[
https://issues.apache.org/jira/browse/HDFS-13112?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16359271#comment-16359271
]
Xiao Chen commented on HDFS-13112:
----------------------------------
Thanks for the fix Daryn and Kihwal. I have not reviewed as careful as Kihwal
did, but from what I see, LGTM. :)
One question:
{code:title=FSNamesystem.java}
public void logUpdateMasterKey(DelegationKey key) {
...
assert hasReadLock();
getEditLog().logUpdateMasterKey(key);
getEditLog().logSync();
}
{code}
I think {{logSync}} is usually done outside of the FSN lock, why not do the
same here?
Also just to confirm my understanding: the comment in
{{logExpireDelegationToken}} says that expiration edits are batched, which is
reasonable. In code there is no {{logSync}} called at the end of the
{{removeExpiredToken}}, but we don't necessarily have to call it because worst
case is we lost it on failover, and new NN will still remove it in the next
interval.
> Token expiration edits may cause log corruption or deadlock
> -----------------------------------------------------------
>
> Key: HDFS-13112
> URL: https://issues.apache.org/jira/browse/HDFS-13112
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: namenode
> Affects Versions: 2.1.0-beta, 0.23.8
> Reporter: Daryn Sharp
> Assignee: Daryn Sharp
> Priority: Critical
> Attachments: HDFS-13112.1.patch, HDFS-13112.patch
>
>
> HDFS-4477 specifically did not acquire the fsn lock during token cancellation
> based on the belief that edit logs are thread-safe. However, log rolling is
> not thread-safe. Failure to externally synchronize on the fsn lock during a
> roll will cause problems.
> For sync edit logging, it may cause corruption by interspersing edits with
> the end/start segment edits. Async edit logging may encounter a deadlock if
> the log queue overflows. Luckily, losing the race is extremely rare. In ~5
> years, we've never encountered it. However, HDFS-13051 lost the race with
> async edits.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
