VSM Primary/Secondary problems

cajalat · ‎05-26-2010

I have a VSM that I originally installed as the primary VSM (via the GUI method). I wanted to install the secondary but they never seem to sync. I saw this trhread:

https://communities.cisco.com/message/33166#33166

And I noticed that my current VSM (the one I configured as primary) was in a state of standalone. The thread pointed out that it shouldn't reboot if configured for primary however, the VSM immediately reboots upon issue of the commenad "system red role pri" and doesn't allow me to save the config. So when it comes up again it becomes standalone again.

I'm currently running 4.0(4)SV1(3)

Here's the state on the "primary"

nexus-2# sh sys red st
Redundancy role
---------------
      administrative:   standalone
         operational:   standalone

Redundancy mode
---------------
      administrative:   HA
         operational:   None

This supervisor (sup-1)
-----------------------
    Redundancy state:   Active
    Supervisor state:   Active
      Internal state:   Active with no standby

Other supervisor (sup-2)
------------------------
    Redundancy state:   Not present
nexus-2#

And here's the status on the secondary:

switch# sh sys red sta
Redundancy role
---------------
      administrative:   secondary
         operational:   secondary

Redundancy mode
---------------
      administrative:   HA
         operational:   None

This supervisor (sup-2)
-----------------------
    Redundancy state:   Active
    Supervisor state:   Active
      Internal state:   Active with no standby

Other supervisor (sup-1)
------------------------
    Redundancy state:   Not present
switch#

Any help/pointers greatly appreciated.

Casey

lwatta · ‎05-26-2010

Casey,

I've seen this. I beleive it is because the secondary is up. Shutdown the secondary VSM and see if you can change the role of the standalone to primary.

I think once the secondary is down you will be able to change the role without it rebooting. Once the role takes then power up the secondary.

I'll play with it here in my lab to confirm.

louis

cajalat · ‎05-26-2010

You hit the nail on the head. That solved it for me

Thanks.

lwatta · ‎05-26-2010

Good. I'm glad that fixed it and I also confirmed in my lab.

I beleive what is happening is that when the standy reboots after running setup it comes up and if it can't find the primary it comes up as HA status of degraded mode. When you change the role of the standalone to primary it instantly starts advertising on the control network for a secondary. As soon as the primary and secondary see each other the secondary reboots the primary because it's last known state for HA is "degraded". I know that seems odd but thats the way HA prevents split brain since we don't have a quorum device.

As you have seen the primary reboots right after you change the role not giving you time to copy run start so it comes back up in standalone mode

louis

cajalat · ‎05-28-2010

Thank you for the explanation. That is consistent with what I saw. The only downside for me was that since the Secondary never received the correct config and upon the reboot of the "primary" the config was lost. The Secondary did in fact start to see all the ESX Servers in the cluster (but no configs of course). If I reloaded the secondary at this point then the "primary" takes control of the ESX hosts and resumes using its original config.

Casey