[
https://issues.apache.org/jira/browse/SOLR-13082?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16725616#comment-16725616
]
Shalin Shekhar Mangar commented on SOLR-13082:
----------------------------------------------
Here's a patch which shuffles triggers on resume to somewhat mitigate this
problem. The scheduled thread pool executor runs scheduled tasks in FIFO order
so we shuffle the list of triggers before submitting them to the executor. I
also added a cautionary note in the ref guide for the scheduled trigger
describing this problem and recommending that scheduled triggers not be used
for very frequent operations.
> A trigger that creates trigger events more frequently than the cool down
> period can starve other triggers
> ---------------------------------------------------------------------------------------------------------
>
> Key: SOLR-13082
> URL: https://issues.apache.org/jira/browse/SOLR-13082
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Components: AutoScaling
> Reporter: Shalin Shekhar Mangar
> Priority: Major
> Fix For: master (8.0), 7.7
>
> Attachments: SOLR-13082.patch
>
>
> A trigger that creates frequent events (such as ScheduledTrigger) with period
> less than the cool down period can starve all other triggers. This is because
> we execute triggers in the same order each time so if the scheduled trigger
> happens to be the first in the list then after every cool down period, the
> scheduled trigger can create another event and the cycle repeats.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
