Deployment Architecture

Why are scheduled reports failing to run after enabling search head clustering in Splunk 6.2.1?

Splunk Employee
Splunk Employee

We just enabled SHC 6.2.1 and I'm finding none of my scheduled reports are running at all.

As a test, I created a simple search (index=_internal | stats count) along with a cron to execute it every minute, but the saved scheduled search does not run at all. I can also confirm this search provides results as an Adhoc search.
Investigating further, it seems nothing has ran since the day I enabled SHC last week. I have also confirmed, default-mode.conf does not have the 'pipeline:scheduler' thread disabled, so I'm assuming this is specific to SHC and my configuration.

1 Solution

Splunk Employee
Splunk Employee

found the problem!!

We had an app get punted over from our DEV environment that had

 server.conf: adhoc_searchhead = true

Once we killed the app and restarted, ALL scheduled searches started up nicely.

adhoc_searchhead = <bool>
    * This setting configures a member as an adhoc searchhead; i.e., the member will not
      run any scheduled jobs.
    * Use the setting captain_is_adhoc_searchhead to reduce compute load on the captain.
    * Defaults to false.

View solution in original post

Splunk Employee
Splunk Employee

found the problem!!

We had an app get punted over from our DEV environment that had

 server.conf: adhoc_searchhead = true

Once we killed the app and restarted, ALL scheduled searches started up nicely.

adhoc_searchhead = <bool>
    * This setting configures a member as an adhoc searchhead; i.e., the member will not
      run any scheduled jobs.
    * Use the setting captain_is_adhoc_searchhead to reduce compute load on the captain.
    * Defaults to false.

View solution in original post

Path Finder

Thanks for this. It pointed me in the right direction. While not the exact cause of our issue, it is similar.

We had a user who couldn't load a dashboard from scheduled search results on a search head cluster. It seemed to only happen on one node of the cluster. Turned out it was alway the captain node. We had the captain configured to only run ad-hoc searches:
https://docs.splunk.com/Documentation/Splunk/6.6.2/DistSearch/Adhocclustermember#Configure_the_capta...

This explains why there were no results from the scheduled search to load the dashboard on that one instance.

0 Karma