[
https://issues.apache.org/jira/browse/HDFS-8658?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15059882#comment-15059882
]
Kanaka Kumar Avvaru commented on HDFS-8658:
-------------------------------------------
Thanks for taking up a look at my patch [~ajisakaa]. Sorry for late response.
Yes logs may confuse when debug level is not enabled. I have updated patch to
use INFO level log for every attempt.
Now the logs should look like
{code}
INFO: Could not complete <path> retrying... current attempt is 1 out of 5
INFO: Could not complete <path> retrying... current attempt is 2 out of 5
INFO: Could not complete <path> retrying... current attempt is 3 out of 5
INFO: Could not complete <path> retrying... current attempt is 4 out of 5
INFO: Could not complete <path> retrying... current attempt is 5 out of 5
{code}
> Improve retry attempt log in DFSOutputStream#completeFile with current
> attempt number
> --------------------------------------------------------------------------------------
>
> Key: HDFS-8658
> URL: https://issues.apache.org/jira/browse/HDFS-8658
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: Kanaka Kumar Avvaru
> Assignee: Kanaka Kumar Avvaru
> Priority: Minor
> Attachments: HDFS-8658-00.patch, HDFS-8658-01.patch
>
>
> In a scenario when block update was delayed from DN to NN, following client
> log is observed which is confusing that file close failed after two attempts
> but it tried 5 times actually. So, it's better to log the current and
> remaining attempt counts.
> {code}
> 15/06/23 17:05:10 INFO hdfs.DFSClient: Could not complete
> /write2/tst5484._COPYING_ retrying...
> 15/06/23 17:05:16 INFO hdfs.DFSClient: Could not complete
> /write2/tst5484._COPYING_ retrying...
> copyFromLocal: Unable to close file because the last block does not have
> enough number of replicas.
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
