You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "Klaus Ma (JIRA)" <ji...@apache.org> on 2016/03/03 03:32:18 UTC

[jira] [Commented] (MESOS-4837) Ignore DOWN machine in "/machine/schedule"

    [ https://issues.apache.org/jira/browse/MESOS-4837?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15177020#comment-15177020 ] 

Klaus Ma commented on MESOS-4837:
---------------------------------

Further more, if the only update one slave's maintenance schedule, all slave will be updated to UP and trigger allocate(slaveId). Refer to the code in {{master/http.cpp}}: https://github.com/apache/mesos/blob/master/src/master/http.cpp#L1996

One proposed solution is to only handle {{DRAIN}} here, to avoid unnecessary; MESOS-4838 will enhance it to trigger allocate in batch.

cc [~jvanremoortere], [~kaysoky], [~hartem]

> Ignore DOWN machine in "/machine/schedule"
> ------------------------------------------
>
>                 Key: MESOS-4837
>                 URL: https://issues.apache.org/jira/browse/MESOS-4837
>             Project: Mesos
>          Issue Type: Bug
>          Components: master
>            Reporter: Klaus Ma
>
> Currently, if "/machine/schedule" did not include the {{DOWN}} machine, the request will be rejected; it's inconvenient for operator to collect all {{DOWN}} machines and add it into scheduler request. The proposal solution is to ignore {{DOWN}} machines in "/machine/schedule".



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