[
https://issues.apache.org/jira/browse/HBASE-13616?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stack updated HBASE-13616:
--------------------------
Attachment: 13616.wip.v5.branch-1.1.txt
Against branch-1.1 for now.
Add ServerCrashState data and states to MasterProcedure protos.
Replaced ServerShutdownHandler with ServerCrashProcedure.
Some renaming so we talk of server crash rather than SSH
In master, removed the 'just-in-case'
this.fileSystemManager.removeStaleRecoveringRegionsFromZK(previouslyFailedServers);
which could actually do premature remove of DLR region no-read znodes Not needed
now we have no longer have 'holes' in DLR if master crashes because of this
move of SSH to procedure-basis.
Did rename of lock methods in pv2. Rather than talk of read/write locks,
instead do exclusive/shared.
In MasterProcedureQueue, have server processing happen before table procedures.
Lots of fixup in TestAssignmentManager.
TODO: Fix queue and fix wait on assign.
> Move ServerShutdownHandler to Pv2
> ---------------------------------
>
> Key: HBASE-13616
> URL: https://issues.apache.org/jira/browse/HBASE-13616
> Project: HBase
> Issue Type: Sub-task
> Components: proc-v2
> Affects Versions: 1.1.0
> Reporter: stack
> Assignee: stack
> Attachments: 13616.wip.txt, 13616.wip.v3.branch-1.txt,
> 13616.wip.v4.branch-1.txt, 13616.wip.v5.branch-1.1.txt, 13616wip.v2.txt
>
>
> Move ServerShutdownHandler to run on ProcedureV2. Need this for DLR to work.
> See HBASE-13567.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
