anyone upgraded to 13-1.63.18 on a HA pair and experience connections failing to re establish when failing over to the freshly upgraded secondary from the not upgraded primary ? i got session reliability on HA failover ticked and it seemed to take quite a while for sessions to re establish and NSGW page was unavailable as well for a min or two. Failing back to original primary once both nodes was upgraded was fine blip for like 1-2 secs and then session re established. Seen issue mentioned before here linen.dev/s/worldofeuc/…/anyone-notice-any-ha-failover-issues-with-the-new-13-1-x-fir

YES, on 14.1 happened again today when doing force failover

had that on 14.1 as well

Do we have a solution?

All our Citrix connections, both applications and desktops, became disconnected when we failed over from the non-upgraded 13.1-59.22 node to the upgraded 13.1-63.18 node. And they never automatically reconnected afaik. Normally, applications and desktop disconnect for just a fraction of a second and then immediately reconnect.

THIS, after upgrade and reboot of the secondary node, force failover does this. AI talks about proc mode on ESXi, are all of you running vCenter? Wonder if this might started to happen after v7 to v8 upgrades?

Had same experience with 2 14.1 HA failovers so far. One on ESX, and one on Azure Local/HyperV. First failover dropped all connections. Second failover once both are on the same firmware is automatic session reconnect as expected. Am trying a 13.1 upgrade tonight and will be interesting to see if it has the same issue, although they use the Secure Access VPN client which should reauth/reconnect to the new NS cleaner that an HDX connection.

@Eric, our production 13.1 VPXes are running on 13.1 SDXes. And it looks like Aaron had this problem^ with Hyper-V. So it is probably an issue with the VPX/NetScaler ADC itself.

Nutanix for me but seen it on ESXI as well

It seems to be "by design" now 🥲, just find this in the 14.1 documentation about the new inline upgrade service: "If both the builds have different HA version numbers, connection failover (even if it is enabled) for existing data connections is not supported. In other words, all existing data connections are lost, which leads to downtime." docs.netscaler.com/en-us/…/issu-high-availability If someone has more information about this...

so what is HA firmware update process now ? Update secondary to latest firmware and reboot it then click the migrate option which syncs the sessions , failover occurs with new connections on updated primary (previous secondary) and exisiting sessions staying on the not updated secondary (previous primary). What you got to do then wait for those sessions to end ? bit crap if you ask me

I'm not really ok with this "new process" (optional) because the main interest of the HA was for me to test the new "updated node" and only if it's ok to upgrade the other node or failover on the other in case of trouble. If old connections are still served on the "old node" we cannot test clearly the new build. For the moment i preferer to keep the "same standard method" (more than 15 years from my side) even if there is a now cut we just have to update the process to communicate on this cut. This cut is also not really a cut because the client seems to try to reconnect and never close even when the timeout has occured. Users needs to reconnect manually so we made a communication before within the sessions through director.