vcs disable - split vchassis in 2 standalone - will result in outage?
fischerdouglas
Member ✭
in System
We have a pair of thunder 3040 used for CGNAT in vChassis mode.
We need to split it in two Standalone nodes.
We will do it using the command: "vcs disable"
A colleague mentioned that no disruption will occur.
But I would like to double-check it.
I looked into "ACOS 4.1.4-GR1-P5 Configuring ACOS Virtual Chassis Systems" pdf...
In there it is explicit that a reload is needed to add a node to a vChassis, but is not explicit if it is needed to reload to disable aVCS.
Any suggestions?
Tagged:
1
Comments
I think this will depend a bit more on the VRRP-A config and routing, are they active active or active/standby?
VCS should just apply to config management between the devices AFAIK. I.E upgrading you may want do disable it so if you upgrade the secondary - it doesn't get immediately downgraded to the primary code.
A crude method to avoid an active takeover, you could disable all the data plane ports on the secondary/standby device, so no traffic will flow to it.
Unfortunately, I'm not able to deviate the active traffic from one box to other...
So, it's not an option by now.
This VCS scenario is not in use to redundancy.
Just for config-sync purposes.
Each node is active on his own services.
So, in theory everything is independent from the other box.
My question is more:
"Does the ACOS will need to reboot or reload because it is leaving behind the Virtual Chassis?"
Hi Fisherdouglas,
if you do vcs disable on the standby appliance the active vmaster appliance will not reload, only the vblade Will SW relax when it’s added back into the VCS cluster. If your not moving the traffic from vmaster to vblade you should be OK.
thanks Eddie Bellamy
SE A10 UK/IRE.
Typo sorry “only the vblade will SW reload” not relax :) as my text above.
This morning we executed the "vcs disable" on two nodes of the vChassis.
ZERO problems!
ZERO downtime related to that.
Thank you all!
Great news thanks for the updates.
eddie