[j-nsp] Juniper ex4200 four member virtual chassis 12.3R3.4 pfem core dumps

[Mark Tees]

The company I work for a has four member ex4200 virtual chassis. Runs v4+v6
OSPFv2 and OSPFv3. Has about 320 RVI's on it. Around 600 routes in total
from different sources.

Everything seems within limits of TCAM etc.

Every now and again (2-3 weeks) sometimes during a commit all four members
of the virtual chassis crash and we get pfem core dumps. Looking at the
core dumps it looks roughly like this vague problem report: PR912889

https://prsearch.juniper.net/InfoCenter/index?page=prcontent&id=PR912889&smlogin=true

In the core dump there was mention of an unexpected null pointer. We moved
from 11.4R5.5 to 12.3R3.4 to try and get away from the problem.

At this stage I am working to break up this virtual chassis down to only
two members and also isolate traffic as much as possible.

Any pointers here would be appreciated. We dont have active J-care on these.

Cheers,

Mark