[
https://issues.apache.org/jira/browse/CLOUDSTACK-8107?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Likitha Shetty resolved CLOUDSTACK-8107.
----------------------------------------
Resolution: Fixed
+Root Cause Analysis+
While taking snapshot of a volume, CS determines which storage pool contains
the volume. CS then chooses the endpoint to perform snapshot backup operation
by selecting any host that has the storage pool mounted on it. In case of
zone-wide primary storage, since every host present in the zone will have the
storage mounted, a random host will be chosen.
And if the host chosen is not in the cluster that contains the VM that the
volume is attached to, the snapshot operation will fail.
What happens in CS VMware resource layer if a host that belongs to a different
cluster is chosen as the endpoint?
During snapshot creation of a volume, we first snapshot the VM that the volume
is attached to. In case a host that doesn't contain the volume is chosen as the
endpoint then the look-up for the VM in the cluster is negative. Since we
couldn't find the VM, we create a worker VM to which we attach the disk to be
snapshot'ed (data path for detached volumes). And this operation fails in
vCenter.
+Proposed Solution+
While taking a snapshot of a volume, CS chooses the endpoint to perform backup
snapshot operation by selecting any host that has the storage containing the
volume mounted on it. Instead, if the volume is attached to a VM, the endpoint
chosen by CS should be the host that contains the VM.
> Failed to create snapshot from volume when the task is performed repeatedly
> in zone wide primary Storage.
> ---------------------------------------------------------------------------------------------------------
>
> Key: CLOUDSTACK-8107
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-8107
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Affects Versions: 4.5.0
> Reporter: Likitha Shetty
> Assignee: Likitha Shetty
> Priority: Critical
> Fix For: Future
>
>
> +Steps to reproduce+
> 1. Setup a CS zone with a VMware DC.
> 2. Ensure the DC has 2 clusters with a host each.
> 3. Add 2 Cluster-Wide primary storage pool to each of the clusters.
> 4. Add a Zone-wide primary storage pool.
> 5. Deploy a VM with a data-disk. Ensure both the ROOT and Data disk of the VM
> is in the zone-wide storage pool.
> 6. Take snapshots for the Data volume till the operation fails.
> In vCenter, the failure will be while Reconfiguring (worker)VM and the error
> will be '<>.vmdk was not found'.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
