If the errors happen with garbage collection then potentially, yes.
You should never pause longer than your zk timeout (both sides).
On Thu, Apr 30, 2020 at 11:03 PM Ganesh Sethuraman
wrote:
>
> Any other JVM settings change possible?
>
> On Tue, Apr 28, 2020, 10:15 PM Sethuraman, Ganesh
> wrot
Any other JVM settings change possible?
On Tue, Apr 28, 2020, 10:15 PM Sethuraman, Ganesh
wrote:
> Hi
>
> We are using SolrCloud 7.2.1 with 3 node Zookeeper ensemble. We have 92
> collection each on avg. having 8 shards and 2 replica with 2 EC2 nodes,
> with JVM size of 18GB (G1 GC). We need you
meters it will certainly help
>
> Regards
> Ganesh
>
> -Original Message-
> From: matthew sporleder
> Sent: Wednesday, April 29, 2020 11:47 AM
> To: solr-user@lucene.apache.org
> Subject: Re: SolrCloud degraded during backup and batch CSV update
>
> CAUTION:
increase the timeout setting in Solr or Zookeeper, if you
can share parameters it will certainly help
Regards
Ganesh
-Original Message-
From: matthew sporleder
Sent: Wednesday, April 29, 2020 11:47 AM
To: solr-user@lucene.apache.org
Subject: Re: SolrCloud degraded during backup and batch CSV
FWIW I've had some luck with strategy 3 (increase zk timeout) when you
overwhelm the connection to zk or the disk on zk.
Is zk on the same boxes as solr?
On Tue, Apr 28, 2020 at 10:15 PM Sethuraman, Ganesh
wrote:
>
> Hi
>
> We are using SolrCloud 7.2.1 with 3 node Zookeeper ensemble. We have 92
Hi
We are using SolrCloud 7.2.1 with 3 node Zookeeper ensemble. We have 92
collection each on avg. having 8 shards and 2 replica with 2 EC2 nodes, with
JVM size of 18GB (G1 GC). We need your help with the Issue we faced today: The
issue is SolrCloud server went into a degraded collections (for
