How to upgrade a stack and then reboot the members one by one?

[cisco]

Hi,

I have 2 Catalyst 2960S-switches configured in a stack. The uplink is configured as a cross-stack etherchannel. Now I want to upgrade the stack with a minimum of downtime for the connected hosts. The hosts are connected to the stack via cross-stack etherchannels. So my question is: How can I upgrade the stack and reboot the members with a minimum of downtime? If I use the archive download-sw command with the /reload-option, the stack is reloaded and the result will be downtime for the hosts while th stack is reloading. Is it possible to upgrade the stack and then reboot the members one by one in a controlled manner?

Best regards,

Thor-Egil

[Leo Laohoo]

Upgrade the iOS using the command "archive download-SW".

Once the IOS is installed and MD5 hash checked, reboot each member using the command "reload module".

Sent from Cisco Technical Support Nintendo App

[cisco]

Hi and thanks for your answer. I cannot see that the "reload module"-command is available, do you mean "reload slot "?

[Leo Laohoo]

Yes "reload slot <1-4>". 

[Reza Sharifi · Accepted solution]

Hi,

I don't think there is way to upgrade one switch at a time and reboot them and still keep the stack function correctly.  If you upgrade one member and reboot, now when that member comes back on line it has a different IOS version then the rest of your stack and that is usually not good.

"archive download-sw [options] ."

The default behavior is to upgrade all members.

The options for the `archive download-sw' command allow for detailed control of the upgrade. Use the options to update only a single member instead of the entire stack or to reload the stack after the upgrade is successful instead of having the reboot be an extra step for the network admin.

http://www.cisco.com/en/US/prod/collateral/switches/ps5718/ps6406/white_paper_c11-578928.html

HTH

[cisco]

I want to upgrade all members, but I want to control the reload-phase so that the connected hosts do not loose connectivity during the reload.

[Reza Sharifi]

The problem is if you reload one device at a time, you will have a different IOS on the device that comes back online and that cause the stack to not function correctly.  I my opinion, if you have an outage window after hours and reboot all at the same type, you are dealing will less headache.

HTH

[Joseph W. Doherty]

Disclaimer

The  Author of this posting offers the information contained within this  posting without consideration and with the reader's understanding that  there's no implied or expressed suitability or fitness for any purpose.  Information provided is for informational purposes only and should not  be construed as rendering professional advice of any kind. Usage of this  posting's information is solely at reader's own risk.

Liability Disclaimer

In  no event shall Author be liable for any damages whatsoever (including,  without limitation, damages for loss of use, data or profit) arising out  of the use or inability to use the posting's information even if Author  has been advised of the possibility of such damage.

Posting

Connected hosts are going to take a hit when the device they connect to reloads, either individually or as the whole stack reloads.  The only advantage to what you propose not all stack connected hosts would drop at the same time, instead you would have a rolling outage.

If the stack management part of the software doesn't differ in version, or at least major version (?), it's possible to run a stack with different IOS versions, but personally, like Reza, I would rather reload the whole stack.

[krahmani323]

Hello Thor,

In addition to all previous posts, the ideal feature in your scenario would have been the RSU (Rolling Stack upgrade) available since 12.2(58)SE.

Unfortunately it is only supported on 3750-E & X models (not on 2960-S) as per the doc below:

http://www.cisco.com/en/US/prod/collateral/switches/ps5718/ps10745/product_bulletin_c25-658743_ps6406_Products_Bulletin.html

Cisco Rolling Stack Upgrade (RSU) on Cisco Catalyst 3750-E and 3750-X: With previous releases, whenever Cisco IOS Software was upgraded on a stack of switches, a complete stack reload was required, which resulted in hosts losing network connectivity for several minutes.

Starting from this release [12.2(58)SE], in networks that use redundant uplinks and downlinks to the hosts, the intrinsic redundancy of the network can be used to perform a stack upgrade one member at a time with minimal traffic disruption. In this case, hosts with connectivity (either wireless or wired) to two different switches in the same stack would experience minimal traffic loss when switching from one uplink path to the redundant one while the rolling upgrade is being performed.

Regards.

Karim