[
https://issues.apache.org/jira/browse/AURORA-1183?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14354045#comment-14354045
]
Maxim Khutornenko commented on AURORA-1183:
-------------------------------------------
FWIW, the NearestFit logic has been reworked in
https://reviews.apache.org/r/31525/ and this is unlikely to happen in a new
build.
> job stay in PENDING: Host drained for maintenance, even though host is ACTIVE
> -----------------------------------------------------------------------------
>
> Key: AURORA-1183
> URL: https://issues.apache.org/jira/browse/AURORA-1183
> Project: Aurora
> Issue Type: Bug
> Components: Scheduler
> Affects Versions: 0.7.0
> Reporter: Bhuvan Arumugam
>
> Example: 1. I schedule a job with 1 instance, 1 cpu and pin the job to
> specific host, the job stay in {{PENDING: Insufficient CPU}} state, if there
> is >1 cpu in the offer. It's expected.
> Example: 2. I schedule a job with 2 instances, 1 cpu each and pin the job to
> specific hosts; one instance is scheduled/running. The other instance remain
> in {{PENDING: Host drained for maintenance}} state, if there is >1 cpu in the
> offer for this slave.
> The error message in example 2 is incorrect. It should indicate that job is
> PENDING due to insufficient CPU.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
