULL,
+ "20s", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
+ N_("How long to wait for a response from other nodes during start-up"),
+ N_("The optimal value will depend on the speed and load of your "
+ "network and the type of switches used."),
+ },
+ {
+ "cluster-recheck-interval", NULL, "time",
+ N_("Zero disables polling, while positive values are an interval in "
+ "seconds (unless other units are specified, for example \"5min\")"),
+ "15min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
+ N_("Polling interval to recheck cluster state and evaluate rules "
+ "with date specifications"),
+ N_("Pacemaker is primarily event-driven, and looks ahead to know when "
+ "to recheck cluster state for failure timeouts and most time-based "
+ "rules. However, it will also recheck the cluster after this "
+ "amount of inactivity, to evaluate rules with date specifications "
+ "and serve as a fail-safe for certain types of scheduler bugs."),
+ },
+ {
+ "fence-reaction", NULL, "select", "stop, panic",
+ "stop", NULL,
+ pcmk__opt_context_controld,
+ N_("How a cluster node should react if notified of its own fencing"),
+ N_("A cluster node may receive notification of its own fencing if "
+ "fencing is misconfigured, or if fabric fencing is in use that "
+ "doesn't cut cluster communication. Use \"stop\" to attempt to "
+ "immediately stop Pacemaker and stay stopped, or \"panic\" to "
+ "attempt to immediately reboot the local node, falling back to "
+ "stop on failure."),
+ },
+ {
+ "election-timeout", NULL, "time", NULL,
+ "2min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
+ N_("*** Advanced Use Only ***"),
+ N_("Declare an election failed if it is not decided within this much "
+ "time. If you need to adjust this value, it probably indicates "
+ "the presence of a bug."),
+ },
+ {
+ "shutdown-escalation", NULL, "time", NULL,
+ "20min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
+ N_("*** Advanced Use Only ***"),
+ N_("Exit immediately if shutdown does not complete within this much "
+ "time. If you need to adjust this value, it probably indicates "
+ "the presence of a bug."),
+ },
+ {
+ "join-integration-timeout", "crmd-integration-timeout", "time",
+ NULL,
+ "3min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
+ N_("*** Advanced Use Only ***"),
+ N_("If you need to adjust this value, it probably indicates "
+ "the presence of a bug."),
+ },
+ {
+ "join-finalization-timeout", "crmd-finalization-timeout",
+ "time", NULL,
+ "30min", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
+ N_("*** Advanced Use Only ***"),
+ N_("If you need to adjust this value, it probably indicates "
+ "the presence of a bug."),
+ },
+ {
+ "transition-delay", "crmd-transition-delay", "time", NULL,
+ "0s", pcmk__valid_interval_spec,
+ pcmk__opt_context_controld,
+ N_("*** Advanced Use Only *** "
+ "Enabling this option will slow down cluster recovery under all "
+ "conditions"),
+ N_("Delay cluster recovery for this much time to allow for additional "
+ "events to occur. Useful if your configuration is sensitive to "
+ "the order in which ping updates arrive."),
+ },
+ {
+ "no-quorum-policy", NULL, "select",
+ "stop, freeze, ignore, demote, suicide",
+ "stop", pcmk__valid_quorum,
+ pcmk__opt_context_schedulerd,
+ N_("What to do when the cluster does not have quorum"),
+ NULL,
+ },
+ {
+ "shutdown-lock", NULL, "boolean", NULL,
+ XML_BOOLEAN_FALSE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ 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."),
+ },
+ {
+ "shutdown-lock-limit", NULL, "time", NULL,
+ "0", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
+ 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."),
+ },
+ {
+ "enable-acl", NULL, "boolean", NULL,
+ XML_BOOLEAN_FALSE, pcmk__valid_boolean,
+ pcmk__opt_context_based,
+ N_("Enable Access Control Lists (ACLs) for the CIB"),
+ NULL,
+ },
+ {
+ "symmetric-cluster", NULL, "boolean", NULL,
+ XML_BOOLEAN_TRUE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ N_("Whether resources can run on any node by default"),
+ NULL,
+ },
+ {
+ "maintenance-mode", NULL, "boolean", NULL,
+ XML_BOOLEAN_FALSE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ N_("Whether the cluster should refrain from monitoring, starting, and "
+ "stopping resources"),
+ NULL,
+ },
+ {
+ "start-failure-is-fatal", NULL, "boolean", NULL,
+ XML_BOOLEAN_TRUE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ 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,
+ XML_BOOLEAN_TRUE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ N_("Whether the cluster should check for active resources during "
+ "start-up"),
+ NULL,
+ },
+
+ // Fencing-related options
+ {
+ "stonith-enabled", NULL, "boolean", NULL,
+ XML_BOOLEAN_TRUE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ 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",
+ PCMK_ACTION_REBOOT, pcmk__is_fencing_action,
+ pcmk__opt_context_schedulerd,
+ 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,
+ pcmk__opt_context_schedulerd,
+ N_("How long to wait for on, off, and reboot fence actions to complete "
+ "by default"),
+ NULL,
+ },
+ {
+ "have-watchdog", NULL, "boolean", NULL,
+ XML_BOOLEAN_FALSE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ 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."),
+ },
+ {
+ /* @COMPAT Currently, unparsable values default to -1 (auto-calculate),
+ * while missing values default to 0 (disable). All values are accepted
+ * (unless the controller finds that the value conflicts with the
+ * SBD_WATCHDOG_TIMEOUT).
+ *
+ * At a compatibility break: properly validate as a timeout, let
+ * either negative values or a particular string like "auto" mean auto-
+ * calculate, and use 0 as the single default for when the option either
+ * is unset or fails to validate.
+ */
+ "stonith-watchdog-timeout", NULL, "time", NULL,
+ "0", NULL,
+ pcmk__opt_context_controld,
+ N_("How long before nodes can be assumed to be safely down when "
+ "watchdog-based self-fencing via SBD is in use"),
+ N_("If this is set to a positive value, lost nodes are assumed to "
+ "self-fence using watchdog-based SBD within this much time. This "
+ "does not require a fencing resource to be explicitly configured, "
+ "though a fence_watchdog resource can be configured, to limit use "
+ "to specific nodes. If this is set to 0 (the default), the cluster "
+ "will never assume watchdog-based self-fencing. If this is set to a "
+ "negative value, the cluster will use twice the local value of the "
+ "`SBD_WATCHDOG_TIMEOUT` environment variable if that is positive, "
+ "or otherwise treat this as 0. WARNING: When used, this timeout "
+ "must be larger than `SBD_WATCHDOG_TIMEOUT` on all nodes that use "
+ "watchdog-based SBD, and Pacemaker will refuse to start on any of "
+ "those nodes where this is not true for the local value or SBD is "
+ "not active. When this is set to a negative value, "
+ "`SBD_WATCHDOG_TIMEOUT` must be set to the same value on all nodes "
+ "that use SBD, otherwise data corruption or loss could occur."),
+ },
+ {
+ "stonith-max-attempts", NULL, "integer", NULL,
+ "10", pcmk__valid_positive_number,
+ pcmk__opt_context_controld,
+ N_("How many times fencing can fail before it will no longer be "
+ "immediately re-attempted on a target"),
+ NULL,
+ },
+ {
+ "concurrent-fencing", NULL, "boolean", NULL,
+ PCMK__CONCURRENT_FENCING_DEFAULT, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ N_("Allow performing fencing operations in parallel"),
+ NULL,
+ },
+ {
+ "startup-fencing", NULL, "boolean", NULL,
+ XML_BOOLEAN_TRUE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ 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."),
+ },
+ {
+ "priority-fencing-delay", NULL, "time", NULL,
+ "0", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
+ 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."),
+ },
+ {
+ "node-pending-timeout", NULL, "time", NULL,
+ "0", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
+ N_("How long to wait for a node that has joined the cluster to join "
+ "the controller process group"),
+ N_("Fence nodes that do not join the controller process group within "
+ "this much time after joining the cluster, to allow the cluster "
+ "to continue managing resources. A value of 0 means never fence "
+ "pending nodes. Setting the value to 2h means fence nodes after "
+ "2 hours."),
+ },
+ {
+ "cluster-delay", NULL, "time", NULL,
+ "60s", pcmk__valid_interval_spec,
+ pcmk__opt_context_schedulerd,
+ 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.")
+ },
+
+ // Limits
+ {
+ "load-threshold", NULL, "percentage", NULL,
+ "80%", pcmk__valid_percentage,
+ pcmk__opt_context_controld,
+ N_("Maximum amount of system load that should be used by cluster "
+ "nodes"),
+ N_("The cluster will slow down its recovery process when the amount of "
+ "system resources used (currently CPU) approaches this limit"),
+ },
+ {
+ "node-action-limit", NULL, "integer", NULL,
+ "0", pcmk__valid_number,
+ pcmk__opt_context_controld,
+ N_("Maximum number of jobs that can be scheduled per node (defaults to "
+ "2x cores)"),
+ NULL,
+ },
+ {
+ "batch-limit", NULL, "integer", NULL,
+ "0", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
+ 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,
+ pcmk__opt_context_schedulerd,
+ N_("The number of live migration actions that the cluster is allowed "
+ "to execute in parallel on a node (-1 means no limit)"),
+ NULL,
+ },
+ {
+ "cluster-ipc-limit", NULL, "integer", NULL,
+ "500", pcmk__valid_positive_number,
+ pcmk__opt_context_based,
+ N_("Maximum IPC message backlog before disconnecting a cluster daemon"),
+ N_("Raise this if log has \"Evicting client\" messages for cluster "
+ "daemon PIDs (a good value is the number of resources in the "
+ "cluster multiplied by the number of nodes)."),
+ },
+
+ // Orphans and stopping
+ {
+ "stop-all-resources", NULL, "boolean", NULL,
+ XML_BOOLEAN_FALSE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ N_("Whether the cluster should stop all active resources"),
+ NULL,
+ },
+ {
+ "stop-orphan-resources", NULL, "boolean", NULL,
+ XML_BOOLEAN_TRUE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ N_("Whether to stop resources that were removed from the "
+ "configuration"),
+ NULL,
+ },
+ {
+ "stop-orphan-actions", NULL, "boolean", NULL,
+ XML_BOOLEAN_TRUE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ N_("Whether to cancel recurring actions removed from the "
+ "configuration"),
+ NULL,
+ },
+ {
+ "remove-after-stop", NULL, "boolean", NULL,
+ XML_BOOLEAN_FALSE, pcmk__valid_boolean,
+ pcmk__opt_context_schedulerd,
+ 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,
+ pcmk__opt_context_schedulerd,
+ 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,
+ pcmk__opt_context_schedulerd,
+ 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,
+ pcmk__opt_context_schedulerd,
+ N_("The number of scheduler inputs without errors or warnings to save"),
+ N_("Zero to disable, -1 to store unlimited."),
+ },
+
+ // Node health
+ {
+ "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,
+ pcmk__opt_context_schedulerd,
+ 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\".")
+ },
+ {
+ "node-health-base", NULL, "integer", NULL,
+ "0", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
+ N_("Base health score assigned to a node"),
+ N_("Only used when \"node-health-strategy\" is set to "
+ "\"progressive\"."),
+ },
+ {
+ "node-health-green", NULL, "integer", NULL,
+ "0", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
+ 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\"."),
+ },
+ {
+ "node-health-yellow", NULL, "integer", NULL,
+ "0", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
+ 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\"."),
+ },
+ {
+ "node-health-red", NULL, "integer", NULL,
+ "-INFINITY", pcmk__valid_number,
+ pcmk__opt_context_schedulerd,
+ 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", pcmk__valid_placement_strategy,
+ pcmk__opt_context_schedulerd,
+ N_("How the cluster should allocate resources to nodes"),
+ NULL,
+ },
+};
+#endif // 0
+
+
/*
* Environment variable option handling
*/
--
2.31.1
From 96fa08b7adc911cce417f7f9889029510ec1c428 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Wed, 10 Jan 2024 23:35:22 -0800
Subject: [PATCH 08/24] Refactor: libcrmcommon: New filter arg to
pcmk__format_option_metadata()
Now each cluster option is in exactly one daemon's metadata. The four
options that were previously in the metadata of both the controller and
the scheduler are now only in the scheduler's metadata.
All daemons still have access to all the options they use.
All function calls in daemons still use the local options arrays rather
than the one in libcrmcommon. That will change in upcoming commits.
Closes T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/controld/controld_control.c | 6 ++++--
daemons/fenced/pacemaker-fenced.c | 17 +++++++++-------
include/crm/common/options_internal.h | 1 +
lib/cib/cib_utils.c | 7 ++++---
lib/common/options.c | 28 +++++++++++++++++++++++++--
lib/pengine/common.c | 7 ++++---
6 files changed, 49 insertions(+), 17 deletions(-)
diff --git a/daemons/controld/controld_control.c b/daemons/controld/controld_control.c
index 4d7cb14..8fe09da 100644
--- a/daemons/controld/controld_control.c
+++ b/daemons/controld/controld_control.c
@@ -723,11 +723,13 @@ static pcmk__cluster_option_t controller_options[] = {
void
crmd_metadata(void)
{
+ const char *name = "pacemaker-controld";
const char *desc_short = "Pacemaker controller options";
const char *desc_long = "Cluster options used by Pacemaker's controller";
- gchar *s = pcmk__format_option_metadata("pacemaker-controld", desc_short,
- desc_long, controller_options,
+ gchar *s = pcmk__format_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_controld,
+ controller_options,
PCMK__NELEM(controller_options));
printf("%s", s);
g_free(s);
diff --git a/daemons/fenced/pacemaker-fenced.c b/daemons/fenced/pacemaker-fenced.c
index b2f4742..d0b6c31 100644
--- a/daemons/fenced/pacemaker-fenced.c
+++ b/daemons/fenced/pacemaker-fenced.c
@@ -742,14 +742,17 @@ static pcmk__cluster_option_t fencer_options[] = {
void
fencer_metadata(void)
{
+ const char *name = "pacemaker-fenced";
const char *desc_short = N_("Instance attributes available for all "
- "\"stonith\"-class resources");
- const char *desc_long = N_("Instance attributes available for all \"stonith\"-"
- "class resources and used by Pacemaker's fence "
- "daemon, formerly known as stonithd");
-
- gchar *s = pcmk__format_option_metadata("pacemaker-fenced", desc_short,
- desc_long, fencer_options,
+ "\"stonith\"-class resources");
+ const char *desc_long = N_("Instance attributes available for all "
+ "\"stonith\"-class resources and used by "
+ "Pacemaker's fence daemon, formerly known as "
+ "stonithd");
+
+ gchar *s = pcmk__format_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_none,
+ fencer_options,
PCMK__NELEM(fencer_options));
printf("%s", s);
g_free(s);
diff --git a/include/crm/common/options_internal.h b/include/crm/common/options_internal.h
index b2525ef..89d27d1 100644
--- a/include/crm/common/options_internal.h
+++ b/include/crm/common/options_internal.h
@@ -70,6 +70,7 @@ const char *pcmk__cluster_option(GHashTable *options,
gchar *pcmk__format_option_metadata(const char *name, const char *desc_short,
const char *desc_long,
+ enum pcmk__opt_context filter,
pcmk__cluster_option_t *option_list,
int len);
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index 9e4060b..2205d15 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -894,13 +894,14 @@ static pcmk__cluster_option_t cib_opts[] = {
void
cib_metadata(void)
{
+ const char *name = "pacemaker-based";
const char *desc_short = "Cluster Information Base manager options";
const char *desc_long = "Cluster options used by Pacemaker's Cluster "
"Information Base manager";
- gchar *s = pcmk__format_option_metadata("pacemaker-based", desc_short,
- desc_long, cib_opts,
- PCMK__NELEM(cib_opts));
+ gchar *s = pcmk__format_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_based,
+ cib_opts, PCMK__NELEM(cib_opts));
printf("%s", s);
g_free(s);
}
diff --git a/lib/common/options.c b/lib/common/options.c
index ff73dcc..d5b6c17 100644
--- a/lib/common/options.c
+++ b/lib/common/options.c
@@ -937,13 +937,32 @@ add_desc(GString *s, const char *tag, const char *desc, const char *values,
free(escaped_en);
}
+/*!
+ * \internal
+ * \brief Format option metadata as an OCF-like XML string
+ *
+ * \param[in] name Daemon name
+ * \param[in] desc_short Short description of the daemon
+ * \param[in] desc_long Long description of the daemon
+ * \param[in] filter If not \c pcmk__opt_context_none, include only
+ * those options whose \c context field is equal to
+ * \p filter
+ * \param[in] option_list Options whose metadata to format
+ * \param[in] len Number of items in \p option_list
+ *
+ * \return A string containing OCF-like option metadata XML
+ *
+ * \note The caller is responsible for freeing the return value using
+ * \c g_free().
+ */
gchar *
pcmk__format_option_metadata(const char *name, const char *desc_short,
const char *desc_long,
+ enum pcmk__opt_context filter,
pcmk__cluster_option_t *option_list, int len)
{
- /* big enough to hold "pacemaker-schedulerd metadata" output */
- GString *s = g_string_sized_new(13000);
+ // Large enough to hold current cluster options with room for growth (2^15)
+ GString *s = g_string_sized_new(32768);
pcmk__g_strcat(s,
"<?xml version=\"1.0\"?>\n"
@@ -964,6 +983,11 @@ pcmk__format_option_metadata(const char *name, const char *desc_short,
const char *opt_desc_short = option_list[lpc].description_short;
const char *opt_desc_long = option_list[lpc].description_long;
+ if ((filter != pcmk__opt_context_none)
+ && (filter != option_list[lpc].context)) {
+ continue;
+ }
+
// The standard requires long and short parameter descriptions
CRM_ASSERT((opt_desc_short != NULL) || (opt_desc_long != NULL));
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index 383c4af..e9aa2e2 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -319,12 +319,13 @@ static pcmk__cluster_option_t pe_opts[] = {
void
pe_metadata(pcmk__output_t *out)
{
+ const char *name = "pacemaker-schedulerd";
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));
+ gchar *s = pcmk__format_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_schedulerd,
+ pe_opts, PCMK__NELEM(pe_opts));
out->output_xml(out, "metadata", s);
g_free(s);
}
--
2.31.1
From 96b59bf0c66fccc0656a9195ebe7580d54083eb3 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Wed, 10 Jan 2024 23:59:39 -0800
Subject: [PATCH 09/24] Test: cts-cli: Update daemon outputs for option
filtering
Now each cluster option is in exactly one daemon's metadata. The four
options that were previously in the metadata of both the controller and
the scheduler are now only in the scheduler's metadata.
All daemons still have access to all the options they use.
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
cts/cli/regression.daemons.exp | 26 --------------------------
1 file changed, 26 deletions(-)
diff --git a/cts/cli/regression.daemons.exp b/cts/cli/regression.daemons.exp
index 6a24089..9b2dd96 100644
--- a/cts/cli/regression.daemons.exp
+++ b/cts/cli/regression.daemons.exp
@@ -101,32 +101,6 @@
<shortdesc lang="en">How many times fencing can fail before it will no longer be immediately re-attempted on a target</shortdesc>
<content type="integer" default=""/>
</parameter>
- <parameter name="no-quorum-policy">
- <longdesc lang="en">What to do when the cluster does not have quorum Allowed values: stop, freeze, ignore, demote, suicide</longdesc>
- <shortdesc lang="en">What to do when the cluster does not have quorum</shortdesc>
- <content type="select" default="">
- <option value="stop" />
- <option value="freeze" />
- <option value="ignore" />
- <option value="demote" />
- <option value="suicide" />
- </content>
- </parameter>
- <parameter name="shutdown-lock">
- <longdesc lang="en">When true, resources active on a node when it is cleanly shut down are kept &quot;locked&quot; 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.</longdesc>
- <shortdesc lang="en">Whether to lock resources to a cleanly shut down node</shortdesc>
- <content type="boolean" default=""/>
- </parameter>
- <parameter name="shutdown-lock-limit">
- <longdesc lang="en">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.</longdesc>
- <shortdesc lang="en">Do not lock resources to a cleanly shut down node longer than this</shortdesc>
- <content type="time" default=""/>
- </parameter>
- <parameter name="node-pending-timeout">
- <longdesc lang="en">Fence nodes that do not join the controller process group within this much time after joining the cluster, to allow the cluster to continue managing resources. A value of 0 means never fence pending nodes. Setting the value to 2h means fence nodes after 2 hours.</longdesc>
- <shortdesc lang="en">How long to wait for a node that has joined the cluster to join the controller process group</shortdesc>
- <content type="time" default=""/>
- </parameter>
</parameters>
</resource-agent>
=#=#=#= End test: Get controller metadata - OK (0) =#=#=#=
--
2.31.1
From fec945824ed11395a8366882c29315c509de80f0 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Sat, 6 Jan 2024 19:07:00 -0800
Subject: [PATCH 10/24] Refactor: libcrmcommon, daemons: New
pcmk__cluster_option_metadata()
This new function is a wrapper for pcmk__format_option_metadata() that
always uses the shared cluster_options array and its length.
Daemons can now call this function to get metadata instead of using
their local options arrays for that purpose.
Soon we'll introduce a command that outputs all cluster option metadata
directly, instead of calling daemon metadata commands.
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/controld/controld_control.c | 7 +++----
include/crm/common/options_internal.h | 4 ++++
lib/cib/cib_utils.c | 6 +++---
lib/common/options.c | 28 +++++++++++++++++++++++++--
lib/pengine/common.c | 6 +++---
5 files changed, 39 insertions(+), 12 deletions(-)
diff --git a/daemons/controld/controld_control.c b/daemons/controld/controld_control.c
index 8fe09da..82aa143 100644
--- a/daemons/controld/controld_control.c
+++ b/daemons/controld/controld_control.c
@@ -727,10 +727,9 @@ crmd_metadata(void)
const char *desc_short = "Pacemaker controller options";
const char *desc_long = "Cluster options used by Pacemaker's controller";
- gchar *s = pcmk__format_option_metadata(name, desc_short, desc_long,
- pcmk__opt_context_controld,
- controller_options,
- PCMK__NELEM(controller_options));
+ gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_controld);
+
printf("%s", s);
g_free(s);
}
diff --git a/include/crm/common/options_internal.h b/include/crm/common/options_internal.h
index 89d27d1..a62015f 100644
--- a/include/crm/common/options_internal.h
+++ b/include/crm/common/options_internal.h
@@ -74,6 +74,10 @@ gchar *pcmk__format_option_metadata(const char *name, const char *desc_short,
pcmk__cluster_option_t *option_list,
int len);
+gchar *pcmk__cluster_option_metadata(const char *name, const char *desc_short,
+ const char *desc_long,
+ enum pcmk__opt_context filter);
+
void pcmk__validate_cluster_options(GHashTable *options,
pcmk__cluster_option_t *option_list,
int len);
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index 2205d15..479a7fb 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -899,9 +899,9 @@ cib_metadata(void)
const char *desc_long = "Cluster options used by Pacemaker's Cluster "
"Information Base manager";
- gchar *s = pcmk__format_option_metadata(name, desc_short, desc_long,
- pcmk__opt_context_based,
- cib_opts, PCMK__NELEM(cib_opts));
+ gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_based);
+
printf("%s", s);
g_free(s);
}
diff --git a/lib/common/options.c b/lib/common/options.c
index d5b6c17..df4a8b4 100644
--- a/lib/common/options.c
+++ b/lib/common/options.c
@@ -43,7 +43,6 @@ pcmk__cli_help(char cmd)
* Option metadata
*/
-#if 0
static pcmk__cluster_option_t cluster_options[] = {
/* name, old name, type, allowed values,
* default value, validator,
@@ -534,7 +533,6 @@ static pcmk__cluster_option_t cluster_options[] = {
NULL,
},
};
-#endif // 0
/*
@@ -1036,6 +1034,32 @@ pcmk__format_option_metadata(const char *name, const char *desc_short,
return g_string_free(s, FALSE);
}
+/*!
+ * \internal
+ * \brief Format cluster option metadata as an OCF-like XML string
+ *
+ * \param[in] name Daemon name
+ * \param[in] desc_short Short description of the daemon
+ * \param[in] desc_long Long description of the daemon
+ * \param[in] filter If not \c pcmk__opt_context_none, include only
+ * those options whose \c context field is equal to
+ * \p filter
+ *
+ * \return A string containing OCF-like cluster option metadata XML
+ *
+ * \note The caller is responsible for freeing the return value using
+ * \c g_free().
+ */
+gchar *
+pcmk__cluster_option_metadata(const char *name, const char *desc_short,
+ const char *desc_long,
+ enum pcmk__opt_context filter)
+{
+ return pcmk__format_option_metadata(name, desc_short, desc_long, filter,
+ cluster_options,
+ PCMK__NELEM(cluster_options));
+}
+
void
pcmk__validate_cluster_options(GHashTable *options,
pcmk__cluster_option_t *option_list, int len)
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index e9aa2e2..c9f1fc1 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -323,9 +323,9 @@ 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(name, desc_short, desc_long,
- pcmk__opt_context_schedulerd,
- pe_opts, PCMK__NELEM(pe_opts));
+ gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_schedulerd);
+
out->output_xml(out, "metadata", s);
g_free(s);
}
--
2.31.1
From 9a7d33003dffea465e7b452abd0388db4a7d73b0 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Thu, 11 Jan 2024 00:06:32 -0800
Subject: [PATCH 11/24] Test: cts-cli: Update daemon outputs for reordering
In the new libcrmcommon options array, some options have been reordered
to be near other similar options.
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
cts/cli/regression.daemons.exp | 40 +++++++++++++++++-----------------
1 file changed, 20 insertions(+), 20 deletions(-)
diff --git a/cts/cli/regression.daemons.exp b/cts/cli/regression.daemons.exp
index 9b2dd96..43393df 100644
--- a/cts/cli/regression.daemons.exp
+++ b/cts/cli/regression.daemons.exp
@@ -51,16 +51,6 @@
<shortdesc lang="en">Polling interval to recheck cluster state and evaluate rules with date specifications</shortdesc>
<content type="time" default=""/>
</parameter>
- <parameter name="load-threshold">
- <longdesc lang="en">The cluster will slow down its recovery process when the amount of system resources used (currently CPU) approaches this limit</longdesc>
- <shortdesc lang="en">Maximum amount of system load that should be used by cluster nodes</shortdesc>
- <content type="percentage" default=""/>
- </parameter>
- <parameter name="node-action-limit">
- <longdesc lang="en">Maximum number of jobs that can be scheduled per node (defaults to 2x cores)</longdesc>
- <shortdesc lang="en">Maximum number of jobs that can be scheduled per node (defaults to 2x cores)</shortdesc>
- <content type="integer" default=""/>
- </parameter>
<parameter name="fence-reaction">
<longdesc lang="en">A cluster node may receive notification of its own fencing if fencing is misconfigured, or if fabric fencing is in use that doesn&apos;t cut cluster communication. Allowed values are &quot;stop&quot; to attempt to immediately stop Pacemaker and stay stopped, or &quot;panic&quot; to attempt to immediately reboot the local node, falling back to stop on failure.</longdesc>
<shortdesc lang="en">How a cluster node should react if notified of its own fencing</shortdesc>
@@ -101,6 +91,16 @@
<shortdesc lang="en">How many times fencing can fail before it will no longer be immediately re-attempted on a target</shortdesc>
<content type="integer" default=""/>
</parameter>
+ <parameter name="load-threshold">
+ <longdesc lang="en">The cluster will slow down its recovery process when the amount of system resources used (currently CPU) approaches this limit</longdesc>
+ <shortdesc lang="en">Maximum amount of system load that should be used by cluster nodes</shortdesc>
+ <content type="percentage" default=""/>
+ </parameter>
+ <parameter name="node-action-limit">
+ <longdesc lang="en">Maximum number of jobs that can be scheduled per node (defaults to 2x cores)</longdesc>
+ <shortdesc lang="en">Maximum number of jobs that can be scheduled per node (defaults to 2x cores)</shortdesc>
+ <content type="integer" default=""/>
+ </parameter>
</parameters>
</resource-agent>
=#=#=#= End test: Get controller metadata - OK (0) =#=#=#=
@@ -259,6 +259,16 @@
<option value="suicide" />
</content>
</parameter>
+ <parameter name="shutdown-lock">
+ <longdesc lang="en">When true, resources active on a node when it is cleanly shut down are kept &quot;locked&quot; 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.</longdesc>
+ <shortdesc lang="en">Whether to lock resources to a cleanly shut down node</shortdesc>
+ <content type="boolean" default=""/>
+ </parameter>
+ <parameter name="shutdown-lock-limit">
+ <longdesc lang="en">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.</longdesc>
+ <shortdesc lang="en">Do not lock resources to a cleanly shut down node longer than this</shortdesc>
+ <content type="time" default=""/>
+ </parameter>
<parameter name="symmetric-cluster">
<longdesc lang="en">Whether resources can run on any node by default</longdesc>
<shortdesc lang="en">Whether resources can run on any node by default</shortdesc>
@@ -279,16 +289,6 @@
<shortdesc lang="en">Whether the cluster should check for active resources during start-up</shortdesc>
<content type="boolean" default=""/>
</parameter>
- <parameter name="shutdown-lock">
- <longdesc lang="en">When true, resources active on a node when it is cleanly shut down are kept &quot;locked&quot; 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.</longdesc>
- <shortdesc lang="en">Whether to lock resources to a cleanly shut down node</shortdesc>
- <content type="boolean" default=""/>
- </parameter>
- <parameter name="shutdown-lock-limit">
- <longdesc lang="en">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.</longdesc>
- <shortdesc lang="en">Do not lock resources to a cleanly shut down node longer than this</shortdesc>
- <content type="time" default=""/>
- </parameter>
<parameter name="stonith-enabled">
<longdesc lang="en">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 &quot;split-brain&quot; situation, potentially leading to data loss and/or service unavailability.</longdesc>
<shortdesc lang="en">*** Advanced Use Only *** Whether nodes may be fenced as part of recovery</shortdesc>
--
2.31.1
From c085ff844deddefe4f00355e2a273f27eb35ce00 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Thu, 11 Jan 2024 00:23:01 -0800
Subject: [PATCH 12/24] Refactor: libcrmcommon, daemons: Use cluster_options
array in getters
Drop the option_list and len arguments from pcmk__cluster_option() and
pcmk__validate_cluster_options(). Use cluster_options in libcrmcommon
instead.
Now, all daemons fetch and validate the full set of cluster options,
even the ones they don't use. This is only slightly less efficient. It
ensures that there's no problem using the same option with multiple
daemons, and it makes it easy to use new options in a given daemon in
the future.
Now that nothing is using the local, per-daemon options arrays anymore,
we can drop them in an upcoming commit.
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/controld/controld_control.c | 9 +++---
include/crm/common/options_internal.h | 8 ++----
lib/cib/cib_utils.c | 7 +++--
lib/common/options.c | 40 ++++++++++++---------------
lib/pengine/common.c | 6 ++--
5 files changed, 31 insertions(+), 39 deletions(-)
diff --git a/daemons/controld/controld_control.c b/daemons/controld/controld_control.c
index 82aa143..4208947 100644
--- a/daemons/controld/controld_control.c
+++ b/daemons/controld/controld_control.c
@@ -515,6 +515,7 @@ do_recover(long long action,
register_fsa_input(C_FSA_INTERNAL, I_TERMINATE, NULL);
}
+#if 0
static pcmk__cluster_option_t controller_options[] = {
/* name, old name, type, allowed values,
* default value, validator,
@@ -719,6 +720,7 @@ static pcmk__cluster_option_t controller_options[] = {
"2 hours.")
},
};
+#endif // 0
void
crmd_metadata(void)
@@ -775,8 +777,7 @@ config_query_callback(xmlNode * msg, int call_id, int rc, xmlNode * output, void
config_hash, CIB_OPTIONS_FIRST, FALSE, now, NULL);
// Validate all options, and use defaults if not already present in hash
- pcmk__validate_cluster_options(config_hash, controller_options,
- PCMK__NELEM(controller_options));
+ pcmk__validate_cluster_options(config_hash);
/* Validate the watchdog timeout in the context of the local node
* environment. If invalid, the controller will exit with a fatal error.
@@ -900,9 +901,7 @@ crm_shutdown(int nsig)
* config_query_callback() has been run at least once, it doesn't look like
* anything could have changed the timer period since then.
*/
- value = pcmk__cluster_option(NULL, controller_options,
- PCMK__NELEM(controller_options),
- XML_CONFIG_ATTR_FORCE_QUIT);
+ value = pcmk__cluster_option(NULL, XML_CONFIG_ATTR_FORCE_QUIT);
default_period_ms = crm_parse_interval_spec(value);
controld_shutdown_start_countdown(default_period_ms);
}
diff --git a/include/crm/common/options_internal.h b/include/crm/common/options_internal.h
index a62015f..b727a58 100644
--- a/include/crm/common/options_internal.h
+++ b/include/crm/common/options_internal.h
@@ -64,9 +64,7 @@ typedef struct pcmk__cluster_option_s {
} pcmk__cluster_option_t;
-const char *pcmk__cluster_option(GHashTable *options,
- const pcmk__cluster_option_t *option_list,
- int len, const char *name);
+const char *pcmk__cluster_option(GHashTable *options, const char *name);
gchar *pcmk__format_option_metadata(const char *name, const char *desc_short,
const char *desc_long,
@@ -78,9 +76,7 @@ gchar *pcmk__cluster_option_metadata(const char *name, const char *desc_short,
const char *desc_long,
enum pcmk__opt_context filter);
-void pcmk__validate_cluster_options(GHashTable *options,
- pcmk__cluster_option_t *option_list,
- int len);
+void pcmk__validate_cluster_options(GHashTable *options);
bool pcmk__valid_interval_spec(const char *value);
bool pcmk__valid_boolean(const char *value);
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index 479a7fb..97f62ac 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -866,6 +866,7 @@ cib_native_notify(gpointer data, gpointer user_data)
crm_trace("Callback invoked...");
}
+#if 0
static pcmk__cluster_option_t cib_opts[] = {
/* name, legacy name, type, allowed values,
* default value, validator,
@@ -890,6 +891,7 @@ static pcmk__cluster_option_t cib_opts[] = {
" multiplied by the number of nodes).")
},
};
+#endif // 0
void
cib_metadata(void)
@@ -909,14 +911,13 @@ cib_metadata(void)
static void
verify_cib_options(GHashTable *options)
{
- pcmk__validate_cluster_options(options, cib_opts, PCMK__NELEM(cib_opts));
+ pcmk__validate_cluster_options(options);
}
const char *
cib_pref(GHashTable * options, const char *name)
{
- return pcmk__cluster_option(options, cib_opts, PCMK__NELEM(cib_opts),
- name);
+ return pcmk__cluster_option(options, name);
}
gboolean
diff --git a/lib/common/options.c b/lib/common/options.c
index df4a8b4..13d58e3 100644
--- a/lib/common/options.c
+++ b/lib/common/options.c
@@ -852,27 +852,21 @@ cluster_option_value(GHashTable *options, bool (*validate)(const char *),
* \internal
* \brief Get the value of a cluster option
*
- * \param[in,out] options Name/value pairs for configured options
- * \param[in] option_list Possible cluster options
- * \param[in] len Length of \p option_list
- * \param[in] name (Primary) option name to look for
+ * \param[in,out] options Name/value pairs for configured options
+ * \param[in] name (Primary) option name to look for
*
* \return Option value
*/
const char *
-pcmk__cluster_option(GHashTable *options,
- const pcmk__cluster_option_t *option_list,
- int len, const char *name)
+pcmk__cluster_option(GHashTable *options, const char *name)
{
- const char *value = NULL;
-
- for (int lpc = 0; lpc < len; lpc++) {
- if (pcmk__str_eq(name, option_list[lpc].name, pcmk__str_casei)) {
- value = cluster_option_value(options, option_list[lpc].is_valid,
- option_list[lpc].name,
- option_list[lpc].alt_name,
- option_list[lpc].default_value);
- return value;
+ for (int lpc = 0; lpc < PCMK__NELEM(cluster_options); lpc++) {
+ if (pcmk__str_eq(name, cluster_options[lpc].name, pcmk__str_casei)) {
+ return cluster_option_value(options,
+ cluster_options[lpc].is_valid,
+ cluster_options[lpc].name,
+ cluster_options[lpc].alt_name,
+ cluster_options[lpc].default_value);
}
}
CRM_CHECK(FALSE, crm_err("Bug: looking for unknown option '%s'", name));
@@ -1061,13 +1055,13 @@ pcmk__cluster_option_metadata(const char *name, const char *desc_short,
}
void
-pcmk__validate_cluster_options(GHashTable *options,
- pcmk__cluster_option_t *option_list, int len)
+pcmk__validate_cluster_options(GHashTable *options)
{
- for (int lpc = 0; lpc < len; lpc++) {
- cluster_option_value(options, option_list[lpc].is_valid,
- option_list[lpc].name,
- option_list[lpc].alt_name,
- option_list[lpc].default_value);
+ for (int lpc = 0; lpc < PCMK__NELEM(cluster_options); lpc++) {
+ cluster_option_value(options,
+ cluster_options[lpc].is_valid,
+ cluster_options[lpc].name,
+ cluster_options[lpc].alt_name,
+ cluster_options[lpc].default_value);
}
}
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index c9f1fc1..f99bd1b 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -21,6 +21,7 @@
gboolean was_processing_error = FALSE;
gboolean was_processing_warning = FALSE;
+#if 0
static pcmk__cluster_option_t pe_opts[] = {
/* name, old name, type, allowed values,
* default value, validator,
@@ -315,6 +316,7 @@ static pcmk__cluster_option_t pe_opts[] = {
NULL
},
};
+#endif // 0
void
pe_metadata(pcmk__output_t *out)
@@ -333,13 +335,13 @@ pe_metadata(pcmk__output_t *out)
void
verify_pe_options(GHashTable * options)
{
- pcmk__validate_cluster_options(options, pe_opts, PCMK__NELEM(pe_opts));
+ pcmk__validate_cluster_options(options);
}
const char *
pe_pref(GHashTable * options, const char *name)
{
- return pcmk__cluster_option(options, pe_opts, PCMK__NELEM(pe_opts), name);
+ return pcmk__cluster_option(options, name);
}
const char *
--
2.31.1
From de834cee2c5d8f4f796633e66f263ad77b9cd2eb Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 03:06:27 -0800
Subject: [PATCH 13/24] Refactor: various: Drop per-daemon cluster opt tables
Ref T746
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/controld/controld_control.c | 207 -------------------
lib/cib/cib_utils.c | 27 ---
lib/pengine/common.c | 297 ----------------------------
3 files changed, 531 deletions(-)
diff --git a/daemons/controld/controld_control.c b/daemons/controld/controld_control.c
index 4208947..40b90f8 100644
--- a/daemons/controld/controld_control.c
+++ b/daemons/controld/controld_control.c
@@ -515,213 +515,6 @@ do_recover(long long action,
register_fsa_input(C_FSA_INTERNAL, I_TERMINATE, NULL);
}
-#if 0
-static pcmk__cluster_option_t controller_options[] = {
- /* name, old name, type, allowed values,
- * default value, validator,
- * context,
- * short description,
- * long description
- */
- {
- "dc-version", NULL, "string", NULL, PCMK__VALUE_NONE, NULL,
- pcmk__opt_context_controld,
- N_("Pacemaker version on cluster node elected Designated Controller (DC)"),
- N_("Includes a hash which identifies the exact changeset the code was "
- "built from. Used for diagnostic purposes.")
- },
- {
- "cluster-infrastructure", NULL, "string", NULL, "corosync", NULL,
- pcmk__opt_context_controld,
- N_("The messaging stack on which Pacemaker is currently running"),
- N_("Used for informational and diagnostic purposes.")
- },
- {
- "cluster-name", NULL, "string", NULL, NULL, NULL,
- pcmk__opt_context_controld,
- N_("An arbitrary name for the cluster"),
- N_("This optional value is mostly for users' convenience as desired "
- "in administration, but may also be used in Pacemaker "
- "configuration rules via the #cluster-name node attribute, and "
- "by higher-level tools and resource agents.")
- },
- {
- XML_CONFIG_ATTR_DC_DEADTIME, NULL, "time",
- NULL, "20s", pcmk__valid_interval_spec,
- pcmk__opt_context_controld,
- N_("How long to wait for a response from other nodes during start-up"),
- N_("The optimal value will depend on the speed and load of your network "
- "and the type of switches used.")
- },
- {
- XML_CONFIG_ATTR_RECHECK, NULL, "time",
- N_("Zero disables polling, while positive values are an interval in seconds"
- "(unless other units are specified, for example \"5min\")"),
- "15min", pcmk__valid_interval_spec,
- pcmk__opt_context_controld,
- N_("Polling interval to recheck cluster state and evaluate rules "
- "with date specifications"),
- N_("Pacemaker is primarily event-driven, and looks ahead to know when to "
- "recheck cluster state for failure timeouts and most time-based "
- "rules. However, it will also recheck the cluster after this "
- "amount of inactivity, to evaluate rules with date specifications "
- "and serve as a fail-safe for certain types of scheduler bugs.")
- },
- {
- "load-threshold", NULL, "percentage", NULL,
- "80%", pcmk__valid_percentage,
- pcmk__opt_context_controld,
- N_("Maximum amount of system load that should be used by cluster nodes"),
- N_("The cluster will slow down its recovery process when the amount of "
- "system resources used (currently CPU) approaches this limit"),
- },
- {
- "node-action-limit", NULL, "integer", NULL,
- "0", pcmk__valid_number,
- pcmk__opt_context_controld,
- N_("Maximum number of jobs that can be scheduled per node "
- "(defaults to 2x cores)")
- },
- { XML_CONFIG_ATTR_FENCE_REACTION, NULL, "string", NULL, "stop", NULL,
- pcmk__opt_context_controld,
- N_("How a cluster node should react if notified of its own fencing"),
- N_("A cluster node may receive notification of its own fencing if fencing "
- "is misconfigured, or if fabric fencing is in use that doesn't cut "
- "cluster communication. Allowed values are \"stop\" to attempt to "
- "immediately stop Pacemaker and stay stopped, or \"panic\" to attempt "
- "to immediately reboot the local node, falling back to stop on failure.")
- },
- {
- XML_CONFIG_ATTR_ELECTION_FAIL, NULL, "time", NULL,
- "2min", pcmk__valid_interval_spec,
- pcmk__opt_context_controld,
- "*** Advanced Use Only ***",
- N_("Declare an election failed if it is not decided within this much "
- "time. If you need to adjust this value, it probably indicates "
- "the presence of a bug.")
- },
- {
- XML_CONFIG_ATTR_FORCE_QUIT, NULL, "time", NULL,
- "20min", pcmk__valid_interval_spec,
- pcmk__opt_context_controld,
- "*** Advanced Use Only ***",
- N_("Exit immediately if shutdown does not complete within this much "
- "time. If you need to adjust this value, it probably indicates "
- "the presence of a bug.")
- },
- {
- "join-integration-timeout", "crmd-integration-timeout", "time", NULL,
- "3min", pcmk__valid_interval_spec,
- pcmk__opt_context_controld,
- "*** Advanced Use Only ***",
- N_("If you need to adjust this value, it probably indicates "
- "the presence of a bug.")
- },
- {
- "join-finalization-timeout", "crmd-finalization-timeout", "time", NULL,
- "30min", pcmk__valid_interval_spec,
- pcmk__opt_context_controld,
- "*** Advanced Use Only ***",
- N_("If you need to adjust this value, it probably indicates "
- "the presence of a bug.")
- },
- {
- "transition-delay", "crmd-transition-delay", "time", NULL,
- "0s", pcmk__valid_interval_spec,
- pcmk__opt_context_controld,
- N_("*** Advanced Use Only *** Enabling this option will slow down "
- "cluster recovery under all conditions"),
- N_("Delay cluster recovery for this much time to allow for additional "
- "events to occur. Useful if your configuration is sensitive to "
- "the order in which ping updates arrive.")
- },
- {
- /* @COMPAT Currently unparsable values default to -1 (auto-calculate),
- * while missing values default to 0 (disable). All values are accepted
- * (unless the controller finds that the value conflicts with the
- * SBD_WATCHDOG_TIMEOUT).
- *
- * At a compatibility break: properly validate as a timeout, let
- * either negative values or a particular string like "auto" mean auto-
- * calculate, and use 0 as the single default for when the option either
- * is unset or fails to validate.
- */
- "stonith-watchdog-timeout", NULL, "time", NULL,
- "0", NULL,
- pcmk__opt_context_controld,
- N_("How long before nodes can be assumed to be safely down when "
- "watchdog-based self-fencing via SBD is in use"),
- N_("If this is set to a positive value, lost nodes are assumed to "
- "self-fence using watchdog-based SBD within this much time. This "
- "does not require a fencing resource to be explicitly configured, "
- "though a fence_watchdog resource can be configured, to limit use "
- "to specific nodes. If this is set to 0 (the default), the cluster "
- "will never assume watchdog-based self-fencing. If this is set to a "
- "negative value, the cluster will use twice the local value of the "
- "`SBD_WATCHDOG_TIMEOUT` environment variable if that is positive, "
- "or otherwise treat this as 0. WARNING: When used, this timeout "
- "must be larger than `SBD_WATCHDOG_TIMEOUT` on all nodes that use "
- "watchdog-based SBD, and Pacemaker will refuse to start on any of "
- "those nodes where this is not true for the local value or SBD is "
- "not active. When this is set to a negative value, "
- "`SBD_WATCHDOG_TIMEOUT` must be set to the same value on all nodes "
- "that use SBD, otherwise data corruption or loss could occur.")
- },
- {
- "stonith-max-attempts", NULL, "integer", NULL,
- "10", pcmk__valid_positive_number,
- pcmk__opt_context_controld,
- N_("How many times fencing can fail before it will no longer be "
- "immediately re-attempted on a target")
- },
-
- // Already documented in libpe_status (other values must be kept identical)
- {
- "no-quorum-policy", NULL, "select",
- "stop, freeze, ignore, demote, suicide", "stop", pcmk__valid_quorum,
- pcmk__opt_context_controld,
- N_("What to do when the cluster does not have quorum"), NULL
- },
- {
- XML_CONFIG_ATTR_SHUTDOWN_LOCK, NULL, "boolean", NULL,
- "false", pcmk__valid_boolean,
- pcmk__opt_context_controld,
- 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,
- pcmk__opt_context_controld,
- 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.")
- },
- {
- XML_CONFIG_ATTR_NODE_PENDING_TIMEOUT, NULL, "time", NULL,
- "0", pcmk__valid_interval_spec,
- pcmk__opt_context_controld,
- N_("How long to wait for a node that has joined the cluster to join "
- "the controller process group"),
- N_("Fence nodes that do not join the controller process group within "
- "this much time after joining the cluster, to allow the cluster "
- "to continue managing resources. A value of 0 means never fence "
- "pending nodes. Setting the value to 2h means fence nodes after "
- "2 hours.")
- },
-};
-#endif // 0
-
void
crmd_metadata(void)
{
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index 97f62ac..b83158c 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -866,33 +866,6 @@ cib_native_notify(gpointer data, gpointer user_data)
crm_trace("Callback invoked...");
}
-#if 0
-static pcmk__cluster_option_t cib_opts[] = {
- /* name, legacy name, type, allowed values,
- * default value, validator,
- * context,
- * short description,
- * long description
- */
- {
- "enable-acl", NULL, "boolean", NULL,
- "false", pcmk__valid_boolean,
- pcmk__opt_context_based,
- N_("Enable Access Control Lists (ACLs) for the CIB"),
- NULL
- },
- {
- "cluster-ipc-limit", NULL, "integer", NULL,
- "500", pcmk__valid_positive_number,
- pcmk__opt_context_based,
- N_("Maximum IPC message backlog before disconnecting a cluster daemon"),
- N_("Raise this if log has \"Evicting client\" messages for cluster daemon"
- " PIDs (a good value is the number of resources in the cluster"
- " multiplied by the number of nodes).")
- },
-};
-#endif // 0
-
void
cib_metadata(void)
{
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index f99bd1b..e96f0b5 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -21,303 +21,6 @@
gboolean was_processing_error = FALSE;
gboolean was_processing_warning = FALSE;
-#if 0
-static pcmk__cluster_option_t pe_opts[] = {
- /* name, old name, type, allowed values,
- * default value, validator,
- * context,
- * short description,
- * long description
- */
- {
- "no-quorum-policy", NULL, "select", "stop, freeze, ignore, demote, suicide",
- "stop", pcmk__valid_quorum,
- pcmk__opt_context_schedulerd,
- N_("What to do when the cluster does not have quorum"),
- NULL
- },
- {
- "symmetric-cluster", NULL, "boolean", NULL,
- "true", pcmk__valid_boolean,
- pcmk__opt_context_schedulerd,
- N_("Whether resources can run on any node by default"),
- NULL
- },
- {
- "maintenance-mode", NULL, "boolean", NULL,
- "false", pcmk__valid_boolean,
- pcmk__opt_context_schedulerd,
- N_("Whether the cluster should refrain from monitoring, starting, "
- "and stopping resources"),
- NULL
- },
- {
- "start-failure-is-fatal", NULL, "boolean", NULL,
- "true", pcmk__valid_boolean,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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",
- PCMK_ACTION_REBOOT, pcmk__is_fencing_action,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- N_("Allow performing fencing operations in parallel"),
- NULL
- },
- {
- "startup-fencing", NULL, "boolean", NULL,
- "true", pcmk__valid_boolean,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- "0", pcmk__valid_interval_spec,
- pcmk__opt_context_schedulerd,
- N_("How long to wait for a node that has joined the cluster to join "
- "the controller process group"),
- N_("Fence nodes that do not join the controller process group within "
- "this much time after joining the cluster, to allow the cluster "
- "to continue managing resources. A value of 0 means never fence "
- "pending nodes. Setting the value to 2h means fence nodes after "
- "2 hours.")
- },
- {
- "cluster-delay", NULL, "time", NULL,
- "60s", pcmk__valid_interval_spec,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- N_("Whether the cluster should stop all active resources"),
- NULL
- },
- {
- "stop-orphan-resources", NULL, "boolean", NULL,
- "true", pcmk__valid_boolean,
- pcmk__opt_context_schedulerd,
- N_("Whether to stop resources that were removed from the configuration"),
- NULL
- },
- {
- "stop-orphan-actions", NULL, "boolean", NULL,
- "true", pcmk__valid_boolean,
- pcmk__opt_context_schedulerd,
- N_("Whether to cancel recurring actions removed from the configuration"),
- NULL
- },
- {
- "remove-after-stop", NULL, "boolean", NULL,
- "false", pcmk__valid_boolean,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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,
- pcmk__opt_context_schedulerd,
- 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", pcmk__valid_placement_strategy,
- pcmk__opt_context_schedulerd,
- N_("How the cluster should allocate resources to nodes"),
- NULL
- },
-};
-#endif // 0
-
void
pe_metadata(pcmk__output_t *out)
{
--
2.31.1
From 9a8bb049fcb49204932e96014c3a63e58fd95d23 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 20:36:34 -0800
Subject: [PATCH 14/24] Refactor: libpe_status: Drop verify_pe_opts()
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
include/crm/pengine/internal.h | 1 -
lib/pengine/common.c | 6 ------
lib/pengine/unpack.c | 2 +-
3 files changed, 1 insertion(+), 8 deletions(-)
diff --git a/include/crm/pengine/internal.h b/include/crm/pengine/internal.h
index 9c8068f..5835ef8 100644
--- a/include/crm/pengine/internal.h
+++ b/include/crm/pengine/internal.h
@@ -209,7 +209,6 @@ pcmk_node_t *native_location(const pcmk_resource_t *rsc, GList **list,
int current);
void pe_metadata(pcmk__output_t *out);
-void verify_pe_options(GHashTable * options);
void native_add_running(pcmk_resource_t *rsc, pcmk_node_t *node,
pcmk_scheduler_t *scheduler, gboolean failed);
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index e96f0b5..402fae9 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -35,12 +35,6 @@ pe_metadata(pcmk__output_t *out)
g_free(s);
}
-void
-verify_pe_options(GHashTable * options)
-{
- pcmk__validate_cluster_options(options);
-}
-
const char *
pe_pref(GHashTable * options, const char *name)
{
diff --git a/lib/pengine/unpack.c b/lib/pengine/unpack.c
index 3429d56..2a9b563 100644
--- a/lib/pengine/unpack.c
+++ b/lib/pengine/unpack.c
@@ -228,7 +228,7 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
pe__unpack_dataset_nvpairs(config, XML_CIB_TAG_PROPSET, &rule_data, config_hash,
CIB_OPTIONS_FIRST, FALSE, scheduler);
- verify_pe_options(scheduler->config_hash);
+ pcmk__validate_cluster_options(config_hash);
set_config_flag(scheduler, "enable-startup-probes",
pcmk_sched_probe_resources);
--
2.31.1
From af79c50b7a5626218bf2a9b34fe631f07b1e2bda Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 08:40:13 -0800
Subject: [PATCH 15/24] Refactor: libpe_status: Drop pe_pref() internally
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/schedulerd/schedulerd_messages.c | 3 ++-
include/crm/pengine/internal.h | 10 +++++++---
lib/pacemaker/pcmk_graph_producer.c | 9 +++++----
lib/pacemaker/pcmk_sched_nodes.c | 5 +++--
lib/pengine/unpack.c | 4 +++-
tools/crm_resource_print.c | 4 ++--
6 files changed, 22 insertions(+), 13 deletions(-)
diff --git a/daemons/schedulerd/schedulerd_messages.c b/daemons/schedulerd/schedulerd_messages.c
index 5a97365..ff31fce 100644
--- a/daemons/schedulerd/schedulerd_messages.c
+++ b/daemons/schedulerd/schedulerd_messages.c
@@ -112,7 +112,8 @@ handle_pecalc_request(pcmk__request_t *request)
series_id = 2;
}
- value = pe_pref(scheduler->config_hash, series[series_id].param);
+ value = pcmk__cluster_option(scheduler->config_hash,
+ series[series_id].param);
if ((value == NULL)
|| (pcmk__scan_min_int(value, &series_wrap, -1) != pcmk_rc_ok)) {
series_wrap = series[series_id].wrap;
diff --git a/include/crm/pengine/internal.h b/include/crm/pengine/internal.h
index 5835ef8..2b7f2eb 100644
--- a/include/crm/pengine/internal.h
+++ b/include/crm/pengine/internal.h
@@ -620,14 +620,18 @@ int pe__node_health(pcmk_node_t *node);
static inline enum pcmk__health_strategy
pe__health_strategy(pcmk_scheduler_t *scheduler)
{
- return pcmk__parse_health_strategy(pe_pref(scheduler->config_hash,
- PCMK__OPT_NODE_HEALTH_STRATEGY));
+ const char *strategy = pcmk__cluster_option(scheduler->config_hash,
+ PCMK__OPT_NODE_HEALTH_STRATEGY);
+
+ return pcmk__parse_health_strategy(strategy);
}
static inline int
pe__health_score(const char *option, pcmk_scheduler_t *scheduler)
{
- return char2score(pe_pref(scheduler->config_hash, option));
+ const char *value = pcmk__cluster_option(scheduler->config_hash, option);
+
+ return char2score(value);
}
/*!
diff --git a/lib/pacemaker/pcmk_graph_producer.c b/lib/pacemaker/pcmk_graph_producer.c
index 59b6176..3006775 100644
--- a/lib/pacemaker/pcmk_graph_producer.c
+++ b/lib/pacemaker/pcmk_graph_producer.c
@@ -1004,16 +1004,17 @@ pcmk__create_graph(pcmk_scheduler_t *scheduler)
GList *iter = NULL;
const char *value = NULL;
long long limit = 0LL;
+ GHashTable *config_hash = scheduler->config_hash;
transition_id++;
crm_trace("Creating transition graph %d", transition_id);
scheduler->graph = create_xml_node(NULL, XML_TAG_GRAPH);
- value = pe_pref(scheduler->config_hash, "cluster-delay");
+ value = pcmk__cluster_option(config_hash, "cluster-delay");
crm_xml_add(scheduler->graph, "cluster-delay", value);
- value = pe_pref(scheduler->config_hash, "stonith-timeout");
+ value = pcmk__cluster_option(config_hash, "stonith-timeout");
crm_xml_add(scheduler->graph, "stonith-timeout", value);
crm_xml_add(scheduler->graph, "failed-stop-offset", "INFINITY");
@@ -1024,12 +1025,12 @@ pcmk__create_graph(pcmk_scheduler_t *scheduler)
crm_xml_add(scheduler->graph, "failed-start-offset", "1");
}
- value = pe_pref(scheduler->config_hash, "batch-limit");
+ value = pcmk__cluster_option(config_hash, "batch-limit");
crm_xml_add(scheduler->graph, "batch-limit", value);
crm_xml_add_int(scheduler->graph, "transition_id", transition_id);
- value = pe_pref(scheduler->config_hash, "migration-limit");
+ value = pcmk__cluster_option(config_hash, "migration-limit");
if ((pcmk__scan_ll(value, &limit, 0LL) == pcmk_rc_ok) && (limit > 0)) {
crm_xml_add(scheduler->graph, "migration-limit", value);
}
diff --git a/lib/pacemaker/pcmk_sched_nodes.c b/lib/pacemaker/pcmk_sched_nodes.c
index 9cf5545..03baa2c 100644
--- a/lib/pacemaker/pcmk_sched_nodes.c
+++ b/lib/pacemaker/pcmk_sched_nodes.c
@@ -360,8 +360,9 @@ pcmk__apply_node_health(pcmk_scheduler_t *scheduler)
{
int base_health = 0;
enum pcmk__health_strategy strategy;
- const char *strategy_str = pe_pref(scheduler->config_hash,
- PCMK__OPT_NODE_HEALTH_STRATEGY);
+ const char *strategy_str =
+ pcmk__cluster_option(scheduler->config_hash,
+ PCMK__OPT_NODE_HEALTH_STRATEGY);
strategy = pcmk__parse_health_strategy(strategy_str);
if (strategy == pcmk__health_strategy_none) {
diff --git a/lib/pengine/unpack.c b/lib/pengine/unpack.c
index 2a9b563..49443c6 100644
--- a/lib/pengine/unpack.c
+++ b/lib/pengine/unpack.c
@@ -50,7 +50,9 @@ struct action_history {
* flag is stringified more readably in log messages.
*/
#define set_config_flag(scheduler, option, flag) do { \
- const char *scf_value = pe_pref((scheduler)->config_hash, (option)); \
+ GHashTable *config_hash = (scheduler)->config_hash; \
+ const char *scf_value = pcmk__cluster_option(config_hash, (option)); \
+ \
if (scf_value != NULL) { \
if (crm_is_true(scf_value)) { \
(scheduler)->flags = pcmk__set_flags_as(__func__, __LINE__, \
diff --git a/tools/crm_resource_print.c b/tools/crm_resource_print.c
index bdf3ad9..26761dd 100644
--- a/tools/crm_resource_print.c
+++ b/tools/crm_resource_print.c
@@ -479,8 +479,8 @@ resource_check_list_default(pcmk__output_t *out, va_list args) {
"'%s' cannot run on unhealthy nodes due to "
PCMK__OPT_NODE_HEALTH_STRATEGY "='%s'",
parent->id,
- pe_pref(checks->rsc->cluster->config_hash,
- PCMK__OPT_NODE_HEALTH_STRATEGY));
+ pcmk__cluster_option(checks->rsc->cluster->config_hash,
+ PCMK__OPT_NODE_HEALTH_STRATEGY));
}
out->end_list(out);
--
2.31.1
From 1e78e617965b1a2e1a5671aa15943ba42487b09a Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 08:43:55 -0800
Subject: [PATCH 16/24] API: libpe_status: Deprecate pe_pref()
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
include/crm/pengine/common.h | 4 +---
include/crm/pengine/common_compat.h | 5 ++++-
lib/pengine/common.c | 20 ++++++++++++++------
lib/pengine/unpack.c | 28 ++++++++++++++--------------
4 files changed, 33 insertions(+), 24 deletions(-)
diff --git a/include/crm/pengine/common.h b/include/crm/pengine/common.h
index 2feac8a..a935aa7 100644
--- a/include/crm/pengine/common.h
+++ b/include/crm/pengine/common.h
@@ -1,5 +1,5 @@
/*
- * Copyright 2004-2023 the Pacemaker project contributors
+ * Copyright 2004-2024 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
@@ -28,8 +28,6 @@ enum rsc_role_e text2role(const char *role);
const char *role2text(enum rsc_role_e role);
const char *fail2text(enum action_fail_response fail);
-const char *pe_pref(GHashTable * options, const char *name);
-
/*!
* \brief Get readable description of a recovery type
*
diff --git a/include/crm/pengine/common_compat.h b/include/crm/pengine/common_compat.h
index 4330ccf..52e11f7 100644
--- a/include/crm/pengine/common_compat.h
+++ b/include/crm/pengine/common_compat.h
@@ -1,5 +1,5 @@
/*
- * Copyright 2004-2023 the Pacemaker project contributors
+ * Copyright 2004-2024 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
@@ -54,6 +54,9 @@ extern "C" {
//! \deprecated Do not use
#define RSC_ROLE_MASTER_S RSC_ROLE_PROMOTED_LEGACY_S
+
+//! \deprecated Do not use
+const char *pe_pref(GHashTable * options, const char *name);
#ifdef __cplusplus
}
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index 402fae9..0a4dfe6 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -35,12 +35,6 @@ pe_metadata(pcmk__output_t *out)
g_free(s);
}
-const char *
-pe_pref(GHashTable * options, const char *name)
-{
- return pcmk__cluster_option(options, name);
-}
-
const char *
fail2text(enum action_fail_response fail)
{
@@ -350,3 +344,17 @@ pe_node_attribute_raw(const pcmk_node_t *node, const char *name)
}
return g_hash_table_lookup(node->details->attrs, name);
}
+
+// Deprecated functions kept only for backward API compatibility
+// LCOV_EXCL_START
+
+#include <crm/pengine/common_compat.h>
+
+const char *
+pe_pref(GHashTable * options, const char *name)
+{
+ return pcmk__cluster_option(options, name);
+}
+
+// LCOV_EXCL_STOP
+// End deprecated API
diff --git a/lib/pengine/unpack.c b/lib/pengine/unpack.c
index 49443c6..d484e93 100644
--- a/lib/pengine/unpack.c
+++ b/lib/pengine/unpack.c
@@ -238,7 +238,7 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
crm_info("Startup probes: disabled (dangerous)");
}
- value = pe_pref(scheduler->config_hash, XML_ATTR_HAVE_WATCHDOG);
+ value = pcmk__cluster_option(config_hash, XML_ATTR_HAVE_WATCHDOG);
if (value && crm_is_true(value)) {
crm_info("Watchdog-based self-fencing will be performed via SBD if "
"fencing is required and stonith-watchdog-timeout is nonzero");
@@ -251,7 +251,7 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
set_if_xpath(pcmk_sched_enable_unfencing, XPATH_ENABLE_UNFENCING,
scheduler);
- value = pe_pref(scheduler->config_hash, "stonith-timeout");
+ value = pcmk__cluster_option(config_hash, "stonith-timeout");
scheduler->stonith_timeout = (int) crm_parse_interval_spec(value);
crm_debug("STONITH timeout: %d", scheduler->stonith_timeout);
@@ -262,8 +262,8 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
crm_debug("STONITH of failed nodes is disabled");
}
- scheduler->stonith_action = pe_pref(scheduler->config_hash,
- "stonith-action");
+ scheduler->stonith_action = pcmk__cluster_option(config_hash,
+ "stonith-action");
if (!strcmp(scheduler->stonith_action, "poweroff")) {
pe_warn_once(pcmk__wo_poweroff,
"Support for stonith-action of 'poweroff' is deprecated "
@@ -280,8 +280,8 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
crm_debug("Concurrent fencing is disabled");
}
- value = pe_pref(scheduler->config_hash,
- XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY);
+ value = pcmk__cluster_option(config_hash,
+ XML_CONFIG_ATTR_PRIORITY_FENCING_DELAY);
if (value) {
scheduler->priority_fencing_delay = crm_parse_interval_spec(value)
/ 1000;
@@ -299,7 +299,7 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
crm_debug("Cluster is symmetric" " - resources can run anywhere by default");
}
- value = pe_pref(scheduler->config_hash, "no-quorum-policy");
+ value = pcmk__cluster_option(config_hash, "no-quorum-policy");
if (pcmk__str_eq(value, "ignore", pcmk__str_casei)) {
scheduler->no_quorum_policy = pcmk_no_quorum_ignore;
@@ -367,7 +367,7 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
crm_trace("Orphan resource actions are ignored");
}
- value = pe_pref(scheduler->config_hash, "remove-after-stop");
+ value = pcmk__cluster_option(config_hash, "remove-after-stop");
if (value != NULL) {
if (crm_is_true(value)) {
pe__set_working_set_flags(scheduler, pcmk_sched_remove_after_stop);
@@ -407,14 +407,14 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
pe__unpack_node_health_scores(scheduler);
- scheduler->placement_strategy = pe_pref(scheduler->config_hash,
- "placement-strategy");
+ scheduler->placement_strategy =
+ pcmk__cluster_option(config_hash, "placement-strategy");
crm_trace("Placement strategy: %s", scheduler->placement_strategy);
set_config_flag(scheduler, "shutdown-lock", pcmk_sched_shutdown_lock);
if (pcmk_is_set(scheduler->flags, pcmk_sched_shutdown_lock)) {
- value = pe_pref(scheduler->config_hash,
- XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT);
+ value = pcmk__cluster_option(config_hash,
+ XML_CONFIG_ATTR_SHUTDOWN_LOCK_LIMIT);
scheduler->shutdown_lock = crm_parse_interval_spec(value) / 1000;
crm_trace("Resources will be locked to nodes that were cleanly "
"shut down (locks expire after %s)",
@@ -424,8 +424,8 @@ unpack_config(xmlNode *config, pcmk_scheduler_t *scheduler)
"shut down");
}
- value = pe_pref(scheduler->config_hash,
- XML_CONFIG_ATTR_NODE_PENDING_TIMEOUT);
+ value = pcmk__cluster_option(config_hash,
+ XML_CONFIG_ATTR_NODE_PENDING_TIMEOUT);
scheduler->node_pending_timeout = crm_parse_interval_spec(value) / 1000;
if (scheduler->node_pending_timeout == 0) {
crm_trace("Do not fence pending nodes");
--
2.31.1
From 866877401075e7ea4c3bc278e69ed94ea3a7af99 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 20:32:36 -0800
Subject: [PATCH 17/24] Refactor: libpe_status, scheduler: Drop pe_metadata()
Replace with static scheduler_metadata() in the scheduler since we don't
rely on a static options array in lib/pengine/common.c anymore.
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/schedulerd/pacemaker-schedulerd.c | 18 ++++++++++++++++--
include/crm/pengine/internal.h | 2 --
lib/pengine/common.c | 14 --------------
3 files changed, 16 insertions(+), 18 deletions(-)
diff --git a/daemons/schedulerd/pacemaker-schedulerd.c b/daemons/schedulerd/pacemaker-schedulerd.c
index 3f2a3e8..27c96da 100644
--- a/daemons/schedulerd/pacemaker-schedulerd.c
+++ b/daemons/schedulerd/pacemaker-schedulerd.c
@@ -1,5 +1,5 @@
/*
- * Copyright 2004-2023 the Pacemaker project contributors
+ * Copyright 2004-2024 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
@@ -46,6 +46,20 @@ pcmk__supported_format_t formats[] = {
void pengine_shutdown(int nsig);
+static void
+scheduler_metadata(pcmk__output_t *out)
+{
+ const char *name = "pacemaker-schedulerd";
+ const char *desc_short = "Pacemaker scheduler options";
+ const char *desc_long = "Cluster options used by Pacemaker's scheduler";
+
+ gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_schedulerd);
+
+ out->output_xml(out, "metadata", s);
+ g_free(s);
+}
+
static GOptionContext *
build_arg_context(pcmk__common_args_t *args, GOptionGroup **group) {
GOptionContext *context = NULL;
@@ -98,7 +112,7 @@ main(int argc, char **argv)
if (options.remainder) {
if (g_strv_length(options.remainder) == 1 &&
pcmk__str_eq("metadata", options.remainder[0], pcmk__str_casei)) {
- pe_metadata(out);
+ scheduler_metadata(out);
goto done;
} else {
exit_code = CRM_EX_USAGE;
diff --git a/include/crm/pengine/internal.h b/include/crm/pengine/internal.h
index 2b7f2eb..5965c1a 100644
--- a/include/crm/pengine/internal.h
+++ b/include/crm/pengine/internal.h
@@ -208,8 +208,6 @@ char *native_parameter(pcmk_resource_t *rsc, pcmk_node_t *node, gboolean create,
pcmk_node_t *native_location(const pcmk_resource_t *rsc, GList **list,
int current);
-void pe_metadata(pcmk__output_t *out);
-
void native_add_running(pcmk_resource_t *rsc, pcmk_node_t *node,
pcmk_scheduler_t *scheduler, gboolean failed);
diff --git a/lib/pengine/common.c b/lib/pengine/common.c
index 0a4dfe6..6551d10 100644
--- a/lib/pengine/common.c
+++ b/lib/pengine/common.c
@@ -21,20 +21,6 @@
gboolean was_processing_error = FALSE;
gboolean was_processing_warning = FALSE;
-void
-pe_metadata(pcmk__output_t *out)
-{
- const char *name = "pacemaker-schedulerd";
- const char *desc_short = "Pacemaker scheduler options";
- const char *desc_long = "Cluster options used by Pacemaker's scheduler";
-
- gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
- pcmk__opt_context_schedulerd);
-
- out->output_xml(out, "metadata", s);
- g_free(s);
-}
-
const char *
fail2text(enum action_fail_response fail)
{
--
2.31.1
From 700c906d621887f257c73ddfd7c82c773cb32c8e Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 20:40:18 -0800
Subject: [PATCH 18/24] Refactor: libcib: Drop verify_cib_options()
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
lib/cib/cib_utils.c | 8 +-------
1 file changed, 1 insertion(+), 7 deletions(-)
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index b83158c..227a50f 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -881,12 +881,6 @@ cib_metadata(void)
g_free(s);
}
-static void
-verify_cib_options(GHashTable *options)
-{
- pcmk__validate_cluster_options(options);
-}
-
const char *
cib_pref(GHashTable * options, const char *name)
{
@@ -913,7 +907,7 @@ cib_read_config(GHashTable * options, xmlNode * current_cib)
options, CIB_OPTIONS_FIRST, TRUE, now, NULL);
}
- verify_cib_options(options);
+ pcmk__validate_cluster_options(options);
crm_time_free(now);
--
2.31.1
From 6284a3a79b88fd20630bfbfe866a4c2c3686a246 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 20:24:19 -0800
Subject: [PATCH 19/24] Refactor: libcib: Drop cib_pref() internally
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
lib/cib/cib_utils.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index 227a50f..5b241ae 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -281,7 +281,7 @@ cib_acl_enabled(xmlNode *xml, const char *user)
GHashTable *options = pcmk__strkey_table(free, free);
cib_read_config(options, xml);
- value = cib_pref(options, "enable-acl");
+ value = pcmk__cluster_option(options, "enable-acl");
rc = crm_is_true(value);
g_hash_table_destroy(options);
}
--
2.31.1
From 1806822590b0060079b94b7d2867722ef2430bf9 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 20:26:29 -0800
Subject: [PATCH 20/24] API: libcib: Deprecate cib_pref()
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
include/crm/cib/util.h | 1 -
include/crm/cib/util_compat.h | 3 +++
lib/cib/cib_utils.c | 12 ++++++------
3 files changed, 9 insertions(+), 7 deletions(-)
diff --git a/include/crm/cib/util.h b/include/crm/cib/util.h
index 18726bb..47894cb 100644
--- a/include/crm/cib/util.h
+++ b/include/crm/cib/util.h
@@ -57,7 +57,6 @@ int set_standby(cib_t * the_cib, const char *uuid, const char *scope, const char
xmlNode *cib_get_generation(cib_t * cib);
void cib_metadata(void);
-const char *cib_pref(GHashTable * options, const char *name);
int cib_apply_patch_event(xmlNode *event, xmlNode *input, xmlNode **output,
int level);
diff --git a/include/crm/cib/util_compat.h b/include/crm/cib/util_compat.h
index 20f1e2d..d6ccd4d 100644
--- a/include/crm/cib/util_compat.h
+++ b/include/crm/cib/util_compat.h
@@ -33,6 +33,9 @@ const char *get_object_parent(const char *object_type);
//! \deprecated Use pcmk_cib_xpath_for() instead
xmlNode *get_object_root(const char *object_type, xmlNode *the_root);
+//! \deprecated Do not use
+const char *cib_pref(GHashTable * options, const char *name);
+
#ifdef __cplusplus
}
#endif
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index 5b241ae..f9c463e 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -881,12 +881,6 @@ cib_metadata(void)
g_free(s);
}
-const char *
-cib_pref(GHashTable * options, const char *name)
-{
- return pcmk__cluster_option(options, name);
-}
-
gboolean
cib_read_config(GHashTable * options, xmlNode * current_cib)
{
@@ -1085,5 +1079,11 @@ get_object_root(const char *object_type, xmlNode *the_root)
return pcmk_find_cib_element(the_root, object_type);
}
+const char *
+cib_pref(GHashTable * options, const char *name)
+{
+ return pcmk__cluster_option(options, name);
+}
+
// LCOV_EXCL_STOP
// End deprecated API
--
2.31.1
From 422fb81250aa733d2601b4d412c3fbcbf5b74420 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 20:29:29 -0800
Subject: [PATCH 21/24] Refactor: based, libcib: Drop cib_metadata() internally
Replace with a static based_metadata() in pacemaker-based since we don't
depend on a static options array in lib/cib/cib_utils.c anymore.
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/based/pacemaker-based.c | 17 ++++++++++++++++-
1 file changed, 16 insertions(+), 1 deletion(-)
diff --git a/daemons/based/pacemaker-based.c b/daemons/based/pacemaker-based.c
index 5dd7938..78bcd51 100644
--- a/daemons/based/pacemaker-based.c
+++ b/daemons/based/pacemaker-based.c
@@ -126,6 +126,21 @@ setup_stand_alone(GError **error)
return pcmk_rc_ok;
}
+static void
+based_metadata(void)
+{
+ const char *name = "pacemaker-based";
+ const char *desc_short = "Cluster Information Base manager options";
+ const char *desc_long = "Cluster options used by Pacemaker's Cluster "
+ "Information Base manager";
+
+ gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_based);
+
+ printf("%s", s);
+ g_free(s);
+}
+
static GOptionEntry entries[] = {
{ "stand-alone", 's', G_OPTION_FLAG_NONE, G_OPTION_ARG_NONE, &stand_alone,
"(Advanced use only) Run in stand-alone mode", NULL },
@@ -204,7 +219,7 @@ main(int argc, char **argv)
if ((g_strv_length(processed_args) >= 2)
&& pcmk__str_eq(processed_args[1], "metadata", pcmk__str_none)) {
- cib_metadata();
+ based_metadata();
goto done;
}
--
2.31.1
From 05b3e08de7c515c38cf42bbbeaf18e3346eb360d Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 20:30:16 -0800
Subject: [PATCH 22/24] API: libcib: Deprecate cib_metadata()
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
include/crm/cib/util.h | 2 --
include/crm/cib/util_compat.h | 3 +++
lib/cib/cib_utils.c | 30 +++++++++++++++---------------
3 files changed, 18 insertions(+), 17 deletions(-)
diff --git a/include/crm/cib/util.h b/include/crm/cib/util.h
index 47894cb..ce6cbeb 100644
--- a/include/crm/cib/util.h
+++ b/include/crm/cib/util.h
@@ -56,8 +56,6 @@ int set_standby(cib_t * the_cib, const char *uuid, const char *scope, const char
xmlNode *cib_get_generation(cib_t * cib);
-void cib_metadata(void);
-
int cib_apply_patch_event(xmlNode *event, xmlNode *input, xmlNode **output,
int level);
diff --git a/include/crm/cib/util_compat.h b/include/crm/cib/util_compat.h
index d6ccd4d..95e0766 100644
--- a/include/crm/cib/util_compat.h
+++ b/include/crm/cib/util_compat.h
@@ -36,6 +36,9 @@ xmlNode *get_object_root(const char *object_type, xmlNode *the_root);
//! \deprecated Do not use
const char *cib_pref(GHashTable * options, const char *name);
+//! \deprecated Do not use
+void cib_metadata(void);
+
#ifdef __cplusplus
}
#endif
diff --git a/lib/cib/cib_utils.c b/lib/cib/cib_utils.c
index f9c463e..75dda16 100644
--- a/lib/cib/cib_utils.c
+++ b/lib/cib/cib_utils.c
@@ -866,21 +866,6 @@ cib_native_notify(gpointer data, gpointer user_data)
crm_trace("Callback invoked...");
}
-void
-cib_metadata(void)
-{
- const char *name = "pacemaker-based";
- const char *desc_short = "Cluster Information Base manager options";
- const char *desc_long = "Cluster options used by Pacemaker's Cluster "
- "Information Base manager";
-
- gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
- pcmk__opt_context_based);
-
- printf("%s", s);
- g_free(s);
-}
-
gboolean
cib_read_config(GHashTable * options, xmlNode * current_cib)
{
@@ -1085,5 +1070,20 @@ cib_pref(GHashTable * options, const char *name)
return pcmk__cluster_option(options, name);
}
+void
+cib_metadata(void)
+{
+ const char *name = "pacemaker-based";
+ const char *desc_short = "Cluster Information Base manager options";
+ const char *desc_long = "Cluster options used by Pacemaker's Cluster "
+ "Information Base manager";
+
+ gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_based);
+
+ printf("%s", s);
+ g_free(s);
+}
+
// LCOV_EXCL_STOP
// End deprecated API
--
2.31.1
From f8ee575a51f6bacf82abb1d1f41eba1092776682 Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Mon, 8 Jan 2024 20:42:37 -0800
Subject: [PATCH 23/24] Refactor: controller: Replace crmd_metadata() with
controld_metadata()
Can be static since we don't rely on an options array that lives in
controld_control.c anymore.
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
daemons/controld/controld_control.c | 14 --------------
daemons/controld/pacemaker-controld.c | 18 ++++++++++++++++--
daemons/controld/pacemaker-controld.h | 3 +--
3 files changed, 17 insertions(+), 18 deletions(-)
diff --git a/daemons/controld/controld_control.c b/daemons/controld/controld_control.c
index 40b90f8..9b54900 100644
--- a/daemons/controld/controld_control.c
+++ b/daemons/controld/controld_control.c
@@ -515,20 +515,6 @@ do_recover(long long action,
register_fsa_input(C_FSA_INTERNAL, I_TERMINATE, NULL);
}
-void
-crmd_metadata(void)
-{
- const char *name = "pacemaker-controld";
- const char *desc_short = "Pacemaker controller options";
- const char *desc_long = "Cluster options used by Pacemaker's controller";
-
- gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
- pcmk__opt_context_controld);
-
- printf("%s", s);
- g_free(s);
-}
-
static void
config_query_callback(xmlNode * msg, int call_id, int rc, xmlNode * output, void *user_data)
{
diff --git a/daemons/controld/pacemaker-controld.c b/daemons/controld/pacemaker-controld.c
index e4a72c2..d80644d 100644
--- a/daemons/controld/pacemaker-controld.c
+++ b/daemons/controld/pacemaker-controld.c
@@ -1,5 +1,5 @@
/*
- * Copyright 2004-2023 the Pacemaker project contributors
+ * Copyright 2004-2024 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
@@ -46,6 +46,20 @@ static pcmk__supported_format_t formats[] = {
{ NULL, NULL, NULL }
};
+static void
+controld_metadata(void)
+{
+ const char *name = "pacemaker-controld";
+ const char *desc_short = "Pacemaker controller options";
+ const char *desc_long = "Cluster options used by Pacemaker's controller";
+
+ gchar *s = pcmk__cluster_option_metadata(name, desc_short, desc_long,
+ pcmk__opt_context_controld);
+
+ printf("%s", s);
+ g_free(s);
+}
+
static GOptionContext *
build_arg_context(pcmk__common_args_t *args, GOptionGroup **group)
{
@@ -96,7 +110,7 @@ main(int argc, char **argv)
if ((g_strv_length(processed_args) >= 2)
&& pcmk__str_eq(processed_args[1], "metadata", pcmk__str_none)) {
- crmd_metadata();
+ controld_metadata();
initialize = false;
goto done;
}
diff --git a/daemons/controld/pacemaker-controld.h b/daemons/controld/pacemaker-controld.h
index 2334cce..ba8dc8f 100644
--- a/daemons/controld/pacemaker-controld.h
+++ b/daemons/controld/pacemaker-controld.h
@@ -1,5 +1,5 @@
/*
- * Copyright 2004-2023 the Pacemaker project contributors
+ * Copyright 2004-2024 the Pacemaker project contributors
*
* The version control history for this file may have further details.
*
@@ -28,7 +28,6 @@
# define controld_trigger_config() \
controld_trigger_config_as(__func__, __LINE__)
-void crmd_metadata(void);
void controld_trigger_config_as(const char *fn, int line);
void controld_election_init(const char *uname);
void controld_configure_election(GHashTable *options);
--
2.31.1
From 282e9eb026699abef5a28fc37f54b9330029da1c Mon Sep 17 00:00:00 2001
From: Reid Wahl <nrwahl@protonmail.com>
Date: Tue, 2 Jan 2024 19:56:11 -0800
Subject: [PATCH 24/24] Test: cts-cli: Update daemon tests after fence-reaction
select
Signed-off-by: Reid Wahl <nrwahl@protonmail.com>
---
cts/cli/regression.daemons.exp | 7 +++++--
1 file changed, 5 insertions(+), 2 deletions(-)
diff --git a/cts/cli/regression.daemons.exp b/cts/cli/regression.daemons.exp
index 43393df..543d62f 100644
--- a/cts/cli/regression.daemons.exp
+++ b/cts/cli/regression.daemons.exp
@@ -52,9 +52,12 @@
<content type="time" default=""/>
</parameter>
<parameter name="fence-reaction">
- <longdesc lang="en">A cluster node may receive notification of its own fencing if fencing is misconfigured, or if fabric fencing is in use that doesn&apos;t cut cluster communication. Allowed values are &quot;stop&quot; to attempt to immediately stop Pacemaker and stay stopped, or &quot;panic&quot; to attempt to immediately reboot the local node, falling back to stop on failure.</longdesc>
+ <longdesc lang="en">A cluster node may receive notification of its own fencing if fencing is misconfigured, or if fabric fencing is in use that doesn&apos;t cut cluster communication. Use &quot;stop&quot; to attempt to immediately stop Pacemaker and stay stopped, or &quot;panic&quot; to attempt to immediately reboot the local node, falling back to stop on failure. Allowed values: stop, panic</longdesc>
<shortdesc lang="en">How a cluster node should react if notified of its own fencing</shortdesc>
- <content type="string" default=""/>
+ <content type="select" default="">
+ <option value="stop" />
+ <option value="panic" />
+ </content>
</parameter>
<parameter name="election-timeout">
<longdesc lang="en">Declare an election failed if it is not decided within this much time. If you need to adjust this value, it probably indicates the presence of a bug.</longdesc>
--
2.31.1
Reference in New Issue View Git Blame Copy Permalink