MSM765 5.7.0.3-11516 AP's reboot then will not re-connect to controller

[Rick McLaughlin]

I have an MSM765 controller with about 90 AP's (MSM460 and MSM466) and after the upgrade to the latest software some of the AP's will start rebooting every three minutes.  After some time the AP's give up and show as disconnected on the MSM765.  If you go to the POE switches (HP 2910-24al and 2910-48al) the AP's are connected to and turn off poe (no power-over-ethernet, and disable the interface) then enable the interface and turn on poe (ena and power-over-ethernet) the AP will come back up and run for a random period of time.  Weekends have less AP's rebooting which points me towards maybe a client causing the AP's to reboot.  I have read about some problems with some Apple products causing similar issues and wonder if anyone is experiencing anything similar?

[Fredrik Lönnman]

I've been recommended by TAC to stay with 5.5.3 until 5.7.1 is released in about a month. Just a thought.

[Evan_ISS]

---

@Rick McLaughlin wrote:

I have an MSM765 controller with about 90 AP's (MSM460 and MSM466) and after the upgrade to the latest software some of the AP's will start rebooting every three minutes.  ...

---

We had the same issue for some time specifically with the MSM460s but also with MSM410s.

It was more pronounced after upgrading to software version 5.7.0.3 but it was going on from the first day we introduced MSM460s to the environment.

The issue has been multifold and only solved today (03/10/2012) when we upgraded to software version 5.7.1 (not yet available publicly, only via support) and PCMv4 AU3 (publicly available).

For us it was a combination of an inherent issue with the MSM460s (I had a face to face conversation with a L3 support engineer in London during a convention and he described it as a chip firmware issue specific to the MSM460, fixed in 5.7.1) and a dodgy upgrade (AU2) to PCM+v4/PMMv4.

If you use PCM with PMM AU2 and try to use SNMPv3 with your controller, the controller (and APs) are flooded with erroneous SNMP requests causing multiple snmp failures and sflow driver errors causing kernel panic and subsequent restarts.

Downgrading to SNMPv2 or upgrading to PCM AU3 solves the issue.

If I was you I would get my hands on 5.7.1 via support and go from there and if using PCM follow the aforementioned solution(s).

Hope the above helps.

[Rick McLaughlin · Accepted solution]

Hi,

Thanks for the replies.  I am using PCM 4.0 AU2, but nor SNMP v3.  We need the functionality of the Captive portal re-authentication feature that does not quite work as desired in the previous version (Welcome Back Screen is not working for us) so I will be upgrading to 5.7.1.0 and hoping for the best.  Will let you know how the upgrade goes.

Thanks,

Rick

[Rick McLaughlin]

Looks like the upgrade has taken care of the random re-boots.  Thanks for all your help!