lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Raintung Li (JIRA)" <>
Subject [jira] [Created] (SOLR-4073) Overseer will miss operations in some cases.
Date Wed, 14 Nov 2012 13:20:11 GMT
Raintung Li created SOLR-4073:

             Summary: Overseer will miss operations in some cases.
                 Key: SOLR-4073
             Project: Solr
          Issue Type: Bug
          Components: SolrCloud
    Affects Versions: 4.0, 4.0-BETA, 4.0-ALPHA
         Environment: Solr cloud
            Reporter: Raintung Li

One overseer disconnect with Zookeeper, but overseer thread still handle the request(A) in
the DistributedQueue. Example: overseer thread reconnect Zookeeper try to remove the Top's
request. "workQueue.remove();".   

Now the other server will take over the overseer privilege because old overseer disconnect.
Start overseer thread and handle the queue request(A) again, and remove the request(A) from
queue, then try to get the top's request(B, doesn't get). In the this time old overseer reconnect
with ZooKeeper, and remove the top's request from queue. Now the top request is B, it is moved
by old overseer server.  New overseer server never do B request,because this request deleted
by old overseer server, at the last this request(B) miss operations.

At best, distributeQueue.peek can get the request's ID that will be removed for workqueue.remove(ID),
not remove the top's request.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message