[
https://issues.apache.org/jira/browse/HDFS-3757?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13430760#comment-13430760
]
Colin Patrick McCabe commented on HDFS-3757:
--------------------------------------------
I agree that there's a risk of undesirable interaction with HDFS-3742. When I
get the chance, I'll try combining the two and seeing if it gives us what we
want.
I think the upside for this change is very high indeed, since it makes problems
debugable that might otherwise not be. If we absolutely can't get symbol
information for some reason, at least printing out the addresses would still be
a huge step forward, because we could then use gdb or something to find out
what functions were associated with those. However, there will always be bug
reports that come in where we don't have easy access to the binaries, so I
would prefer not to go this route if possible.
> libhdfs: improve native stack traces
> ------------------------------------
>
> Key: HDFS-3757
> URL: https://issues.apache.org/jira/browse/HDFS-3757
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: libhdfs
> Affects Versions: 2.2.0-alpha
> Reporter: Colin Patrick McCabe
> Assignee: Colin Patrick McCabe
> Priority: Minor
> Attachments: HDFS-3757.001.patch
>
>
> When libhdfs crashes, we often don't get very good stack traces. It would be
> nice to get a better stack trace for the thread that crashed.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
