[jira] [Commented] (ACCUMULO-2899) WAL handling fails to deal with 1.4 -> 1.5 -> 1.6

["ASF subversion and git services (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/ACCUMULO-2899?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14029725#comment-14029725 ] 

ASF subversion and git services commented on ACCUMULO-2899:
-----------------------------------------------------------

Commit 31aea2ad89b9b988995a9b8f9096091788eead74 in accumulo's branch refs/heads/1.6.1-SNAPSHOT from [~busbey]
[ https://git-wip-us.apache.org/repos/asf?p=accumulo.git;h=31aea2a ]

ACCUMULO-2899 Correctly handle older internals.

Make sure that we behave correctly when our cluster has been upgraded 1.4 -> 1.5 -> 1.6

* dfsloggers need to track the cq used to create them so they can be used to issue putDeletes
* rewrite 1.4 WAL paths to the top level wal directory
* adjust recovery path to compensate for 1.4 recovered wals
* fix GC to properly handle upgraded WALs

Author: Mike Drob <md...@cloudera.com>
Author: Sean Busbey <bu...@cloudera.com>


> WAL handling fails to deal with 1.4 -> 1.5 -> 1.6
> -------------------------------------------------
>
>                 Key: ACCUMULO-2899
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-2899
>             Project: Accumulo
>          Issue Type: Bug
>          Components: gc, logger
>    Affects Versions: 1.6.0
>            Reporter: Sean Busbey
>            Assignee: Sean Busbey
>            Priority: Blocker
>             Fix For: 1.6.1, 1.7.0
>
>
> After doing a 1.4 -> 1.5 -> 1.6 upgrade that still has WALs for some tables, the 1.6 instance fails to correctly handle the 1.4 recovered WALs.
> This can happen either through not waiting long enough after the upgrade to 1.5 or because of an offline table brought online on 1.6 (ala ACCUMULO-2816).



--
This message was sent by Atlassian JIRA
(v6.2#6252)