Help understanding ISSU workflow

[tom.barat@dimensiondata.com]

Hello,

I am preparing for upgrades without downtime for Catalyst 9800 series Wireless controllers with the ISSU workflow.

I am specifically wondering how the process works on access points.

If i understand correctly the WLC will upgrade APs in a staggered manner during the ISSU process to maintain connectivity.

According to the configuration guide, we can select the percentage of APs to be upgraded in each iteration, but how does the WLC select which APs are going to be upgraded ?

I am trying to avoid the type of situation where i tell the WLC to upgrade 10% APs at a time and the WLC "randomly" selects all APs in the same floor for one of the iterations, thus interrupting connectivity for that floor.

Is there an "intelligent" AP selection process or is it random ?

Thank you,

Have a nice day,

Best regards.

[balaji.bandi]

This yet to be test myself - in principle i see some good document how this processed.

I  am sure you may seen this document, in case if not for reference  (if that read already ignore my comments)

[tom.barat@dimensiondata.com]

Hello,

Thank you for your answer.

Yes, i am indeed familiar with this document but unless i missed an important section it looks like the AP selection process is not explained.

[Scott Fella · Accepted solution]

I believe that the controller determines the ap neighbors so it understands how to choose which ap's will be rebooted in a staggered fashion.  The controller will also send a deauth to the clients and if the client adheres to the deauth code, will try to move the client to a different ap.  Now this is not 100% because the device has to adhere to the deauth status code.  I have tried this in my lab with around 8 access points, but was more interested in how well it works, with the understanding that clients may drop.  I had the same question and this is how I remembered Cisco telling me when I was part of the EFT.