04-01-2013 07:53 AM
Is it possible to update the software of the 8212zl switch with dual management modules without any downtime. We do have the - redundancy management-module nonstop-switching - enabled.
I can't seem to find any documentation online that supports updating the software without rebooting the switch or causing any downtime. The only document I could find was from 2009 this one:
If I follow the procedures in this document with the nonstop-switching option enabled will I experience any downtime?
Solved! Go to Solution.
04-02-2013 04:37 AM
Ok.. I have an answer for myself. The answer is NO. Updating software with dual management modules cannot be performed with out a reboot... i.e loss of service. Here is why:
My switch is running in "non stop switching" mode
When you update the active primary flash with new software the changes are replicated to the standby. You are then instructed to boot the standby with "boot standby". When you reboot the standby module the software version becomes active... great, right? Well now because "non stop switching" requires both management modules to be on the same version you are informed that you are no longer operating in "non stop switching" mode but have been set to "warm standby" because ... since the standby reboot you now have different software versions running on your two management modules.
According to the manual performing a "redundancy failover" at this point will result in loss of service because the switch is operating in "warm standby" mode.
So from this I am guessing that there is no way to perform a "hitless" upgrade of software for the 8212. If anyone knows a way around this please let me know.
04-02-2013 01:15 PM
it is quite difficult to perform an in-box hitless firmware update, so the procurve/provision 8200 is no exception on this. At this moment (as you have found already), there is only a limited downtime, not a hitless update. This means you still need an additional network host/protocol (like xstp or vrrp) to handle the failover to another box during the update.
The update will be completed quite fast, but it will be too long for any tcp session to be terminated, so an additional box/unit is really required for the quick failover,