Controllerless Networks

I just upgraded my IAP-103 devices to ArubaInstant_Pegasus_6.5.4.1_61494 and now the VC is rebooting at times. I'm getting "Reboot Reason: Reboot caused by out of memory" and times throughout the day. Any ideas? It looks like the memory on the VC continues to decrease over time until it finally runs out of memory and reboots. There are only 6 clients connected to it when this happens.

 

Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): ip_time_handler: ignore uplink port state, prefer master 1- standalone 0- wifi uplink 0- 3g uplink 0- enforce eth0 0
Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): ip_time_handler: Got ip and packets on bond0 Started master election 0-0, rand 17
Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): (08:00:39) master provision, 1 vs. 0
Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): (08:00:40) master provision, 1 vs. 0
Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): (16:46:58) master provision, 1 vs. 0
Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): (16:46:59) master provision, 1 vs. 0
Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): (16:47:00) master provision, 1 vs. 0
Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): (16:47:01) master provision, 1 vs. 0
Oct 2 11:47:02 KERNEL(ldc-wifi1@192.168.3.11): (16:47:02) master provision, 1 vs. 0
Oct 2 11:47:02 syslog: <341005> <ERRS> |AP ldc-wifi1: | open core directory failed
Oct 2 11:47:02 syslog: <341005> <ERRS> |AP ldc-wifi1: | terminal access is enabled but sshd is not running, starting
Oct 2 11:47:03 KERNEL(ldc-wifi1@192.168.3.11): (16:47:03) master provision, 1 vs. 0
Oct 2 11:47:04 KERNEL(ldc-wifi1@192.168.3.11): (16:47:04) master provision, 1 vs. 0
Oct 2 11:47:05 nanny[2131]: <303022> <WARN> |AP ldc-wifi1: | Reboot Reason: Reboot caused by out of memory

 

Gregg

hi Gregg

Log a tac case for sure, you can help speed up the resolution by collecting some pertinant info such as:

1. does the leak increase linearly with time ?

2. does the leak only happen when clients are connected ?

3. does it happen on the VC or non VC or all ?

 

Also collect the tech-support dump (ensuring to target an affected AP) over say 3 fixed intervals (if the AP reboots within 6 hours, then collect the tech-support dump 3 x 30minutes, but not more frequently) and send that to TAC - it will give the R&D people something to go on.

 

Most likely with a leak as bad as you describe that it's a regression bug and someone will have probably already found the whoopsie and fixed it, but could also be a sizing issue where someone forgot the smaller ap-103 so may take a TAC case to get R&D onto it.

 

and then there's always rollback :)

 

 

I have the same problem

 

when someone connects about 20 or more clients, the AP immediately restarts

 

how to solve this problem? Please help me

The solution depends on the version of instant you are running. You should consider upgrading, because this has been solved for several versions.