cancel
Showing results for 
Search instead for 
Did you mean: 
Post a Question
Ugur_Urel
Ivory

Jumbo Hotfix Installation Procedure on Full HA R77.30

Hello all,

We have two 5200 appliances running full ha(mgmt and gateway roles installed on both appliances). What is the correct procedure for installing jumbo hotfixes? I couldn't find any documentation. Is it possible to first install hotfix on standby node and then switch this node to active and test the functionality or should I start with the active node? And also what should be the recovery plan of this procedure? Taking snapshots for both nodes sufficient?

0 Kudos
2 Replies
Employee+
Employee+

Re: Jumbo Hotfix Installation Procedure on Full HA R77.30

I have always been overly cautious when I do maintenance. I take a backup of the Gaia config. A backup from within Gaia, and a snapshot too. Just make sure they are all copied off box in the event of a complete failure. I/E: hard drive. It has never happened to me, but I have always been more safe than sorry.

So as for the procedure on upgrading. I generally recommend doing the maintenance on the standby member first. Leverage CPUSE to do the install for you. The process is very straight forward and clean. Use sk92449 to make sure you have the latest build of CPUSE. Really early versions of R77.30 didn't always upgrade on their own. After the JHF has installed, the gateway will reboot. You can manually fail the cluster member over and test inbound/outbound connectivity. If everything tests ok. Repeat the process on the other (now standby) member. 

Re: Jumbo Hotfix Installation Procedure on Full HA R77.30

There is a documantation about Jumbo HF in sk106162 - Jumbo Hotfix Accumulator for R77.30 (R77_30_jumbo_hf), see e.g. cap. 3 When to Backup and Restore your Gaia OS machine? and Can you install the Jumbo Hotfix Accumulator on cluster members while preserving the current connect...

  • Yes, As CCP version does not change, sync is on after installing the Jumbo Hotfix Accumulator on the first member and rebooting it (Standby member in case of cluster HA), and the connections are preserved.
0 Kudos