AutoSys Workload Automation

Hi Autosys Experts,

We are facing a new(?) problem in our Autosys 4.5 instance:

When the instance is restarted, the EP count is less than the expected:

*** WARNING: 8 Primary Event Processors were detected ***
*** 11 Primary Event Processors should be running ***
*** THIS SHOULD BE INVESTIGATED MANUALLY! ***

And after some time, the EP goes down automatically:

[01:57:07.2565] [1] EVENT: STOP_DEMON

[01:57:07.2645] [1] Warning: failed to signal EP with pid <18791>: No such process. May require manual intervention.
[01:57:07.2650] [1] Warning: failed to signal EP with pid <18798>: No such process. May require manual intervention.
[01:57:07.2656] [1] Warning: failed to signal EP with pid <18814>: No such process. May require manual intervention.
[01:57:07.2665] [1] Shutting down the shadow ping process.

[01:57:07.2751] [10] ______________________________________________________________________________
[01:57:07.2757] [10] Received SIGTERM - orderly shutdown
[01:57:07.2758] [10] ______________________________________________________________________________
[01:57:07.2837] [10] ______________________________________________________________________________
[01:57:07.2839] [10] The Event Processor #10 IS SHUTTING DOWN.

Not sure how it is receiving the STOP_DEMON.

Regards

Guru

Hi Guruprasad,

Was there a MULTI_EP_SHUTDOWN alarm as well?

Do you have RestartEPs=1 parameter set in the config.$AUTOSERV file?

Thanks,

Chandru

Hi Chandru,

Thank you for the reply.

There are no MULTIPLE_EP_SHUTDOWN alarms. Only EP_ROLLOVER and then EP_SHUTDOWN.

Also in the config:

RestartEPs=1

The EPs are shutting down quite frequently.

Is there a way Event Processor shuts down itself if there are any internal issues.. like number of EPs less etc?

One of the known cases is both the DBs becoming unavailable, but that does not seem to be the case.

Are there any tables/entries that have the EP pid process details that i can flush before restarting.. I assume the number of EPs started is less because the other slots/counts are still occupied by previous EP processes, although not running.

Regards

Guru

Chandru,

The alamode table seems to contain the list of EPs and their corresponding pids.

If the restart of EPs is not clean (as in above) is it OK to delete these entries from the alamode and then restart the instance.

Hi,

Manually clearing alamode table isn't advisable. I am trying to remember, there used to be a file under $AUTOUSER, some name like *as_semid* that can be removed. If you find it, then delete that file and then try to start EPs.

Thanks,

Chandru

Thank you Chandru!

Will be checking this.

Regards

Guru