VSAN health warning discussion.

[ManivelR]

Hi Team,

Im getting some warning messages in my 4 node vSAN setup.Could you please suggest to clear this warning?

1) Disk usage on storage controller-->Non vSAN disks(OS disks only connected from all these 4 nodes) and VSAN disks connected in the same single RAID controller. As per VMware "ESXi host installation is permitted on non-vSAN disks attached to same controller". which means can i ignore this warning?

2) Need to install storcli on all the 4 ESXi  hosts to do addidtional checkups ? is it mandatory ?

3) What we need to do here?

Please suggest.

Thank you,

Manivel R

[sk84]

Hi ManivelR​

1.) Yes, installing ESXi on these disks is supported, so you could ignore this warning. See: VMware Knowledge Base

2.) It's not mandatory to install these controller utilities on the hosts, but they are used for additional health checks, for example. So it is at least recommended to install them. With vSAN 6.6 and 6.7, these tools can be installed for some controller cards directly from the vSphere Web Client. Just have a look at the section "Updates" under Cluster -> Configure -> vSAN -> Updates.

If you need to install the utility manually, this KB article will help: VMware Knowledge Base

3.) Does your vCenter have internet connectivity? The release catalog should be updated automatically if there are new patches or vSAN releases.

Another option is to upload the release catalog manually to you vCenter: VMware Knowledge Base

But I think the upload process is not very intuitive. You need the new vSphere client (HTML5) and you need to select your vCenter -> Configure -> vSAN -> Update.

If that doesn't help, you can try the following workaround: VMware Knowledge Base

And why don't you click on "Ask VMware"? There you will see exactly the relevant KB articles for the respective error message or warning. 😉

[TheBobkin]

Hello,

sk84

"1.) Yes, installing ESXi on these disks is supported, so you could ignore this warning. See: VMware Knowledge Base"

Not necessarily the case - for instance it is not supported to have ANY VMFS on H730 controllers used for vSAN, even just for logging/dumps and even the small VMFS created by default on install on boot device if not unselected/removed after:

VMware Knowledge Base

(and yes I am aware that kb says 'with lsi_mr3 driver' but I still wouldn't mix VMFS with vSAN on these controllers regardless)

I see Avago and MegaRAID there so in addition to them having never applied HCL JSON update there is a good chance these are H730 controllers (with ancient drivers+firmware).

ManivelR​

There appear to be at least two VMFS partitions in your screenshot (expanding the tabs and/or zooming out your browser helps here), you are using multiple controller (vmhba0 + vmhba2) on each host or different vmhba numbers on different hosts? If this is the case you will notice both are stating they have VMFS devices on them so likely not just default VMFS from boot device, that plus one of them looks to have a default-ish name and the other a given-name (VCS-...). Then again the test only indicates '5 items' and you said it is a 4-node cluster so this would indicate this is inconsistently misconfigured :smileygrin: .

While not mandatory to install storcli, this would be advised for visibility and improvements in interactability with the controller.

And as sk84, if you are ever in doubt about what a health check does or means, this is why we include handy URL links to kb articles providing additional details about every check (which can of course be copied if you don't have direct internet access from there).

Bob

[sk84]

TheBobkin​

Not necessarily the case - for instance it is not supported to have ANY VMFS on H730 controllers used for vSAN, even just for logging/dumps and even the small VMFS created by default on install on boot device if not unselected/removed after:

Oh, wow. I wasn't aware of the problems with the H730 controller and vSAN and that there are special limitations for this combination. Good to know. Thanks.

[ManivelR]

Thanks so much Sebastian and Bob for continuous feedback and response.

Yes Bob,it was five data stores(4 ESXi OS DS + one VMFS DS) and removed the VMFS DS from one ESXi host.

My vCenter server dont have internet connectivity and updated the HCL DB via offline 4 days ago.

But still im seeing some warning messages.Can we ignore this warning as of now? whenever VMware releases any critical patches/updates for VMware vSAN,we can download and update through offline. Am i right?

Thanks,

Manivel R