You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Guanghao Zhang (JIRA)" <ji...@apache.org> on 2016/12/13 13:43:58 UTC

[jira] [Created] (HBASE-17303) Let master to check and transfer the dead rs's replication queues

Guanghao Zhang created HBASE-17303:
--------------------------------------

             Summary: Let master to check and transfer the dead rs's replication queues
                 Key: HBASE-17303
                 URL: https://issues.apache.org/jira/browse/HBASE-17303
             Project: HBase
          Issue Type: Bug
          Components: Replication
            Reporter: Guanghao Zhang


Dump replication queues result from our cluster.
{code}
Found 8 deleted queues, run hbck -fixReplication in order to remove the deleted replication queues
    hostname,24610,1481528189915/80-hostname,24620,1476784763605
    hostname,24620,1476784763605/70-hostname,24630,1470418208092-hostname,24600,1476773709589
    hostname,24630,1481528526258/17000-hostname,24620,1470044455538-hostname,24630,1470037674231-hostname,24600,1476773708489-hostname,24620,1476784763605
    hostname,24620,1481528358531/70-hostname,24600,1476773709589-hostname,24620,1476784763605
    hostname,24600,1481528021595/70-hostname,24630,1470421093464-hostname,24630,1476773708939-hostname,24610,1476779010928-hostname,24620,1476784747260
    hostname,24600,1481528021595/17000-hostname,24620,1476784763605
    hostname,24600,1481528021595/17000-hostname,24630,1475381530644-hostname,24600,1476773709589-hostname,24620,1476784763605
    hostname,24600,1481528021595/17000-hostname,24600,1476773709589-hostname,24620,1476784763605
Found 2 dead regionservers, restart one regionserver to transfer the queues of dead regionservers
    hostname,24600,1481547616148
    hostname,24620,1476784763605
{code}
Now for dead rs's replication znode, you need restart one regionserver to transfer the replication queues of dead regionservers. Same idea with HBASE-16336, we can let master to periodically check the dead rs znode, too. And send the transfer replication queues request to any regionserver. Then the dead rs's replication queues can be transfer automatically and don't need to wait a regionserver restart. Any suggestions are welcomed.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)