[SCM] CTDB repository - branch 1.0.112 updated - ctdb-1.0.111-47-g392f243

[Ronnie Sahlberg]

The branch, 1.0.112 has been updated
       via  392f243fdbd5085da5c3c57fde044df3dfe79d03 (commit)
       via  ebd4c3eac5b536f7a32b3be3f7a6fa699b7b4b7f (commit)
       via  0432b1cdd7d3792f3a00ff539651381efe7f4935 (commit)
      from  f8b32385257c9526999a229e84020677deb79eaf (commit)

http://gitweb.samba.org/?p=sahlberg/ctdb.git;a=shortlog;h=1.0.112


- Log -----------------------------------------------------------------
commit 392f243fdbd5085da5c3c57fde044df3dfe79d03
Author: Ronnie Sahlberg <ronniesahlberg at gmail.com>
Date:   Thu Feb 25 12:34:27 2010 +1100

    New version 1.0.112-11
    
    * Thu Feb 25 2010 : Version 1.0.112-11
     - Increase default script timeout to 90 seconds.
       BZ 61113
     - Add a new tunable EventScriptLogTimeout which will log a message everytime
       a monitor script takes longer than this to finish.
       BZ 61118

commit ebd4c3eac5b536f7a32b3be3f7a6fa699b7b4b7f
Author: Ronnie Sahlberg <ronniesahlberg at gmail.com>
Date:   Thu Feb 25 12:22:32 2010 +1100

    Add a tunable EventScriptLogTimeout that will log an error everytime a
    monitoring script took longer than this.
    
    BZ 61118

commit 0432b1cdd7d3792f3a00ff539651381efe7f4935
Author: Ronnie Sahlberg <ronniesahlberg at gmail.com>
Date:   Thu Feb 25 11:58:07 2010 +1100

    Increase the default script timeout to 90 seconds.
    
    BZ 61113

-----------------------------------------------------------------------

Summary of changes:
 include/ctdb_private.h     |    1 +
 packaging/RPM/ctdb.spec.in |    8 +++++++-
 server/ctdb_tunables.c     |    3 ++-
 server/eventscript.c       |    8 ++++++++
 4 files changed, 18 insertions(+), 2 deletions(-)


Changeset truncated at 500 lines:

diff --git a/include/ctdb_private.h b/include/ctdb_private.h
index 6107bdf..f5e6de8 100644
--- a/include/ctdb_private.h
+++ b/include/ctdb_private.h
@@ -105,6 +105,7 @@ struct ctdb_tunable {
 	uint32_t script_timeout;
 	uint32_t script_timeout_count; /* allow dodgy scripts to hang this many times in a row before we mark the node unhealthy */
 	uint32_t script_unhealthy_on_timeout; /* obsolete */
+	uint32_t script_log_timeout; /* Log an ERROR when a script took longer than this */
 	uint32_t recovery_grace_period;
 	uint32_t recovery_ban_period;
 	uint32_t database_hash_size;
diff --git a/packaging/RPM/ctdb.spec.in b/packaging/RPM/ctdb.spec.in
index 9a2f10b..658c522 100644
--- a/packaging/RPM/ctdb.spec.in
+++ b/packaging/RPM/ctdb.spec.in
@@ -5,7 +5,7 @@ Vendor: Samba Team
 Packager: Samba Team <samba at samba.org>
 Name: ctdb
 Version: 1.0.112
-Release: 10
+Release: 11
 Epoch: 0
 License: GNU GPL version 3
 Group: System Environment/Daemons
@@ -123,6 +123,12 @@ rm -rf $RPM_BUILD_ROOT
 %{_docdir}/ctdb/tests/bin/ctdb_transaction
 
 %changelog
+* Thu Feb 25 2010 : Version 1.0.112-11
+ - Increase default script timeout to 90 seconds. 
+   BZ 61113
+ - Add a new tunable EventScriptLogTimeout which will log a message everytime
+   a monitor script takes longer than this to finish.
+   BZ 61118
 * Tue Feb 23 2010 : Version 1.0.112-10
  - revert the change in 10.0.0.112-9 and make a new attempt to make the scripts behave.
  - make writing the ticklelist in 61.nfstickle a background task to avoid
diff --git a/server/ctdb_tunables.c b/server/ctdb_tunables.c
index cca270b..d1c7212 100644
--- a/server/ctdb_tunables.c
+++ b/server/ctdb_tunables.c
@@ -37,9 +37,10 @@ static const struct {
 	{ "TakeoverTimeout",      5,  offsetof(struct ctdb_tunable, takeover_timeout) },
 	{ "MonitorInterval",     15,  offsetof(struct ctdb_tunable, monitor_interval) },
 	{ "TickleUpdateInterval",20,  offsetof(struct ctdb_tunable, tickle_update_interval) },
-	{ "EventScriptTimeout",  30,  offsetof(struct ctdb_tunable, script_timeout) },
+	{ "EventScriptTimeout",  90,  offsetof(struct ctdb_tunable, script_timeout) },
 	{ "EventScriptTimeoutCount", 1,  offsetof(struct ctdb_tunable, script_timeout_count) },
 	{ "EventScriptUnhealthyOnTimeout", 0, offsetof(struct ctdb_tunable, script_unhealthy_on_timeout) },/* OBSOLETE */
+	{ "EventScriptLogTimeout",  3,  offsetof(struct ctdb_tunable, script_log_timeout) },
 	{ "RecoveryGracePeriod", 120,  offsetof(struct ctdb_tunable, recovery_grace_period) },
 	{ "RecoveryBanPeriod",  300,  offsetof(struct ctdb_tunable, recovery_ban_period) },
 	{ "DatabaseHashSize", 10000,  offsetof(struct ctdb_tunable, database_hash_size) },
diff --git a/server/eventscript.c b/server/eventscript.c
index e0908e1..8a6d779 100644
--- a/server/eventscript.c
+++ b/server/eventscript.c
@@ -472,6 +472,14 @@ static void ctdb_event_script_handler(struct event_context *ev, struct fd_event
 	}
 
 	current->finished = timeval_current();
+
+	/* log something if it took too long */
+	if (ctdb->tunable.script_log_timeout > 0) {
+		if ((state->call == CTDB_EVENT_MONITOR) && (timeval_delta(&current->finished, &current->start) > ctdb->tunable.script_log_timeout)) {
+			DEBUG(DEBUG_ERR,("Monitor script %s took %.3lfs to complete\n", current->name, timeval_delta(&current->finished, &current->start)));
+		}
+	}
+
 	/* valgrind gets overloaded if we run next script as it's still doing
 	 * post-execution analysis, so kill finished child here. */
 	if (ctdb->valgrinding) {


-- 
CTDB repository