[
https://issues.apache.org/jira/browse/CLOUDSTACK-7360?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14100353#comment-14100353
]
Sateesh Chodapuneedi commented on CLOUDSTACK-7360:
--------------------------------------------------
Following procedure to add host doesn't work if the cluster is using Nexus
1000v as backend for atleast one traffic type.
1) Add ESXi host to a VMware cluster in vCenter
2) Add host using CCP UI.
Side effect of the failure is the existing cluster's record in database is
being deleted as a result of handling discovery failure while adding host. This
should be avoided such that a host add attempt failure should not affect the
existing cluster.
While adding a VMware cluster (which is using Nexus 1000v as network backend)
CCP does validation of Nexus VSM's IP address & it's credentials to be used by
CCP later for virtual network orchestration. addCluster API does supply meta
data related to VSM which is available for validation during discovery of
cluster. But in case of addHost the work flow doesn't supply the VSM metadata.
This missing information is resulting in failure while validating during
discovery process.
We need 2 changes,
1) Ensure host discovery / validation failure while adding host should not
affect existing cluster in any way.
2) Skip validation of VSM metadata during host addition to cluster. It is
admin's responsibility to add ESXi host to the cluster and the associated VSM
correctly before trying to add it using CCP UI.
> [vmware] Add host to existing cluster fails if the cluster is using Nexus
> 1000v as backend for atleast one traffic type.
> ------------------------------------------------------------------------------------------------------------------------
>
> Key: CLOUDSTACK-7360
> URL: https://issues.apache.org/jira/browse/CLOUDSTACK-7360
> Project: CloudStack
> Issue Type: Bug
> Security Level: Public(Anyone can view this level - this is the
> default.)
> Components: VMware
> Affects Versions: 4.2.0, 4.3.0, 4.4.0
> Environment: vCenter/ESXi 5.0, Nexus 1000v 1.4
> Reporter: Sateesh Chodapuneedi
> Assignee: Sateesh Chodapuneedi
> Priority: Critical
> Fix For: 4.5.0
>
> Original Estimate: 72h
> Remaining Estimate: 72h
>
> Steps to reproduce:
> 1) Before creating zone, set global params vmware.use.nexus.vswitch &
> vmware.use.dvswitch is to "true".
> 2) Complete zone wizard
> 3) Using vCenter add ESXi host to a cluster, which is already being managed
> by CCP
> 4) Using CCP UI add the same host to that cluster - This fails
--
This message was sent by Atlassian JIRA
(v6.2#6252)
