What’s the best step-by-step process to upgrade firmware on a high-availability pair of 1350’s in order to minimize disruption? Since the Ha pair is not stateful (slave unit doesn’t mirror connections of the master), there are going to be connection drops when the master reboots, but there must be some best practices in order to avoid or minimize service outages. What steps do you guys follow for HA pair upgrades?
We suggest to upgrade the slave unit first. Ensure it starts up completely, then upgrade the master one. This will interrupt twice (failover + failback). If the option “Resume Master Role Upon Recovery” is off, then there will be one failover interruption only. Of course you may still want to failback sometime later, and it will still result in a second interruption, but you could schedule this for after business hours.
Is the failover immediate when the active unit reboots, or is there some heartbeat timeout period where the active unit has rebooted and is not handling traffic, and the standby unit is waiting to confirm the primary unit is down, so it’s not handling traffic either?
The failover should be almost immediate, as soon as the standby unit detects the active unit is down it will start routing traffic.
Please reference this post for High Availability (HA) router pair firmware upgrade procedures.