topic Re: AP Reboot: errno 2 in ZoneDirector

AP Reboot: errno 2

drew_baldock — Thu, 30 Jan 2014 21:38:58 GMT

We have a few APs that reboot with a 'heartbeat' lost message, although sometimes it's 'application reboot' in the web interface (of the ZD).

There's nothing obvious in the AP supportinfo file other than it lists reboot_reason: user reboot instead of the above reasons.

This is what the syslog output shows at the reboot time:

Jan 30 15:15:59 wireless4 stamgr: stamgr_AP_xml_house_keeping():complete updating ap cfg to emfd
Jan 30 15:16:00 wireless4 syslog: Request():IPC connection close. errno 2
Jan 30 15:16:00 wireless4 syslog: Request():Deplete everything in the pipe... TODO
Jan 30 15:16:00 wireless4 syslog: main():handleRequest() failed or done...^M
Jan 30 15:16:10 wireless4 stamgr: stamgr_init_AP():AP c0:8a:de:26:10:60 radio(6) default channelset_enable=1 channelset=1057 Line:1421
Jan 30 15:16:10 wireless4 stamgr: stamgr_init_AP():AP c0:8a:de:26:10:60 radio(5) default channelset_enable=0 channelset=0 Line:1421
Jan 30 15:16:19 wireless4 stamgr: stamgr_AP_xml_house_keeping():complete updating ap cfg to emfd

Any insight as to what these syslog messages mean? Specifically what causes errno 2?

I had a support call open a few months back but couldn't replicate it reliably and it disappears most of the time when we move one of the rebooting APs to a new drop. It's almost like some APs are more picky about locations (power dips maybe?), because another AP moved to the same drop will be fine.

I should add that I've tried fw force-update on the rebooting APs with no luck.

Re: AP Reboot: errno 2

jeff_williams_6 — Thu, 30 Oct 2014 13:34:04 GMT

I get this from time to time and it's usually preceeded by the network switches disabling the ports that a few of the APs are on- in essence rebooting them.

I'd love to have a resolution to this.

The switches are cutting off the APs because STP is detecting traffic storms from the APs. After the ports recover and are reactivated things seem ok.

Re: AP Reboot: errno 2

michael_brado — Thu, 30 Oct 2014 22:09:04 GMT

I can't tell from that snip what is occurring, but I can advise that if you have 64mb RAM APs in use (7363/7962/7762) that you should segregate your ZD/APs from client traffic VLANs, with what we'd call the management VLAN.

Client VLANs often contain lots of broadcast/multicast traffic, and APs have to
inspect all packets to determine client traffic from the other bcast/mcast. We have
seen IPv6 floods having an impact and causing APs to lose heartbeats with ZD
too. Filter IPv6 at your VLAN router/switches if not used.

Avoid the use of TKIP/Auto settings in PSK WLANs, and use WPA2/AES combo
only. Otherwise, if you continue to encounter application reboots, open a ticket
with Tech Support for deeper investigation.