[
https://issues.apache.org/jira/browse/HADOOP-2141?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12703224#action_12703224
]
eric baldeschwieler commented on HADOOP-2141:
---------------------------------------------
Can't you use the the task dispatch time from the JT as the base time to
estimate progress? That would be consistent and a lot simpler than push
responsibility to the TT.
The assertion that it is ok to run a speculative remotely is a little
worrisome. It seems to me we should try to run them locally if possible.
After all, we might be running speculative execution exactly because a task is
slow due to networking issues... I'd like to see this tested. Another issue
is that the disk the source data is on may be overloaded or slow. So a
preference to run local to another copy of the data be optimal for maps...
Seems like a rich area for further research once this first patch is in. Andy,
Devaraj - Let's coordinate the validation of this patch on Y! resource. We
have some good test cases from the terasort contest.
> speculative execution start up condition based on completion time
> -----------------------------------------------------------------
>
> Key: HADOOP-2141
> URL: https://issues.apache.org/jira/browse/HADOOP-2141
> Project: Hadoop Core
> Issue Type: Improvement
> Components: mapred
> Affects Versions: 0.21.0
> Reporter: Koji Noguchi
> Assignee: Andy Konwinski
> Attachments: 2141.patch, HADOOP-2141-v2.patch, HADOOP-2141-v3.patch,
> HADOOP-2141-v4.patch, HADOOP-2141-v5.patch, HADOOP-2141-v6.patch,
> HADOOP-2141.patch
>
>
> We had one job with speculative execution hang.
> 4 reduce tasks were stuck with 95% completion because of a bad disk.
> Devaraj pointed out
> bq . One of the conditions that must be met for launching a speculative
> instance of a task is that it must be at least 20% behind the average
> progress, and this is not true here.
> It would be nice if speculative execution also starts up when tasks stop
> making progress.
> Devaraj suggested
> bq. Maybe, we should introduce a condition for average completion time for
> tasks in the speculative execution check.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
