Random reboot of AP

I have noticed that our AP's will randomly reboot them selves. All customers come back online and it doesn't appear to be down for too long but still very weird that they do this.  Why?

James - what band radio?  What software version are you running?  How long is the cable run between the AP and power?  Are you using a CMM4 for power & timing?  Do you have anything plugged into the timing port?  Are there any surge supressors in between the AP and CMM?  Are the AP, Surges Supressor, and CMM grounded?  If so what gauge wire?

We are using 2.4 GHz radios, with software version 13.2. The cable is ran between 100 and approximately 170 feet. We are not using a CMM but instead a CTM-1M. There is nothing plugged into the timing port and there are surge suppressors in place and everything is grounded with 12 gauge wire.

James,

I had this issue with one of my AP as well, We found out that one of our SMs was using the wrong filter setting then we went ahead and sat the filters to the following and it stoped the AP from rebooting. hope this helped. 

we don’t know which filter would have fixed the problem. We didn’t have time to sniff traffic. We only know that one out of 39 SMs did not have its filters set and once that traffic was stopped our problem went away.

No screen shot but the filters are…

SMB

SNMP

Bootp Server

IPv4 Multicast

And the Filter Direction is ‘Upstream’.

No offense, but I'm hoping that isn't it - here's why:

We've been having this happen for a long time, and our standard SM config does apply Protocol Filtering for SMB, BootP Server, and IPv4 Multicast - upstream, naturally. I do not want to block SNMP at the SM, as we use SNMP to get data from certain SMs.

If the only difference between our configurations is that SNMP filter, and you are no longer randomly rebooting, while we still are, what is the mechanism? In most cases, the SM's which we monitor using SNMP are being polled on their own internal public IP, not directly through the AP itself. The reboots do predate such SM monitoring anyway.

Does your AP and/or SM's have public IP addresses? If they do, and you don't do any firewalling at your router, they'll reboot regularly.

1 Like

All our APs use internal-only IP addresses, and none of them (nor any IP addreses within their subnets) are accessible from outside our network at all. They don't even have outgoing internet access, as those subnets are not source-natted on the edge. We assign source-nat rules to each of the  offset subnets for which the DHCP server hands out addresses for the NAT-mode SMs and bridged devices. When we do SNMP monitoring of SMs, we assign them an interface IP address on the same private range as the AP it's connected to.

Blocking SNMP on the uplink will only stop people from using SNMP from the ethernet side.  It won't stop you from using SNMP from the RF side.

2 Likes

dshea, Do you have a support case open? If so, please let me know the case#. I would like to see the CNUT capture from the AP that is rebooting. If not, can you open a case please?

Eric, The crash with public IP address was narrowed down to an incorrectly formatted telnet packet (probably from some bot scanning telnet ports on internet). The next software release will have the fix.

--Rajesh

2 Likes

Rajesh, I have been working on and off with John Mehling on this issue for some time - the case number is SC11792. It has been happeniing to many (but not all) of our APs since at least April of 2014. The issue has been observed across multiple firmware versions, and there are many CNUT captures for the APs on this case (and some for SMs, as well).

Found the information I was looking for from the support case. You were seeing a “watchdog reset” with 13.2. I would like you to try a private beta which includes a fix for that specific problem. Will reach out to you within a couple of days with a build.

I have AP's that do this as well. 430 and 450 AP's drives us nuts. If there is a fix for it, I would like to know. we opened a support ticket a while ago, but weren't really able to follow through with it as work got a bit hectic.

Thanks

Sam

I have absolutely no clue if it was just conincidence or an actual solution but we had an AP that was doing this for a year or so on and off. After some time I noticed that the event log would indicate an NTP time update each time before the AP would reboot. We had the AP pulling NTP from a linux server, after disabling it the AP stopped rebooting. Again I'm not sure if this was just a coincidence.

dshea,

Can you please try the 13.2.2 open beta posted at http://community.cambiumnetworks.com/t5/PMP-Beta/BETA-System-Software-13-2-2-Build-5-is-now-available/m-p/39203#U39203?

David,

The fix I am talking about is unrelated to NTP. Did you have a support case open? Can you provide the case# and I will follow up. I am spoecifically looking for a cnut capture.

--Rajesh