Original Message:
Sent: Tue August 19, 2025 03:11 PM
From: Eric Derr
Subject: Policy Based HA and VmWare
Thiago,
Thank you very much for your response. It worked. I couldn't reboot the server as it's in production, but both creating
a new volume, and un-mapping and remapping the hosts to a current volume both worked. And yes, that was the situation that we did the code upgrade after the volumes had been created. Again, thank you very much.
Eric
------------------------------
Eric Derr
Original Message:
Sent: Tue August 19, 2025 02:12 PM
From: Thiago Lucas
Subject: Policy Based HA and VmWare
If a VMFS filesystem was in place at the volume (so already mapped and formatted) previously you update to 8.7.3.x I believe it wont update WEAR support on VMware. I would try first rebooting the hosts, if no success unmap and map back the volume, if still no Clustered VMDK support, present a new volume to check, you will see if the device supports while creating the datastore on it.
------------------------------
Thiago Lucas
Original Message:
Sent: Mon August 18, 2025 12:26 PM
From: Eric Derr
Subject: Policy Based HA and VmWare
Chris,
Thank you for your reply. I have opened a case and will update the thread with their response.
Eric
------------------------------
Eric Derr
Original Message:
Sent: Fri August 15, 2025 06:27 AM
From: Chris Bulmer
Subject: Policy Based HA and VmWare
Hi Eric,
SCSI Persistent Reservations with policy-based high availability is supported in that code version so it's odd that it's showing as unsupported.
The Broadcom article you've linked to has three criteria:
- Support SCSI-3 reservation type WEAR (i.e., write exclusive all registrant)
- Support ATS commands
- Have 512/512e sector size disks
and all of these are supported in this code version with PBHA.
Please open a support ticket so we can take a look at what's going on.
Thanks,
Chris
------------------------------
Chris Bulmer
Software architect, IBM Storage Virtualize
Original Message:
Sent: Thu August 14, 2025 01:19 PM
From: Eric Derr
Subject: Policy Based HA and VmWare
We are in the process of migrating from a v5030 to a FlashSystem 5300. I'm the storage/linux administrator , so excuse my ignorance with vmware.
We are using PBHA with the new FlashSystems, we have two, at release 8.7.3.3. I was given this question by our vmware administrators:
We aren't seeing the option to Enable "Clustered VMDK" on the new LUNS in vmware. Checked the requirements via the article below and doesn't look like WEAR reservations are supported. Is there a setting on your side that would need adjusting?
Clustered vmdk datastore LUN support check
Broadcom |
remove preview |
|
Clustered vmdk datastore LUN support check |
To check if a LUN meets these requirements, use vsish on the ESXi commandline: e.g. for device eui.xxxxxxxxxxxxxxxxxxxxxxxxxxxxx006 vsish -e get /storage/scsifw/devices/eui.xxxxxxxxxxxxxxxxxxxxxxxxxxxxx006/info vsish -e get /storage/scsifw/devices/eui.xxxxxxxxxxxxxxxxxxxxxxxxxxxxx006/info Sample output: Note for VAAI ATS status and is WEAR reservation supported? , a value of 1 => supported, value of 0 => unsupported |
View this on Broadcom > |
|
|
I'm assuming this may be because there are no clusters with PBHA at this time.
If anyone has any thoughts on this, it would be greatly appreciated.
Thanks,
Eric
------------------------------
Eric Derr
------------------------------