[
https://issues.apache.org/jira/browse/HBASE-11726?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14393939#comment-14393939
]
sridhararao mutluri commented on HBASE-11726:
---------------------------------------------
Hi Enis,
We are using hadoop 2.6 ha with hbase 1.0 in distributed mode.The master is
dyeing out speedily with below error:
My HMASTER is dyeing out speedily as hmaster log shows below error:
2015-04-02 03:43:43,588 FATAL [vxa1:16020.activeMasterManager] master.HMaster:
Failed to become active master
java.net.ConnectException: Call From vxa1.cloudc.com/10.1.178.86 to
vxa1.cloudc.com:9000 failed on connection exception: java.net.ConnectException:
Connection refused; For more details see:
http://wiki.apache.org/hadoop/ConnectionRefused
Please suggest
> Master should fail-safe if starting with a pre 0.96 layout
> ----------------------------------------------------------
>
> Key: HBASE-11726
> URL: https://issues.apache.org/jira/browse/HBASE-11726
> Project: HBase
> Issue Type: Bug
> Components: master
> Affects Versions: 0.96.2, 0.99.0, 0.98.5, 2.0.0
> Reporter: Esteban Gutierrez
> Assignee: Esteban Gutierrez
> Priority: Critical
> Fix For: 0.99.0, 2.0.0, 0.98.6
>
> Attachments: HBASE-11726.v0.diff, HBASE-11726.v1.diff,
> HBASE-11726.v2.diff
>
>
> We recently saw this: If user inadvertently starts the HBase Master after
> deploying new HBase binaries (any version that supports namespaces), the
> HMaster will start the migration to PBs the the {{hbase.version}} file per
> HBASE-5453 and that will write a new version file PB-serialized but with the
> old version number. Further restarts of the master will fail because the
> hbase version file has been migrated to PBs and there will be version
> mismatch. The right approach should be to fail safe the master if we find an
> old {{hbase.version}} file in order to force user to run upgrade tool.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
