[
https://issues.apache.org/jira/browse/SLIDER-930?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15003508#comment-15003508
]
Gour Saha commented on SLIDER-930:
----------------------------------
[~sherryxg] the patch looks ok. From the nature of this bug, seems like it
needs some functional tests. But then it might not be very straight forward to
validate that the window is working, since we need to know the value of
{{yarn.resourcemanager.am.max-retries}} of the cluster the fun test will be run
against.
Say the value is n, then we could kill the AM n-1 times within the window, and
then let the window expire and kill it 1 more time, and the app should continue
to run. Of course we should kill AM n times as well and validate that the app
shuts down. But then how do we write the test without knowing the value of n.
Is my understanding wrong? If yes, then I am happy, since it means that we can
write some deterministic fun tests. If not, then is it possible for you to do
some manual testing on this?
> Incorporate Yarn feature of resetting AM failure count into Slider AM
> ---------------------------------------------------------------------
>
> Key: SLIDER-930
> URL: https://issues.apache.org/jira/browse/SLIDER-930
> Project: Slider
> Issue Type: Bug
> Components: appmaster
> Affects Versions: Slider 0.80
> Reporter: Gour Saha
> Assignee: Sherry Guo
> Fix For: Slider 0.90
>
> Attachments: SLIDER-930-001.patch
>
>
> YARN-611 provides this feature. Currently Slider apps are bound by the number
> set for yarn.resourcemanager.am.max-retries in the cluster. By default this
> value is set to 2, which is very low for long running services.
> Slider AM should use the feature provided in YARN-611 and set an interval
> after which the failure count will be reset to 0.
> I believe the API to call on ApplicationSubmissionContext is
> attemptFailuresValidityInterval. To start with Slider can set it to 5 mins
> which should be a reasonable default.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
