[c-nsp] Const2 IPv6 Proc

Steve Lalonde steve at enta.net
Thu Jun 2 10:42:27 EDT 2011


Hi

I am trying to debug several random crashes we have experienced on our 7600 based network. 
I have narrowed it down to a process on the SP "Const2 IPv6 Proc". It grows in size over time, Starting at around 8M currently the largest I have on the network is 388M. Eventualy the SP runs out of memory and triggers a reload.

after 1 day
PID TTY  Allocated      Freed    Holding    Getbufs    Retbufs Process
408   0    8641644    5147540    8188404          0          0 Const2 IPv6 Proc

after 26 weeks
PID TTY  Allocated      Freed    Holding    Getbufs    Retbufs Process
383   0  411086712  319278684  388950676          0          0 Const2 IPv6 Proc


I have tried multiple IOS versions from SRD4 up to SRE4 and all have the same problem, the only reference i can find on CCO or google are unrelated articles containing "show proc mem" output.

Network is configured as MPLS-TE 6PE with FRR link protection, OSPF as igp and customer routes in iBGP. The "Const2 IPv6 Proc" grows in steps with events on the network, reloading a router causes  multiple routers to use a chunk more memory as does a core circuit failure (fibre cut).

This affects some routers much more than others and does not seem to be load related. 

Does anyone know what this process is/does and can it be restarted without a reload.


Steve


More information about the cisco-nsp mailing list