6509 SUP UPGRADE - Can I reduce downtime ??

dclee · ‎03-08-2007

Alright, I will be upgrading my production 6509 SUP1A's to SUP2's. What is the best possible method to reduce downtime ?

- Is there any scenario where I can eliminate downtime and ensure that all servers switching on modules 6,7,8 remain connected. This would be invaluable between our APP and DB servers.

- Can I install one of the SUP2's into slot 2 and then perform a switch supervisor ? How does the automatic sync work b/c the SUP2 is running a newer rev.

Any help would be appreciated.

Cheers

Dave

steve.busby · ‎03-08-2007

The redundant supervisor engines must be of the same type with the same model feature card.

http://www.cisco.com/en/US/products/hw/switches/ps708/products_configuration_guide_chapter09186a00803f586a.html#wp1079325

In your case, if you have spare chassis, I would take the configuration from your Sup1A and load that configuration on your Lab 65XX Chassis with the Sup2, then you can simply power down the Sup1A chassis and replace it's sup with the newly configured Sup2.

If you have no spare Chassis, then your options become a bit more limited. You can copy your current startup configuration to a flash card and once you powerdown and swap supervisors, then you could copy the backed up configuration from the flash card to the switch. Or have the configuration backed up to a TFTP server, swap out the supervisors, get minimal IP connectivity on the newly installed Sup2, then TFTP the configuration on to the Sup2.

But unfortunately, to answer your subject, there really isn't a painless way to upgrade your Supervisors. Either way will entail some downtime, your total outage time will depend on how quickly you can get IP communications up and your old configuration back in place.

HTH

Steve

dclee · ‎03-08-2007

Thanks Steve, the upgrade process itself is fairly easy for me b/c I do have the other 6509 up and running in the lab with the production config. However it seems our application servers to oracle DB connections are so "finnicky" that even the 10 minutes of downtime during the SUP swap is enough to cause issues. Which means about 10 minutes of work for me but about 6 hours of work for our DB team b/c of the graceful shutdown and startup process. At least that is what they are telling me. So I was hoping to make it a little easier for those guys..

So be it..Thanks again.

Cheers

Dave

steve.busby · ‎03-08-2007

Understand, our DB folks have the same complaint. Seems it would save time if they do a graceful shutdown before you commence your switchover instead of after.

Steve

dclee · ‎03-08-2007

No doubt, and that will be the gameplan. The downside is b/c we have so many environments between QA/DEV and production as well as Peoplesoft and Oracle that it takes about 2 - 3 hours for shutdown and about 3-4 hours to bring it all back up..For literally about 10 - 15 minutes of Cisco upgrade time...

Which is why I was looking for a work around..

Thanks again.

Dave