[jira] [Commented] (CLOUDSTACK-6218) VR commands need to be serialized in VR resource

["ASF subversion and git services (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CLOUDSTACK-6218?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13930889#comment-13930889 ] 

ASF subversion and git services commented on CLOUDSTACK-6218:
-------------------------------------------------------------

Commit b399c315a3fceae699fa24dc1ff2bbae191ddd07 in cloudstack's branch refs/heads/master from [~yasker]
[ https://git-wip-us.apache.org/repos/asf?p=cloudstack.git;h=b399c31 ]

CLOUDSTACK-6218: Serialize VR commands in VR resource


> VR commands need to be serialized in VR resource
> ------------------------------------------------
>
>                 Key: CLOUDSTACK-6218
>                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-6218
>             Project: CloudStack
>          Issue Type: Bug
>      Security Level: Public(Anyone can view this level - this is the default.) 
>          Components: Virtual Router
>    Affects Versions: 4.4.0
>            Reporter: Sheng Yang
>            Assignee: Sheng Yang
>             Fix For: 4.4.0
>
>
> In very large scale parallel deployment, multiple commands executed at VR would result in VR itself's lock time out, since the later commands would need to wait longer and longer for execution.
> This fix would put a lock in VR resource part, and make sure every commands to VR would be queued before executed in VR, prevent timeout in the VR. However, the timeout would still possible with agent manager level, if VR resource has too many commands to deal with and cannot give back answer in time. But this timeout can be adjusted from mgmt server side, so the problem is mitigated. 



--
This message was sent by Atlassian JIRA
(v6.2#6252)