Virtual Switch on 10K switches

[Jackle_08]

Thanks for information.

[jcbelmonte]

Any other comment on Virtual chassis ? We are going to upgrade our 4x 10K from multi-chassis to virtual-chassis.
We've got repetitive bugs on the L2 linkagg between each pair of 10k... with some black out...

Now we hope that Virtual chassis will stabilize our core network, with the 7.3.2.344...

Did you set up RCD?
Did you have specific pbs on L2 ? L3 ?
Do you use static linkagg / dynamic ?
Do you use VRF ?

Thanks for your help.

[devnull]

I had problems with MC-LAG (on 6900) and i have a setup utilizing VC on OS10k, so far *knocking on wood* without any issues i am aware of.
I had in the past problems with flapping LACP links on 10K which were either 6850-AOS or SFP related.


Did you set up RCD?
Yes.

Did you have specific pbs on L2 ? L3 ?
Do not understand.

Do you use static linkagg / dynamic ?
dynamic linkagg

Do you use VRF ?
no

I would NOT use 7.3.2.384 but the latest maintenace release 7.3.2.548

As always: plan a maintanence window, have a disaster recovery in place (e.g. create a new vc-dir, create all needed configs/files in the new directory "reload from vc-dir no rollback-timeout" have patience, it will take about 4-12 Minutes for the VC to form, whilst the console seems to be stuck (e.g. up to 3 minutes after it initializes the BCM)
if possible do a "dry-run/test" so you get a feeling for what should happen.

If it fails do a reload from MClag-dir to get back to your know state.

I must admit: i am unsure how the update will work with dual CMM.

[ydeschoe]

JCBelmonte

answers to your questions

Did you set up RCD?
Yes.

Did you have specific pbs on L2 ? L3 ?
currently we are running the 10Ks as the default gateways for our server VLAN's and no problem seem, OSPF is also configured

Do you use static linkagg / dynamic ?
dynamic and static linkagg are in use

Do you use VRF ?
no

The way we have done our upgrade is the following

We removed one chassis out of the MC-LAG setup and configured it as a VC device all interface in status admin down, ones the VC was ready we brought down all the interfaces onto the still existing MC-LAG device and enable in the same time the newly upgraded VC his interfaces, this has generated a downtime of about 5 sec, time to bring up all the links and linkaggs configured, end user didn't noticed anything, was done during normal business hours.
We left the 2nd chassis for the coming 30 min still in the MC-LAG config (all interfaced down) for the case a problem would pop up (quick fail-back)
After 30 min we migrated the 2nd MC-LAG chassis to VC and brought it together with the already migrated chassis as slave chassis, not downtime seem with the process (boot.cfg and vc-boot.cfg already the same files onto the 2 systems), and all links and LACP's came up without any problem on the 2 chassis (1 link on 1 switch 2nd link on the 2nd switch).
We did also the upgrade process by creating 2 extra dir on the switch (old MC-setup, normal switch with all interfaces down and no config in it (this gave us the time to configure the emp interfaces and setting up the lic for the VC setup, dir with the VC setup).
This is now running for more than 2 months without any failure/problem, system is stable.

Software version we run is 7.3.2.439.R01

Yves