[j-nsp] MS-PIC Sampling and Fragmented Packets

[Stefan Fouant]

Ok folks,

Seeing an issue here w/ an M120 running Junos 9.6R3.8 doing 1:100 sampling
on an MS-PIC.  It appears that the flow records being generated are not
including details about fragmented packets passing through the device.  I
was wondering if anyone knows of any issues in Junos 9.6R3.8 or of any other
systemic issues in Junos code which might prevent the flow records from
properly accounting for fragmented packets.

I'm off to scour PRs to see if there are any known issues, but thought I'd
reach out in the event anyone has seen this issue before.

Thanks in advance,
 
Stefan Fouant