[
https://issues.apache.org/jira/browse/HDFS-5590?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13840988#comment-13840988
]
Suresh Srinivas commented on HDFS-5590:
---------------------------------------
bq. Do I understand correctly that removal of this parameter will be affecting
downstream tools like CM and Ambari, but we don't care?
This is as indicated in comments above, an undocumented configuration. It was
only used within hdfs code for optimization, which we now know causes data
loss.
This should not cause issues for downstream tools.
> Block ID and generation stamp may be reused when persistBlocks is set to false
> ------------------------------------------------------------------------------
>
> Key: HDFS-5590
> URL: https://issues.apache.org/jira/browse/HDFS-5590
> Project: Hadoop HDFS
> Issue Type: Bug
> Affects Versions: 2.2.0
> Reporter: Jing Zhao
> Assignee: Jing Zhao
> Fix For: 2.3.0
>
> Attachments: HDFS-5590.000.patch, HDFS-5590.001.patch
>
>
> In a cluster with non-HA setup and dfs.persist.blocks set to false, we may
> have data loss in the following case:
> # client creates file1 and requests a block from NN and get blk_id1_gs1
> # client writes blk_id1_gs1 to DN
> # NN is restarted and because persistBlocks is false, blk_id1_gs1 may not be
> persisted in disk
> # another client creates file2 and NN will allocate a new block using the
> same block id blk_id1_gs1 since block ID and generation stamp are both
> increased sequentially.
> Now we may have two versions (file1 and file2) of the blk_id1_gs1 (same id,
> same gs) in the system. It will case data loss.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
