I thought I would make my first VxRail post over a topic I get asked frequently about, which is how difficult is it to upgrade a VxRail. Having performed many VxRail upgrades over the years I’d like to start by saying these are my own personal notes and style for upgrading VxRail. This post is the process followed for an external PSC and VCSA. There are additional steps if you have an embedded PSC/VCSA. One last thing to take into consideration is the code level of your external vCenter and make sure it is line with the VxRail version you are upgrading too.
First, we need to have an understanding of the importance of slack space. vSAN “slack space” is free space that is set aside for operations such as host maintenance mode data evacuation, component rebuilds, rebalancing operations, and VM snapshots. Activities such as rebuilds and rebalancing can temporarily consume additional raw capacity. Host maintenance mode temporarily reduces the total amount of raw capacity a cluster has. This is because the local drives on a host that is in maintenance mode do not contribute to vSAN datastore capacity until the host exits maintenance mode.
Second, we need to do a little math. With vSAN you must reserve a percentage of your capacity for slack space. VxRail best practice is 25%. Since upgrading a VxRail requires putting a node into maintenance mode one at a time we must honor a N+1 architecture in respect to slack space.
The math is simple. If you have 5 nodes contributing 25TB each, then during an upgrade you will have 4 nodes contributing 25TB for 100TB. Take 25% off the top and you have a remainder of 75TB. If you are currently utilizing 80TB of storage space then I recommend a storage vMotion (Another datastore or cluster) temporarily of some workloads to prevent a re-sync from occurring during the upgrade.
Luckily as part of the VxRail upgrade process it will set default timeouts to a responsible amount of time now preventing timeouts and delays in the upgrade process.
The first step in any upgrade process should be to take a snapshot of your VxRail Manager VM
There are two ways to select the upgrade bundle. Local upgrade allowing for you to choose the file stored locally or Internet upgrade and pulling down the bundle from DellEMC
Once the bundle has been uploaded through either method you will need to enter you administrator@vsphere.local and root credentials. Once this step is completed the upgrade begins.
Don’t be alarmed as you will be logged out of the VxRail Manager but wait two minutes and log back in to monitor the upgrade.
Another option for monitoring the reboot of nodes is to log into the iDRAC of each host and bring up their console and watch the reboot process.
Your choice or you can choose to walk away and come back a few hours later to see that the VxRail has been successfully upgraded
Congratulations you now have a healthy upgraded cluster.
In a later blog I intend to mention the ruby shell and some helpful commands you can run should you believe your upgrade is taking longer than normal. Please feel free to reach out should you have any questions or need help preparing for your next VxRail upgrade.