[
https://issues.apache.org/jira/browse/MAPREDUCE-1167?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vinod K V updated MAPREDUCE-1167:
---------------------------------
Status: Open (was: Patch Available)
The patch looks OK. Only minor comments:
- ProcessTreeDump is printing Rss size in pages but the header reads
RSSMEM_USAGE(BYTES). This should be fixed.
- In TestProcfsBasedProcessTree, Long.ParseLong(String) is used in many
places. This is costly, you can use the long value directly if long type is
needed, otherwise Long.valueOf(long) if Long type is. For example, see +408
after applying your patch.
- TestProcfsBasedProcessTree failed. You need to modify the pattern at
TestProcfsBasedProcessTree.java +188.
bq. The reason for 3 is because getCumulativeVmem and getCumulativeRssmem
should be consistent. When /proc/ is not available, getCumulativeVmem will
return 0 instead of throwing Exception.
I think this was wrongly done and should be changed. Will file a new issue. We
can keep whatever you've done for now.
> Make ProcfsBasedProcessTree collect rss memory information
> ----------------------------------------------------------
>
> Key: MAPREDUCE-1167
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-1167
> Project: Hadoop Map/Reduce
> Issue Type: New Feature
> Components: tasktracker
> Affects Versions: 0.22.0
> Reporter: Scott Chen
> Assignee: Scott Chen
> Fix For: 0.22.0
>
> Attachments: MAPREDUCE-1167-v2.patch, MAPREDUCE-1167-v3.patch,
> MAPREDUCE-1167-v4.patch, MAPREDUCE-1167.patch
>
>
> Right now ProcfsBasedProcess collects only virtual memory. We can make it
> collect rss memory as well.
> Later we can use rss in TaskMemoryManagerThread to obtain better memory
> management.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
