https://worldofeuc.com logo
Join Slack
Powered by
I have a question about the upgrade from KVM based...
# citrix-netscaler
n
I have a question about the upgrade from KVM based VPX to AHV officially supported image. This doc references two different update scenarios, one for standalone VPX and the other for HA pair. The latter talks about ejecting the secondary node from the HA pair and adding a newly deployed AHV VPX using the same NSIP in its place, as "zero downtime". I'm probably missing something obvious, but I'm not seeing how it's functionally different from just running the "standalone" upgrade method against the secondary HA node then "failing over" to it, which is also required in the VPX replacement method as well? https://docs.netscaler.com/en-us/vpx/current-release/install-vpx-on-nutanix/migrate-from-citrix-ready-kvm-to-official-vpx-nutanix.html
To me it seems like just running the upgrade bundle from the NetScaler VPX shell is less work than having to deploy a new appliance and do an ejection of the original KVM VPX from the HA pair. I used this method on a test HA VPX instance we have and didn't run into any issues.
r
It's have an HA pair and split it and remake or take a single and make an HA pair. The core that NetScaler is built on was BSD for the normal and a newer Linux distribution for the AHV version. So, best way is to let the config replicate and replace the vm.
👍 1
m
I upgraded my Nutanix HA pair based on the KVM package to the Nutanix native package with no issue. Upgraded the secondary first as usual, failed over, then upgraded the primary. I believe the latest release supports an in place upgrade.
👍 1
n
That's the route I ended up going as well. Seemed to work fine upgrading a test HA pair we typically start upgrades with, then I upgraded a standalone VPX in a secondary site with the "in place" upgrade model, then also did a production HA pair in our primary site like you did, starting with secondary, failing over, then updating former primary. The only "problem" I ran into was the custom certificate applied to the management IP/services had to be re-bound, and we had to re-activate with LAS afterwards, even though it'd already been done.
Did anyone have issues during the upgrade to the latest 14.1 build addressing the vulns disclosed last week, or even just when doing the KVM to AHV "upgrade" procedure where it either lost config or modified the config? I've identified a couple issues after going to the latest 14.1 build, even when just upgrading a standalone VPX, where default certs were reverted to for the mgmt IP and our custom cert had to be re-bound, or with included ciphers in cipher groups being modified post upgrade (I'm comparing the two configs side by side in ExamDiff and can plainly see it).
seems to be mainly related to the "default" cipher group being added back to custom cipher groups after upgrade
c
We had many issues doing the upgrade from KVM to AHV version, we ended up reinstalling two VM's and restoring a backup. Even with that we had to modify the portal them otherwise we would not get to the Netscaler login page. For a last customer we could not even upgrade as the Native OTP feature is not supported on AHV image (we are kind of stuck here).
👍 1
Open in Slack
PreviousNext
Powered by