[ 
https://issues.apache.org/jira/browse/MESOS-4102?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15097721#comment-15097721
 ] 

Klaus Ma edited comment on MESOS-4102 at 1/14/16 8:04 AM:
----------------------------------------------------------

[~alexr]/[~neilc]/[~bmahler], I think the major gap is that allocator consider 
{{slaveIds}} includes all slaves when calculating 
{{remainingClusterResources}}; but it's not true when new slave added. So   one 
option is to calculate it based on all active slaves. The performance impact 
should be acceptable, because this loop is simple.

Draft RR to show the option: https://reviews.apache.org/r/42289/



was (Author: klaus1982):
[~alexr], +1 for persist laid aside; it reduce the idle time of resources. For 
your comments on "quota first": if there's quota'ed framework, the Quota is 
first; if there's no quota'ed framework, allocator should reserve slave's 
resources for laid aside and allocate other resources in DRF.

> Quota doesn't allocate resources on slave joining
> -------------------------------------------------
>
>                 Key: MESOS-4102
>                 URL: https://issues.apache.org/jira/browse/MESOS-4102
>             Project: Mesos
>          Issue Type: Bug
>          Components: allocation
>            Reporter: Neil Conway
>            Assignee: Alexander Rukletsov
>              Labels: mesosphere, quota
>         Attachments: quota_absent_framework_test-1.patch
>
>
> See attached patch. {{framework1}} is not allocated any resources, despite 
> the fact that the resources on {{agent2}} can safely be allocated to it 
> without risk of violating {{quota1}}. If I understand the intended quota 
> behavior correctly, this doesn't seem intended.
> Note that if the framework is added _after_ the slaves are added, the 
> resources on {{agent2}} are allocated to {{framework1}}.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to