This is a continuation of my upgrade series I was requested to post. In an earlier article you saw how to upgrade your external vCenter from 6.5 to 6.7U3, which must be performed prior to upgrading your VxRail. I know many engineers who have been waiting for 6.7U3 to go GA prior to upgrading their clusters from 4.5 code. Recently, I began upgrading my production clusters to 4.7.300 (GA release of 6.7U3). Below you will find the step by step process I followed to upgrade my clusters. I hope y’all find this article helpful.
- Ensure you have enough remaining capacity to put a host in maintenance mode and not require a resync
- Review the 4.7.300 release notes for known issues in 4.7.300 and have work around in place: https://support.emc.com/docu91467_VxRail-Appliance-Software-4.7.x-Release-Notes.pdf?language=en_US&request=akamai
- Confirm vCenter is within range of the desired ESX code level you want to go to
- Confirm all 3rd party software is compatible with the version of ESX you are going to
- Confirm all other VMware Components (vRealize, NSX, etc) are compatible with the ESX version you are going to
Additional but not required Pre-Req:
You should probably change the vSANClomRepairDelay from 60 to 360 during the upgrade to avoid timeouts when going from anything earlier than 4.5.314 to 4.7.x.
Since VMs are vMotioned as part of the upgrade, please ensure that VMs are capable of being vMotioned in advance. Examples of issues which may prevent vMotion:
• VM with an ISO mounted
• VM with external storage locally mounted
• VM pinned to a host
Side note: Don’t forget to take a snapshot of the VxRail Manager VM, vCenter and PSC (If external).
Due vCenter API changes in 6.7 you must be at VxRail 4.0.520 / VxRail 4.5.215 or later prior to upgrading your vCenter to 6.7 U1 or later. If you are below these minimums then upgrade your VxRail first followed by your vCenter.
If you have NSX – Upgrade first then run the following command on ESXi host to confirm NSX was upgraded:
esxcli software vib list |grep nsx
Useful ruby shell commands to run Prior to upgrade:
Check health of your cluster:
vsan.health.health_summary <cluster number>
Since you must put a host in maintenance mode you should also run:
vsan.whatif_host_failures 1 -s
Confirm your disks are balanced and none are close to 80% full, if close then run a proactive rebalance prior to upgrade:
Time to upgrade
Perform internet upgrade – if not possible then local upgrade
Enter your credentials to begin upgrade
Monitor the upgrade until completed
How to monitor VxRail upgrade via putty:
SSH into the VxRail Manager
#tail -f /var/log/mystic/lcm.log
Once the upgrade finishes it will display the new plug-in for 6.7 U3.
You can now see the VxRail plug-in the menu
Run a post health check to confirm health of the cluster
Go into VxRail plugin and confirm health
Finally, you can repeat the ruby commands again as a final health check
***Don’t forget to remove the snapshot of the VxRail Manager
Now you should be able to go through the vCenter console and see all the new monitoring and reporting tools in 6.7 U3 and start taking advantage of the new features.
As you can see if the proper prep work is done then the upgrade process of VxRail should be smooth. I hope this walk through answers any questions you may have had prior to planning your VxRail upgrade to take advantage of all the new features in 6.7 U3.