10.2r4 firmware causing AP650's to lose connectivity?

[KevinP1]

Okay, I don’t know what’s happening here but this isn’t looking pretty. All of our AP650’s are losing connectivity multiple times throughout the day. 

Look at this! Some AP’s will only do this once or twice and others are doing it often, like this ? There’s nothing indicating any type of error or issue. We didn’t have this problem on firmware 10.09rb.

I have a case open with GTAC, but are we the only ones seeing this issue?

[w1f1n00b]

Are any of you seeing definitive evidence of client disconnect or loss of network during the graph memory dip? Is it possibly only affecting some clients and not all?

On the AP I’m monitoring I just had another memory dip in the 8 hr graph (3rd one today) and I have a windows laptop connected and pinging 8.8.8.8 w/ logging. I’m also monitoring roam events locally on the device. I had a graph dip at 12:30 and it didn’t effect the client at all as far as I can tell.

What are the chances that this is just a UI bug and not an AP/FW issue?

[a_huerzeler]

@kevin.piazza I sent Sam my debugs of my test lab couple of hours ago.. Lets hope we get results soon!

[w1f1n00b]

Unfortunately 99% of my AP Models are not supported by the 10.2 FW family. I still haven’t had time to test 10.0r10a. So for now rolling back seemed like the best option.

[KevinP1]

@javabomberman The list of inconsistencies between each firmware release, and XIQ releases are making us feel like we’re watching Tom and Jerry chase each other.

 Oh, are any of you using 802.11K, V, and R on your SSID’s? I was told by GTAC to disable them. I had been using these for months without any issues because Apple devices utilize these protocols. I turned them off last night, and I’m already having calls about clients being dropped.

[Hans_SD44]

So again, I updated to 10.2r4 ~a month ago, and we still haven’t had any reports of disconnects or dropped devices at any sites. Comparing this to the 2 days of issues at all sites when we went from 10.0r9b to 10.0r10, I am skeptical that the reporting of memory dumps is actually causing the same issues that it was previously for us.

Just another fun bit for this whole scenario.