vsx_util change_private_net

[KFR]

About to extend the private IP network (real IP / Funny IP) on a 2 node VSX cluster (25vs') to support 128 interfaces instead of default 64.
This is a production cluster so need to understand the downtime and impact doing this.
In SK99121 downtime is stated to happen in step 13, however a reboot is required in step 17.

Questions:
How much downtime can be expected in step 13 (seconds, minuttes?)?
Will sync (state table) remain or be lost ?
Whats the impact until after step 17 ?

Thanks!

[the_rock]

Hey @KFR 

I will give you answers based on one customer that did this recently, but take it for what its worth, as they say.

Andy

Questions:
How much downtime can be expected in step 13 (seconds, minuttes?)? About 15 minutes

Will sync (state table) remain or be lost ? It was intact

Whats the impact until after step 17 ? 

They followed below as per sk and it was very minimal.

To minimize an impact on the traffic:

• In the VSX High Availability Cluster (not VSLS), reboot the Standby cluster members.
  Then run the "cphastop" command on the Active cluster member and reboot it.

• In the VSX VSLS Cluster, it does not matter in which order you reboot the cluster members.

[the_rock]

Sorry, forgot to add, TAC did tell them downtime timeframe also hugely depends on number of VSs and also subnet size you are "extending"

Andy