cloudstack-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF subversion and git services (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CLOUDSTACK-7359) Volume migration between pools times out on cloudstack, but the migration completes on Xenserver
Date Mon, 18 Aug 2014 06:27:20 GMT

    [ https://issues.apache.org/jira/browse/CLOUDSTACK-7359?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14100337#comment-14100337
] 

ASF subversion and git services commented on CLOUDSTACK-7359:
-------------------------------------------------------------

Commit f099732be70c8ffa6d69870cc4b57f1ac3142cc0 in cloudstack's branch refs/heads/master from
[~devdeep]
[ https://git-wip-us.apache.org/repos/asf?p=cloudstack.git;h=f099732 ]

CLOUDSTACK-7359. Make sure timeout value is set for the MigrateVolumeCommand. This will
make sure the agent waits for the resource to complete (success/failure) migration of
volume on the hypervisor.


> Volume migration between pools times out on cloudstack, but the migration completes on
Xenserver
> ------------------------------------------------------------------------------------------------
>
>                 Key: CLOUDSTACK-7359
>                 URL: https://issues.apache.org/jira/browse/CLOUDSTACK-7359
>             Project: CloudStack
>          Issue Type: Bug
>      Security Level: Public(Anyone can view this level - this is the default.) 
>            Reporter: Devdeep Singh
>            Assignee: Devdeep Singh
>            Priority: Critical
>
> When a volume is live migrated from one storage pool to another, cloudstack sends a migratecommand
to the resource. The agent waits for wait*3 seconds for the command to complete. Whereas the
resource waits for migratewait seconds for the migration to complete. If volume migration
doesn't complete within migratewait interval the operation is cancelled. (wait and migratewait
for global configurable parameters). If wait parameter is much less than migratewait; the
agent assumes the operation hasn't completed and moves ahead. On the resource side the operation
would go ahead and complete at a later point in time. This leads to the volume being moved
to a different storage pool, but the location isn't updated on the db.
> Subsequent operation on the vm/volume like stop start etc. will fail.



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

Mime
View raw message