HPE CLX Virtual Machine Protection Fails

Recently a colleague was attempting to configure the cluster extension protection (CLX) which is required by our Active/Active HPE 3PAR storage setup. Our Microsoft Failover Clusters are stretched over geographically separate sites and the Cluster Share Volumes (CSV) are spread over the 3PAR storage arrays in each site. We use the CLX software to manage which site a CSV is active in, this also requires us to add a custom resource to each virtual machine.

In this instance my colleague was able to add the custom CLX resource to the virtual machine but when he tried to bring that resource online it would fail with a fairly generic message in the Failover Cluster Manager. Having worked on this setup and seen all sorts of CLX errors I had a feeling I knew what it was. Of course the easiest thing is to check the CLX logs which will tell you far more than the Microsoft interface can.

By default the CLX logs can be found here –

C:\Program Files\Hewlett-Packard\Cluster Extension 3PAR\log

As you can see there are a number of log files, in this case as we are trying dealing with a virtual machine we want to look at the file VMClxMscs.log

CLX Logs

With the HP/HPE split I wouldn’t be surprised if this location changes so please be mindful when looking for it. Once you’ve opened the log the easiest thing to do is scroll to the very bottom to see the most recent entries. If we look at my example from the issue below we can see the system indicates the CSV disk is actually running from a different datacenter (DC) to the virtual machine. This will prevent us from bringing the CLX resource online. To resolve this issue we need to bring the CSV to the same DC or move the virtual machine – either way they both need to be in the ‘same place’ for this to work.

[09/19/17 14:35:35][5200][VM_CLX: CLX_3PAR_SCVMM BSA-LAZARUS Resources][ERROR] Virtual Machine resource owner node and CSV disk resource owner node are NOT in the same DataCenter. Hence the Cluster Resource {CLX_3PAR_SCVMM BSA-LAZARUS Resources} and dependent Virtual Machine {SCVMM BSA-LAZARUS} cannot be brought online in this Cluster Node
[09/19/17 14:35:36][5200][VM_CLX: CLX_3PAR_SCVMM BSA-LAZARUS Resources][INFO] Virtual Machine Configuration Information
Virtual Machine Name : SCVMM BSA-LAZARUS
Virtual Machine Configuration Name : SCVMM BSA-LAZARUS Configuration
CLX Resource Name managing VM : CLX_3PAR_SCVMM BSA-LAZARUS Resources
Virtual Machine Role Name : SCVMM BSA-LAZARUS Resources
CSV Disk Name hosting VM : SP_BSA_hv_vv17
CLX Resource managing CSV Disk : CLX_3PAR_BSA_hv_vv17
CSV Role Name : 40d991ed-0b1b-4805-a367-641820cd97fd
Cluster Nodes in Datacenter A : BSA-gphv01 BSA-gphv02 BSA-GPHV03 BSA-gphv04
Cluster Nodes in Datacenter B : BSA-sphv01 BSA-SPHV02 BSA-sphv03 BSA-sphv04
Cluster Node owning VM Role : BSA-sphv04
Cluster Node owning CSV Role : BSA-gphv04

These days I would suggest using PeerPersistence with both VMware and Hyper-V environments. When we built all our Microsoft clusters 5 years ago this wasn’t an option so we had to go with CLX. Always be mindful that Failover Cluster Manager may not give the best error/log information, if you are using custom resources remember to check their specific logs when encountering problems.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.