You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by Paddy Krishnamoorthy <kp...@gmail.com> on 2015/05/15 04:07:33 UTC

Full Copy during cluster restarts

We are running a solr 4.10.2 cluster with dozens of nodes. When ever we
restart the cluster we see that 2-3 nodes goes in to a fullCopy mode, which
takes for ever to recover. Because of the recovery our entire cluster runs
very slow and eventually becomes unresponsive.

Any thoughts?


Thanks
Paddy
-- 
Did you breathe today?
Learn from here: http://us.artofliving.org/index.html

Re: Full Copy during cluster restarts

Posted by Erick Erickson <er...@gmail.com>.
The most likely cause is that you were indexing when the nodes were
down and/or have not issued a hard commit. If you stop indexing _and_
issue a hard commit (openSearcher=true or false doesn't matter), then
the nodes should come back up without having to replicate the index.

This might help:
http://lucidworks.com/blog/understanding-transaction-logs-softcommit-and-commit-in-sorlcloud/

Best,
Erick

On Thu, May 14, 2015 at 9:07 PM, Paddy Krishnamoorthy <kp...@gmail.com> wrote:
> We are running a solr 4.10.2 cluster with dozens of nodes. When ever we
> restart the cluster we see that 2-3 nodes goes in to a fullCopy mode, which
> takes for ever to recover. Because of the recovery our entire cluster runs
> very slow and eventually becomes unresponsive.
>
> Any thoughts?
>
>
> Thanks
> Paddy
> --
> Did you breathe today?
> Learn from here: http://us.artofliving.org/index.html