Skip to Content.
Sympa Menu

grouper-users - Re: [grouper-users] CHANGE_LOG_changeLogTempToChangeLog jobs stop running

Subject: Grouper Users - Open Discussion List

List archive

Re: [grouper-users] CHANGE_LOG_changeLogTempToChangeLog jobs stop running


Chronological Thread 
  • From: Scott Koranda <>
  • To: Shilen Patel <>
  • Cc: grouper-users <>
  • Subject: Re: [grouper-users] CHANGE_LOG_changeLogTempToChangeLog jobs stop running
  • Date: Mon, 29 Apr 2019 20:55:20 -0500

Hi,

Before receiving your note below, I found this wiki page

https://spaces.at.internet2.edu/display/Grouper/Grouper+-+Loader

I then stopped the loader and using GSH ran

GrouperSession.startRootSession();
GrouperLoader.schedulerFactory().getScheduler().unscheduleJob(org.quartz.TriggerKey.triggerKey("triggerChangeLog_grouperChangeLogTempToChangeLog"));

I restarted the loader and saw that the ChangeLogTempToChangeLog jobs
were running as expected.

I then stopped the loader again and looked at the grouper_QZ_TRIGGERS
table. I did not find anything else in the BLOCKED state, but I found
others in the ACQUIRED state, even though they had not run for quite
some time.

So I then ran using GSH

GrouperLoader.schedulerFactory().getScheduler().unscheduleJob(org.quartz.TriggerKey.triggerKey("triggerChangeLog_CHANGE_LOG_consumer_pspng_attributes"));

GrouperLoader.schedulerFactory().getScheduler().unscheduleJob(org.quartz.TriggerKey.triggerKey("triggerChangeLog_CHANGE_LOG_consumer_pspng_groupOfNames"));

GrouperLoader.schedulerFactory().getScheduler().unscheduleJob(org.quartz.TriggerKey.triggerKey("triggerChangeLog_CHANGE_LOG_consumer_grouperRules"));

GrouperLoader.schedulerFactory().getScheduler().unscheduleJob(org.quartz.TriggerKey.triggerKey("triggerChangeLog_CHANGE_LOG_consumer_syncGroups"));

That left only triggers in the WAITING state.

I then started the loader again, and now I see in grouper_loader_log the
sequence of jobs I expect for the change log and the PSP-NG.

More below inline...

> There’s a known issue with quartz where jobs may get stuck in a BLOCKED
> state. There’s a separate job whose purpose is to fix those (every 30
> minutes). Maybe that got stuck too? Do you have anything else that’s
> blocked?

I belive it is this:

https://github.com/quartz-scheduler/quartz/pull/146

It looks like that is scheduled to be included in a release relatively
soon.

Will it be possible to release a patch for 2.3.x that upgrades the quartz jar?

> To manually run the job that unblocks it:
>
> groovy:000> loaderRunOneJob("OTHER_JOB_schedulerCheckDaemon")

I get this:

groovy:000> GrouperSession.startRootSession();
===> 3d29e472266e4c53a008e10d2de31b63,'GrouperSystem','application'
groovy:000> loaderRunOneJob("OTHER_JOB_schedulerCheckDaemon")
ERROR java.lang.RuntimeException:
java.lang.RuntimeException: Problem loading class: null
at
edu.internet2.middleware.grouper.app.loader.GrouperLoader.runOnceByJobName
(GrouperLoader.java:1625)
at edu.internet2.middleware.grouper.app.gsh.loaderRunOneJob.invoke
(loaderRunOneJob.java:95)
at
edu.internet2.middleware.grouper.app.gsh.loaderRunOneJob$invoke.call (Unknown
Source)
at groovysh_evaluate.loaderRunOneJob (groovysh_evaluate:4)

I don't see any quartz trigger with that name:

MySQL [grouper]> select job_name from grouper_QZ_TRIGGERS;
+----------------------------------------+
| job_name |
+----------------------------------------+
| CHANGE_LOG_changeLogTempToChangeLog |
| CHANGE_LOG_consumer_grouperRules |
| CHANGE_LOG_consumer_pspng_attributes |
| CHANGE_LOG_consumer_pspng_groupOfNames |
| CHANGE_LOG_consumer_syncGroups |
| MAINTENANCE_cleanLogs |
| MAINTENANCE__builtinMessagingDaemon |
| MAINTENANCE__enabledDisabled |
| MAINTENANCE__rules |
| OTHER_JOB_attestationDaemon |
| OTHER_JOB_deprovisioningDaemon |
| OTHER_JOB_pspng_attributes_full |
| OTHER_JOB_pspng_groupOfNames_full |
+----------------------------------------+
13 rows in set (0.00 sec)

Any ideas?

Thanks,

Scott

> It just basically changes the state from BLOCKED to WAITING for anything
> that is not currently firing (i.e. actually blocked). Does that fix the
> problem?
>
> Thanks!
>
> - Shilen
>
> On 4/29/19, 8:53 PM, "Scott Koranda" <> wrote:
>
> Hi,
>
> > Are any of your other jobs (that run via the daemon) running
> correctly?
>
> The only other daemon jobs would be the PSP-NG jobs that follow the
> changelog. No, those jobs are not running correctly. They appear to have
> stopped 5 days ago based on the last entries in grouper_loader_log.
>
> > What do the quartz tables show?
> >
> > select * from grouper_qz_triggers where
> trigger_name='triggerChangeLog_grouperChangeLogTempToChangeLog'
> >
> > select * from grouper_qz_fired_triggers where
> trigger_name='triggerChangeLog_grouperChangeLogTempToChangeLog'
>
> MySQL [grouper]> select * from grouper_QZ_TRIGGERS where
> trigger_name='triggerChangeLog_grouperChangeLogTempToChangeLog'\G
> *************************** 1. row ***************************
> sched_name: DefaultQuartzScheduler
> trigger_name: triggerChangeLog_grouperChangeLogTempToChangeLog
> trigger_group: DEFAULT
> job_name: CHANGE_LOG_changeLogTempToChangeLog
> job_group: DEFAULT
> description: NULL
> next_fire_time: 1556569425000
> prev_fire_time: 1556569365000
> priority: 5
> trigger_state: BLOCKED
> trigger_type: CRON
> start_time: 1551790380000
> end_time: 0
> calendar_name: NULL
> misfire_instr: 0
> job_data:
> 1 row in set (0.00 sec)
>
> MySQL [grouper]> select * from grouper_QZ_FIRED_TRIGGERS where
> trigger_name='triggerChangeLog_grouperChangeLogTempToChangeLog'\G
> Empty set (0.00 sec)
>
> Thanks,
>
> Scott
>
>



Archive powered by MHonArc 2.6.19.

Top of Page