[
https://issues.apache.org/jira/browse/GIRAPH-291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13431204#comment-13431204
]
Eli Reisman commented on GIRAPH-291:
------------------------------------
Sorry I did not get that impression from the other threads, sounds great. I
figured verifying the progress calls in the lock code would relate to trying to
solve the timeout problem with the existing solution, and we already know it
doesn't work, so I didn't really follow the purpose.
I believe the main contribution of 246 is making the timeouts go away until we
know why they are happening. If we want to pursue the predicate lock solution,
that sounds great but the annoyance of this problem is that you don't know if
you've solved it without running a bunch of long jobs, and so guessing at this
has proven time consuming. Once I hit something that worked, I was done
guessing as I have bigger fish to fry to get Giraph to the scale I need and
this seemed like an interfacing-with-Hadoop problem, not a Giraph problem.
As far as this patch, looks great to me.
> PredicateLock should have a constructor to take in a custom waiting time and
> additional testing
> -----------------------------------------------------------------------------------------------
>
> Key: GIRAPH-291
> URL: https://issues.apache.org/jira/browse/GIRAPH-291
> Project: Giraph
> Issue Type: Improvement
> Reporter: Avery Ching
> Assignee: Avery Ching
> Attachments: GIRAPH-291.patch
>
>
--
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
