[
https://issues.apache.org/jira/browse/YARN-6050?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15872868#comment-15872868
]
Robert Kanter commented on YARN-6050:
-------------------------------------
{quote}If host1-3/rack1 have relaxLocality set to true, and host1-3 belongs to
rack1, this method get usable nodes = 3 + #rack1.{quote}
I don't think that's correct. We're using Max, not adding, so it would be
Max(Max(1, Max(1, 1)), #rack1) = #rack1. However, ignoring the rack, it should
be 1+1+1, not Max(1, Max(1, 1)). So this doesn't really handle resource
requests with multiple nodes.
{quote}If rack1 has 29 nodes, rack2 has 35 nodes, if relaxLocality set to true
for both of them, usable nodes = 35 instead of 29+35{quote}
You're right. This doesn't handle resource requests with multiple racks.
I'm guessing I'll need to either move this logic into {{nodeTracker}} or expose
more of the node/rack info from {{nodeTracker}} in order to do this properly.
Without knowing which resources are racks and which are nodes, I don't think
this can get any more accurate.
> AMs can't be scheduled on racks or nodes
> ----------------------------------------
>
> Key: YARN-6050
> URL: https://issues.apache.org/jira/browse/YARN-6050
> Project: Hadoop YARN
> Issue Type: Bug
> Affects Versions: 2.9.0, 3.0.0-alpha2
> Reporter: Robert Kanter
> Assignee: Robert Kanter
> Attachments: YARN-6050.001.patch, YARN-6050.002.patch,
> YARN-6050.003.patch, YARN-6050.004.patch, YARN-6050.005.patch,
> YARN-6050.006.patch, YARN-6050.007.patch, YARN-6050.008.patch
>
>
> Yarn itself supports rack/node aware scheduling for AMs; however, there
> currently are two problems:
> # To specify hard or soft rack/node requests, you have to specify more than
> one {{ResourceRequest}}. For example, if you want to schedule an AM only on
> "rackA", you have to create two {{ResourceRequest}}, like this:
> {code}
> ResourceRequest.newInstance(PRIORITY, ANY, CAPABILITY, NUM_CONTAINERS, false);
> ResourceRequest.newInstance(PRIORITY, "rackA", CAPABILITY, NUM_CONTAINERS,
> true);
> {code}
> The problem is that the Yarn API doesn't actually allow you to specify more
> than one {{ResourceRequest}} in the {{ApplicationSubmissionContext}}. The
> current behavior is to either build one from {{getResource}} or directly from
> {{getAMContainerResourceRequest}}, depending on if
> {{getAMContainerResourceRequest}} is null or not. We'll need to add a third
> method, say {{getAMContainerResourceRequests}}, which takes a list of
> {{ResourceRequest}} so that clients can specify the multiple resource
> requests.
> # There are some places where things are hardcoded to overwrite what the
> client specifies. These are pretty straightforward to fix.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
