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 "balaji.gandhi" <ji...@gmail.com> on 2012/10/31 14:53:53 UTC
After to SOLR 4.0.0 Upgrade - ClusterState says we are the leader,
but locally we don't think so
Hi,
After upgrading from Solr 4.0.0-Beta to Solr 4.0.0 we are getting this error
from ALL the leader nodes:-
Oct 31, 2012 6:44:03 AM
org.apache.solr.update.processor.DistributedUpdateProcessor
doDefensiveChecks
SEVERE: ClusterState says we are the leader, but locally we don't think so
Is there a configuration/schema change we are missing? Please let us know.
Thanks,
Balaji
--
View this message in context: http://lucene.472066.n3.nabble.com/After-to-SOLR-4-0-0-Upgrade-ClusterState-says-we-are-the-leader-but-locally-we-don-t-think-so-tp4017277.html
Sent from the Solr - User mailing list archive at Nabble.com.
Re: After to SOLR 4.0.0 Upgrade - ClusterState says we are the
leader, but locally we don't think so
Posted by "balaji.gandhi" <ji...@gmail.com>.
Mark,
We have tried the following:-
1. Removing everything in the ZooKeeper snapshot directory
2. Removing the indexes
We get the same error in both the cases.
Attached the cloud dumps:- cloud_dump.json
<http://lucene.472066.n3.nabble.com/file/n4017315/cloud_dump.json>
Thanks,
Balaji
--
View this message in context: http://lucene.472066.n3.nabble.com/After-to-SOLR-4-0-0-Upgrade-ClusterState-says-we-are-the-leader-but-locally-we-don-t-think-so-tp4017277p4017315.html
Sent from the Solr - User mailing list archive at Nabble.com.
Re: After to SOLR 4.0.0 Upgrade - ClusterState says we are the
leader, but locally we don't think so
Posted by Mark Miller <ma...@gmail.com>.
Hmm...not sure - perhaps you could share you zookeeper data (there is
a dump to clipboard option in the admin ui, cloud tab). Anything else
interesting in the logs?
Some zookeeper layout changes were made, and I'm not sure an upgrade
from beta/alpha was ever really tested with them (I didn't make the
changes) - you might try dropping everything in zookeeper (not your
Solr indexes, just ZooKeeper data) and restarting the cluster. Just a
guess though.
The error itself is as it says - a defensive measure - each SolrCore
tracks whether it thinks it's the leader locally in RAM - the cluster
state also says who the leader is - if there is not agreement, the
core won't act as the leader because something must be off.
- Mark
On Wed, Oct 31, 2012 at 9:53 AM, balaji.gandhi <ji...@gmail.com> wrote:
> Hi,
>
> After upgrading from Solr 4.0.0-Beta to Solr 4.0.0 we are getting this error
> from ALL the leader nodes:-
>
> Oct 31, 2012 6:44:03 AM
> org.apache.solr.update.processor.DistributedUpdateProcessor
> doDefensiveChecks
> SEVERE: ClusterState says we are the leader, but locally we don't think so
>
> Is there a configuration/schema change we are missing? Please let us know.
>
> Thanks,
> Balaji
>
>
>
> --
> View this message in context: http://lucene.472066.n3.nabble.com/After-to-SOLR-4-0-0-Upgrade-ClusterState-says-we-are-the-leader-but-locally-we-don-t-think-so-tp4017277.html
> Sent from the Solr - User mailing list archive at Nabble.com.
--
- Mark