You may have seen my Fusion upgrade article and now this one is about upgrading NexentaStor to the current GA (184.108.40.206). Regardless of the version this article will be updated so you know it works.
I can confirm it is OK to upgrade Fusion first, and NexentaStor second, and I think that is the right way to do it.
I also recommend you follow the exact steps of my upgrade as in the past they worked and slightly different steps did not.
There is no other prep work other than having a backup of what you are running on your Nexenta. Which I do but as I am not in an HA cluster I will be rebooting the array so I need to have the VMs on it powered off. If I was in a cluster there would be no outage.
- Log in on the console via SSH.
- Execute software version to see what your current version is.
- We can see that 5.0.2 is my current. (It was actually 220.127.116.11 but I have not had time to redo all the screenshots.)
- I also have a cluster error or two, but I don’t have a cluster and I am told by Nexenta that it is not an issue.
- Now use the software upgrade -n to see what is available.
- We can see there is outstanding updates.
- So we now use software upgrade to do the upgrade.
- The download finishes soon.
- We see the upgrade start. In 18.104.22.168 it was 258.2 MB and it took longer than 22.214.171.124.
- Before you know it (with 5.1 there was likely 25 minutes wait) you have a prompt to restart.
- And we do. We see at boot in the loader the 5.0.3 choice so things look good.
(This occurred in 5.0.3) I go off and do some work and forget what is going on. I look back and I see on the console that it is still 5.0.2 so I do a manual restart. When it restarts and we can see the console looks healthy we use Fusion to see if things worked. With the 5.0.3 to 5.1 upgrade it took perhaps 25 minutes after saying Y to reboot before it actually did.
When we log in to Fusion, we can see on the Administration page of the appliance that it is using 5.1 now.
Or, if we look on the console after we have logged in, we can also see the version.
So I start my VMs and all is good.
However, I did have a problem. It did not restart on its own. When I manually restarted the array it came up with some errors on the console (with SMF mentioned), and when I forced a reboot again, there were other errors. It restarted once or twice on its own, and by morning it was done and in perfect shape. The console looked good and so did Fusion when I logged in. I got support and engineering invoked and shared logs and they have not determined exactly what happened. So the upgrade took longer, and there were some odd moments with strange error messages but it ended well. I have only one issue outstanding and that is my statistics are not working – so total IOPS on my NFS share is blank. Support / Engineering at Nexenta is investigating. This was fixed and did not occur in the current upgrade.
- 11/5/17 – this was used to upgrade from 5.0.6 to 126.96.36.199.
- 4/16/17 – release. This was the 5.0.2 to 188.8.131.52 upgrade
=== END ===