[
https://issues.apache.org/jira/browse/YARN-4287?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14975410#comment-14975410
]
Wangda Tan commented on YARN-4287:
----------------------------------
Thanks for sharing your thoughts, [~nroberts]!
bq. iiuc, we didn't get good locality before the patch either. i.e. canAssign()
would return true for NODE-LOCAL and OFF-SWITCH without delay.
Yes, you're correct, I think we can safely use min(computed-offswitch,
configured-offswitch) as final offswitch/rack delay.
bq. 1) I need to change the way rackLocalityDelay is specified because it
doesn't handle the case where the configuration value is larger than the
cluster size. I was thinking of just scaling it. Let's say
node-locality-delay=5000, rack-locality-delay=5100, cluster_size is 3000. In
the existing code, node-locality-delay would automatically get lowered to 3000.
Instead, it will lower rack-locality-delay to 3000, and node-locality-delay
will be proportionally adjusted (5000 * 3000 / 5100) = 2941.
I think instead of scaling, I suggest to simply cap rack/offswitch delay by the
cluster size, so:
- rack-delay = min(offswitch, node-locality-delay, cluserSize)
- offswitch-delay = min(offswitch, clusterSize)
The scaling behavior could be hard to explain to end users.
bq. 2) Add a configurable boolean that controls whether a rack-local assignment
resets missed_opportunities to 0 (old behavior), OR node-locality-delay (new
behavior). Default of new behavior.
This is fine to me since this is a configurable item and you have done tests
for this change already.
> Capacity Scheduler: Rack Locality improvement
> ---------------------------------------------
>
> Key: YARN-4287
> URL: https://issues.apache.org/jira/browse/YARN-4287
> Project: Hadoop YARN
> Issue Type: Improvement
> Components: capacityscheduler
> Affects Versions: 2.7.1
> Reporter: Nathan Roberts
> Assignee: Nathan Roberts
> Attachments: YARN-4287-v2.patch, YARN-4287-v3.patch, YARN-4287.patch
>
>
> YARN-4189 does an excellent job describing the issues with the current delay
> scheduling algorithms within the capacity scheduler. The design proposal also
> seems like a good direction.
> This jira proposes a simple interim solution to the key issue we've been
> experiencing on a regular basis:
> - rackLocal assignments trickle out due to nodeLocalityDelay. This can have
> significant impact on things like CombineFileInputFormat which targets very
> specific nodes in its split calculations.
> I'm not sure when YARN-4189 will become reality so I thought a simple interim
> patch might make sense. The basic idea is simple:
> 1) Separate delays for rackLocal, and OffSwitch (today there is only 1)
> 2) When we're getting rackLocal assignments, subsequent rackLocal assignments
> should not be delayed
> Patch will be uploaded shortly. No big deal if the consensus is to go
> straight to YARN-4189.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
