VCS Maintenance mode (x7.2 cluster)

[Martin Koch]

Hello!

I have a cluster pair of two VCS-E X7.2.

Just wonder if this is a bug or the wanted behavior (at least the xstatus seems to be a bug):

If I set cluster peer 1 to maintenance mode (via tsh:  xConfiguration SystemUnit Maintenance Mode: on)

peer 2 also goes down.

I see the alert on the first vcs, the calls and registrations are active till they expire.

An alarm is shown on both that the first one is in maintenance mode.

If I acknowledge all alerts on the second one, suddenly it also shows that the second one is in maintenance mode.

(same date and time but it did not show up before).

Interesting enough when I check on the second VCS in the tsh about the state:

xConfiguration SystemUnit Maintenance Mode

*c xConfiguration SystemUnit Maintenance Mode: Off

It still shows it as off, ...

Regstrations and calls are not accepted by both.

If i then type on the secon one:

xConfiguration SystemUnit Maintenance Mode: Off

systems can register on peer2 ...

My expected behavior would be that the maintenance mode is a peer specific setting so I can remove one peer gracefully.

Also that the mode itself is reported proper in xstatus.

[Michael Boscia]

Martin,

I tried upgrading to x7.2 and it was increadibly destructive/disruptive to my environment.

If you're going to try it, make sure you have console access to your VCS-E, as I lost one, then both of the network interfaces on the device until I deleted and re-added the dual NIC license key.

I don't know why, but it also broke my AD integration on my VCS-C, so watch out.

Oh, and TMSPE broke on both devices.

I was only able to get everything working by downgrading back to x7.1

YMMV.

[Martin Koch]

Hi Michael!

Thank you very much for your feedback / caveat +5 for that!

Yes, I was skeptical and I am happy that this is a lab setup.

I did not had any problems there, the upgrade went fine, also with my VCS expressway

starter pack at home with dial interface option, though I only use it only for nat and not the

secondary interface. What were the symptoms you faced?

I can recommend having a serial terminal connected to the VCS so you can reach it if

something goes wrong with the network config.

The authentication part is also what worried me most, why I need to do some more testing

in the Lab. Not so sure if its a golden combination with TMSPE which forces the accounts

as local user accounts to the VCS and the local database always active if you use LDAP.

Maybe its something which you faced yourself.

Doing a backup upfront so you can revert to the older version is highly recommended.

Had to google what YMMV is ;-)