[
https://issues.apache.org/jira/browse/SOLR-7280?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Erick Erickson updated SOLR-7280:
---------------------------------
Attachment: SOLR-7280-5x.patch
Work in progress patch for back-porting to 5x, just re-doing the lambdas.
WIP because I just noticed that the new test class _also_ has lambdas that have
not been changed in this patch. Even so, I could still compile the server and
run my test suite on it.
Results:
This is _vastly_ better than the old 5x code, I don't get OOM errors and the
like. Occasionally I'll get 1 or two replicas (out of 1,600) that stay in the
"down" state, but do come back up when I restart the Solr instance hosting them.
At Ishan's suggestion I applied my 5x patch to a clean 6x code base and don't
see any replicas staying down there (so far) so it looks like some other
changes between 5x and 6x are making the startup process more robust as well.
> Load cores in sorted order and tweak coreLoadThread counts to improve cluster
> stability on restarts
> ---------------------------------------------------------------------------------------------------
>
> Key: SOLR-7280
> URL: https://issues.apache.org/jira/browse/SOLR-7280
> Project: Solr
> Issue Type: Bug
> Components: SolrCloud
> Reporter: Shalin Shekhar Mangar
> Assignee: Noble Paul
> Fix For: 6.2
>
> Attachments: SOLR-7280-5x.patch, SOLR-7280.patch, SOLR-7280.patch
>
>
> In SOLR-7191, Damien mentioned that by loading solr cores in a sorted order
> and tweaking some of the coreLoadThread counts, he was able to improve the
> stability of a cluster with thousands of collections. We should explore some
> of these changes and fold them into Solr.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
