On 26/06/2014 22:38, Nick Hilliard wrote:
I've seen several boxes running 15.0(2)SE and rebuilds suddenly die with
memory problems. The box would be fine for many months, then one day the
monitoring system would show that it had suddenly started losing ~1% memory
per hour and then would crash a couple of days later with OOM. I caught
one in the act last week, and "show memory" suggested that the dot1x
process was chewing up fragments all over the place. dot1x was disabled on
that switch.
I've seen similar with 15.0(2)SE on the 2960G but I suspected the SSH
process according to "sh mem", possibly exacerbated by monitoring the
SSH port for availability. When ours died only a power cycle brought the
thing back, as opposed to IOS crashing/rebooting... Unpleasant.
I'm planning a downgrade to 12.2(55)SE9 on two units this weekend. I've
had the code running on a test switch to my satisfaction, I'll attempt
to remember to note any oddities here.
Cheers,
David.
_______________________________________________
cisco-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
