[c-nsp] Possible memory corruption w/ 12.2(18)S11

David Coulson david at davidcoulson.net
Tue Mar 21 15:04:59 EST 2006


Rodney,

	Just for reference. We also had the following symptoms in addition to
the OSPF errors.

* High CPU load on the RSP (60%, rather than 10%)
* Packet loss of process switched data (e.g. pings to the router, telnet
sessions were flaky)
* Really slow SNMP processing to the point where a snmp walk would throw
out about 20 entries, then stall, then another 20. Scripts to monitor
interfaces would take literally 20mins to run, where previously they
took less than 5s.
* Resets of OSPF from other gear on the network - They'd reset the
session to the router we're seeing problems with.

Mar 20 23:38:53 adsl-gw 18449: Mar 20 23:38:52: %OSPF-5-ADJCHG: Process
10, Nbr 207.166.219.1 on BVI100 from LOADING to FULL, Loading Done
Mar 20 23:49:16 adsl-gw 18458: Mar 20 23:49:15: %OSPF-5-ADJCHG: Process
10, Nbr 207.166.219.1 on BVI100 from LOADING to FULL, Loading Done
Mar 21 00:51:12 adsl-gw 18529: Mar 21 00:51:11: %OSPF-5-ADJCHG: Process
10, Nbr 207.166.219.1 on BVI100 from LOADING to FULL, Loading Done

Uncached IO solved all of these problems (apart from the CPU load issue,
which actually got worse although process switched packets moved quicker)


David




More information about the cisco-nsp mailing list