Wired Intelligent Edge

Scenario: Rolling upgrade from ArubaOS-CX 10.01 (say 10.01.0030) to 10.02 (say 10.020010).

Two doubts reading ArubaOS-CX 10.02.0010 in preparation to a scheduled upgrade:

Why the no keepalive command is needed to be executed on both VSX members as a preparatory step before booting new software version? ...is it still necessary if only VSX LAGs are used to downlink hosts/devices?

How can the ISL Keepalive be established automatically again after the upgrade (as written on ArubaOS-CX VSX Guide for 10.02: "Keepalive is restored later when the VSX configuration is applied" Step 3 page 8) if it is then requested to save configuration after the issue of no keepalive command (Step 5 page 9)? ...there isn't a following step on the described rolling upgrade procedure where the keepalive command is required to be reapplied to reestabilish the ISL Keepalive again.

Am I wrong? Am I misunderstanding listed rolling upgrade preliminary steps?

Side note: just an example, I really don't understand the "VSX LAG secondary switch" terminology used in VSX Guide...shouldn't it simply be "VSX secondary switch"?

Hi Davide,

I suspect that this was kept from requirement of first version 10.0.

I performed many VSX upgrades without removing keepalive and it does not impact failover time. I'll check internally about this sepcific point in the release note.

Hi Vincent, I did that too on every update performed since first ArubaOS-CX 10.01.0001 of July 2018 (once I first deployed our VSX)...without any issue!

That's why reading about the no keepalive step was a little bit worrying.

Interestingly ArubaOS-CX 10.02.0001 or ArubaOS-CX 10.02.0010 Release Notes didn't mention about the no keepalive step.

I got confirmation that this point was specific for 10.0 to 10.1 migration. (no mclag keepalive). The documentation will reflect this correction soon. Thanks !

That's good. Thank you Vincent.