[
https://issues.apache.org/jira/browse/ACCUMULO-4000?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14790609#comment-14790609
]
Eric Newton commented on ACCUMULO-4000:
---------------------------------------
Yes, I think so. The file had a block allocated to it and it was 93 bytes in
length (how the NN knew this, I have not yet determined), and no copy of the
block on any other server. I could not copy the file locally. The first issue
was that the lease release failed.
I should mention that several racks were being decommissioned, so there was a
lot of block management at the time.
> log recovery failed after hard reset
> ------------------------------------
>
> Key: ACCUMULO-4000
> URL: https://issues.apache.org/jira/browse/ACCUMULO-4000
> Project: Accumulo
> Issue Type: Bug
> Affects Versions: 1.6.2
> Environment: very large cluster, accumulo 1.6.2, hadoop 2.5.0 (cdh
> 5.3)
> Reporter: Eric Newton
> Assignee: Eric Newton
>
> Had a hardware failure on a single node within a large cluster. Tablets were
> migrated away, but one tablet would not recover. The Closer run by the
> master to release the write lease on the WAL failed repeatedly.
> Afterwards, it was determined the file was small, probably just opened and
> used at the moment the machine failed. The block could not be recovered from
> any replicas.
> One question raised: does the write pipeline acknowledge the sync, before the
> write pipeline completes?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
