xinglin commented on PR #6446:
URL: https://github.com/apache/hadoop/pull/6446#issuecomment-1897924023
> > > Ok so it seems we didn't remove any exception logging right? We are
just additionally keeping track of them and logging them again at the end?
> >
> >
> > correct. store all exceptions in a map and then print them if we fail
the request at the end.
>
> Its both right. You keep each stack trace as a warn
>
> ```
> String msg = String.format("Failed to read block %s for file %s from
datanode %s. "
> + "Exception is %s. Retry with the current or next
available datanode.",
> getCurrentBlock().getBlockName(), src,
currentNode.getXferAddr(), e);
> DFSClient.LOG.warn(msg);
> ```
>
> AND you add it to the map and print them later right?
>
> ```
> exceptionMap.get(datanode).add(e);
> ```
>
> So this part of the PR description "The existence of exception stacktrace
in the log has caused multiple hadoop users at Linkedin to consider this WARN
message as the RC/fatal error for their jobs. We would like to improve the log
message and avoid sending the stacktrace to dfsClient.LOG when a read
succeeds." doesn't seem like is actually happening since we are still printing
the warn each time and these will still get printed regardless of if the call
fails. Perhaps you meant to change this to info?
Hi Tom,
we will continue to print a WARN log line but it won't print the callStack.
We only print the callStack when we fail a read request. I updated the
description and added an example.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org
