We have 4 Hosts in vSAN stretched cluster where configured 2 hosts are in preferred site and 2 hosts are in secondary site and Witness VM running on dedicated standalone esxi host .
Now we planning to upgrade vSAN infra with below High level plan
So my questions are below at respective steps that
Step 2 where deciding to place host in Maintenance Mode (MM) ,so should i keep host in MM by selecting ensure accessibility or full data migration as we have only two hosts in preferred and secondary site (attached screen shot know disk space details) ?
As Fw and ESXI upgrade takes around 4 hours,so Can I set VSAN.ClomRepairDelay to higher value (4 Hours) If i chose "ensure accessiblity while placing host in MM.
Step 3 where upgrading standalone esxi host on which witness VM is running, as it’s standalone host, so I have to power off witness VM to upgrade ESXi FW and ESXi host .Is any risk involved in power off witness and performing maintains on standalone host ?
Step 5 Once standalone host and witness VM up, then I will perform witness appliance upgrade via Update manager ,Is it correct and recommended method ?
Thanks in advance for your time and help
I would use "ensure accessibility". I would indeed increase ClomRepair, as that would prevent a potentially unneeded rebuild to kick off, which may cause congestion on the network.
I probably also wouldn't bother upgrading the Witness VM with VM, I would deploy a new one and add that instead of the existing as a witness.
Thanks for your valuable suggestion .
So When updating the cluster, it is preferable to upgrade the Witness ESXi first and then deploy new the virtual witness or vice versa ?? Any recommendation ?
The recommendation is to upgrade all the hosts in Site A first, for instance, then all the hosts in Site B, and finally the Witness appliance. You can upgrade the witness using VUM in some case (smaller upgrades/patches); however, while moving into bigger releases such as 6.5 to 6.7. you need to deploy a new Witness appliance as some requirements need to be met. In this case, the CPUs from a 5.5-6.5 witness are not supported by vSphere 6.7.
Like Duncan said, it is easier/faster just to deploy a new appliance.
Here is some additional info:
Hi, thanks for your valuable suggestion and now I’m ready for upgrade of vSAN Infra.
I have one query about HW compatibility since we have C240-M4SX servers and having Cisco 12G SAS Modular Raid Controller(UCSC-MRAID12G) which is supported matrix from Cisco and Vmware /VSAN HCL for ESXi 6.7 U2 and vSAN 6.7 U2
As we are not using vSAN ready node, so while checking HW compatibility, found strange thing about Cisco 12G SAS Modular Raid Controller ( since I targeted upgrade from 6.0 U2 to 6.7 U2,but I looked 6.7 U3 compatibility ) that as per vSAN HCL,RAID controller(UCSC-MRAID12G) is supported till ESXi 7.0/vSAN7.0,however when I checked Footnotes for vSAN 6.7 U3 stating that“Firmware 4.620.00-7322 is packaged within container 24.12.1-0433.Please use one of the following UCS releases for Storage controller firmware & Driver - UCS C-series Standalone releases – 4.0(2d) , UCSM-Managed C series releases – 4.0(2d)”
But as per Cisco compatibility Matrix,6.7 U3 does not support UCSC-MRAID12G controller for firmware version 4.0(2d),but Vmware Matrix says that use 4.0(2d) for ESXi6.7 U3/vSAN6.7 U3
So quite confused H/W compatibility of vSAN HCL/Cisco HCL, So should it safe/supported to run vSAN 6.7 U3 on C240-M4SX with 4.0(2d) as per vSAN HCL ?
Interesting so are you saying that if you go from 6.5 u3 to 6.7 u3 you need to deploy a new witness. I have upgraded a few stretched clusters using Virtual witnesses with update manager and it upgraded without any issues. I can understand 6 - 6.7 though