[
https://issues.apache.org/jira/browse/CLOUDSTACK-9707?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15904405#comment-15904405
]
ASF GitHub Bot commented on CLOUDSTACK-9707:
--------------------------------------------
Github user sateesh-chodapuneedi commented on the issue:
https://github.com/apache/cloudstack/pull/1868
The point of specifying a host is in deploy request itself means
inclination towards using that host for instance. Hence I think default value
should be false.
While global setting makes sense, rather than a global setting, it would be
better to add optional parameter in API itself which indicates whether deploy
should strictly try deployment on specific host, if any, or look for other
hosts as well. That way more granular control in case by case perspective.
> deployVirtualMachine API should fail if hostid is specified and host doesn't
> have enough resources
> --------------------------------------------------------------------------------------------------
>
> Key: CLOUDSTACK-9707
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-9707
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Reporter: Anshul Gangwar
> Assignee: Anshul Gangwar
>
> While using hostid parameter, vm gets deployed on another if the host
> given is running out of capacity. If host id is specified the deployment
> should happen
> on the given host and it should fail if the host is out of capacity. We are
> retrying
> deployment on the entire zone without the given host id if we fail once. The
> retry,
> which will retry on other hosts, should only be attempted if host id isn't
> given.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
