Hello I will be doing my first upgrade tomorrow after moving up to version 8. We have 2 mobility Masters and 3 sets of cluster 7200 controllers with about 1200 APs. This upgrade is to resolve code bug for meshing 200 series portals. From reading the user guide I can't use the upgrade cluster feature because it doesn't support meshing, and the time could take several hours even at night. I'm going to upgrade the controllers using the method of each controllers non boot partition and reboot. The steps I have are
Pre step backup flash on MM and all controllers record number of APs connected ap's
1. upgrade primary / backup MM reboot wait until they come back up
2. Log into each MD controller load new code on non boot partition
3. Reboot MD on new code partitions
4. verify Ap's are up
The question is pre-loading I tested this in the lab with 2 7000 controllers in a cluster and with 2 AP-225's it took over 15 minutes to preload the image. Should I just forget about pre-loading seemed the old way of upgrading code fast about 30 minute outage once controllers came up until all AP's were upgraded. I read here AP-105s and AP-135s' are not able to pre-load.
2nd question after I upgrade and reboot one MD in the cluster do I wait until it comes back up or do I upgrade and reboot both in the cluster. Do AP's fail over to the other cluster MD or will they also reboot like they did on version 6.x code.
I plan on having TAC involved also but just had a few questions for my CAB planning meeting originally I was going to pre-load but seeing how long that takes with 1200 AP's might not be worth the trouble.