#Ontap upgrade
1 messages · Page 1 of 1 (latest)
The upgrade to 8.3x is generally not fun. There are lots of rules that need to be followed for the upgrade to work. I just did one last week. It took about 45 minutes of trial/error until I got the networking to the point where the upgrade would go. Honestly it depends on your network setup. Download the 8.3 upgrade/revert guide and pay attention and follow the instructions.
Can't believe I'm going to say this, but do you really need to upgrade? Versus migrating the data off, wiping, and then install a newer release? If @pure grotto did one last week, then he is more up to date than even anyone in Support.
After getting on 8.3, upgrades are easy although it’s cli only until updated to the latest 9.1 release
I’ve also done three in the past six months
Is there any guide to doing this?
Log into the support site. Go to Data ONTAP 8 (not: ONTAP 9).
Look around. There is documentation for Data ONTAP 8.3.1 ands one of the document is the Upgrade/revert guide.
If you are going to do this, please update to the latest 8.2.4P release. It should make the upgrade to 8.3.x better
If you can migrate the data off, then access the boot menu via halting and pressing "ctrl + c" when prompted, then you should be able to just install the new SW using boot menu option.
In current versions of ONTAP, this is option 7 "Install new Software" ..
There is a KB on it here:
https://kb.netapp.com/onprem/ontap/os/How_to_perform_the_option_7_Install_new_software_first_from_the_ONTAP_boot_menu
I think there was a bug back in 8.3 with broadcast domains and this led to problems when going to 9.1, its been far to long for me to remember.
the usual prechecks will also tell you what will happen. The mainissue is that all client queries (DNS, LDAP, Kerberos, ...) might go through a completelly different LIF after the upgrade. so if these servers are not reachable through your SVM's existing LIFs, you will run into trouble. In a fully-routed network you will probably be fine
The one I did last week, four nodes. I missed it. Three of the nodes updated just fine (set diag; system node upgrade-revert show -node local). On the last node, the command failed! Turned out to be one of the intercluster lifs had a different mtu than the other three.