[
https://issues.apache.org/jira/browse/HDFS-9426?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15004384#comment-15004384
]
Vinayakumar B commented on HDFS-9426:
-------------------------------------
bq. The remaining datanodes should stop populating trash. Conceptually we could
make it clear trash on upgrade or right away. But the older version of
datanodes can only do the latter.
Yes, thats correct. But it should be made clear that, If some problem in
upgrading those remained nodes, then data will not be recovered, since upgrade
will be immediately finalized in first heartbeat itself.
IMO, proof of concept patch looks good.
> Rollingupgrade finalization is not backward compatible
> ------------------------------------------------------
>
> Key: HDFS-9426
> URL: https://issues.apache.org/jira/browse/HDFS-9426
> Project: Hadoop HDFS
> Issue Type: Bug
> Reporter: Kihwal Lee
> Priority: Blocker
> Attachments: HDFS-9426.branch-2.7.poc.patch, HDFS-9426.trunk.poc.patch
>
>
> After HDFS-7645, the namenode can return non-null {{rollingUpgradeInfo}} in
> heatbeat reponses. 2.7.1 or 2.6.x datanodes won't finalize the upgrade
> because it's not null.
> NN might have to check the DN version and return different
> {{rollingUpgradeInfo}}.
> HDFS-8656 recognized the compatibility issue of the changed semantics, but
> unfortunately did not address the semantics of the heartbeat response.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
