vSAN Enabled Cluster - Mix RAID0 with HBA mode

flaxseed · ‎09-14-2016

Hi all,

We currently have the following setup.

vSAN Cluster (6 hosts)

- 3 hosts running under RAID0

- 3 hosts running under HBA mode

The first 3 nodes have been running for almost a year without any issues at all. The moment we have introduced the other 3 nodes, running under HBA mode, the vSAN datastore is no longer responsive, as it cannot be browsed anymore and we cannot migrate any VMs from the first 3 nodes to the other 3 new. Do you think it's because of the different running configuration? All 6 nodes are on the same ESXi version, 6.0U1 3073146.

**When trying to list the contents of the vSAN datastore from one of the new nodes, we get the below.

And when we issue esxcli vsan cluster get, all the nodes seem to be present.

Thank you in advance.

zdickinson · ‎09-14-2016

Good morning, are they different because the cards a different one does not support HBA mode? Thank you, Zach.

darcidinovmw · ‎09-14-2016

There shouldn't be any real issues with this configuration from what I have seen. Did you add the hosts from the vsphere client or from the CLI? You didn't lose connectivity when adding the first new host?

Honestly, this seems to be a networking issue. Can you launch into the ruby vsphere console and do a health check?

http://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/products/vsan/vmware-ruby-vsphere-c...

Doug Arcidino

Doug Arcidino VCP-DCV 4/5/6, VCP-DTM 5/6/7, VCAP-DCV Deploy/Design 6 If this answer was helpful, please mark it as answer I work for VMware Disclaimer: Any views or opinions expressed here are strictly my own. I am solely responsible for all content published here. Content published here is not read, reviewed or approved in advance by VMware and does not necessarily represent or reflect the views or opinions of VMware.

flaxseed · ‎09-14-2016

Hi,

Thank you for the reply.

All 6 hosts have the same RAID controller, PERC H730P Mini.

The last update is that we have separated the hosts in two clusters instead, the first running in RAID0 and the other one in HBA. Everything seems to be fine now, no issues, so maybe there is indeed an issue with mixing the modes, not sure.

flaxseed · ‎09-15-2016

Hi Doug,

Thank you for your reply.

The hosts were added to the cluster through the web-based vSphere Client and there weren't any issues related with networking, as in "host x cannot communicate with the cluster" and so on. As I have mentioned above though, we have separated the clusters and everything seems to be fine now.

Will try the health check and report back my findings, thanks anyway.

darcidinovmw · ‎09-19-2016

All I can say is that is extremely weird. I'd love to know the root cause. Is it possible for you to grab logs from vCenter and all of your hosts to create a case with GSS? It's very valuable for our customers to share the logs with us so we can attempt to improve our products.

Doug

Doug Arcidino VCP-DCV 4/5/6, VCP-DTM 5/6/7, VCAP-DCV Deploy/Design 6 If this answer was helpful, please mark it as answer I work for VMware Disclaimer: Any views or opinions expressed here are strictly my own. I am solely responsible for all content published here. Content published here is not read, reviewed or approved in advance by VMware and does not necessarily represent or reflect the views or opinions of VMware.

flaxseed · ‎11-08-2016

Hi Doug,

Just wanted to get back to you, even though some time passed. The issue was some misconfiguration from our side which has been resolved, therefore I would like to confirm that you may have RAID0 and HBA mode enabled hosts in the same vSAN cluster.

Thank you and everyone else for their contribution to the thread.

flaxseed · ‎11-08-2016

Hi,

Apologies for the late reply.

The hosts are identical in specs, but as explained below, it was a misconfiguration at the virtual network from our side.

Thank you.