[c-nsp] Parity Errors and Cosmic Rays

[Bill Wichers]

> Back in February, I had a VIP freak out in my 7513 due to a memory
parity error. TAC said that it appeared to be an isolated event and not
due to an actual hardware problem. It happened again in April and TAC
still said that it was most likely cosmic rays or something and not an
actual hardware problem. However, we replaced the hardware anyway just
to be safe.

I remember reading somewhere once that a neutrino hitting a DRAM cell will
release just enough energy to flip the state of one bit...

> Is this actually a common problem? Or at least common enough that I
should expect to see it every other month or so? It seems strange that
this router has run for years and we've never seen a memory parity error
and now we've seen three in three months.
>
> Any thoughts?

You are either incredibly unlucky or there is something else on... We run
two 75xx routers (both in shielded facilities though), and have never had
a problem with either of them on either the RSPs or the VIPs. I have seen
parity errors in switches that are near (as in on the other side of a
shared wall) power rooms and I think the EMI from the transformers and
electrical panels may be to blame. This is especially the case if you have
any primary-side power (kilovolts from the utility) near your equipment.
Static damage has also caused problems like this since static damage
usually manifests itself as unreliability or shortened lifetimes rather
than just plain dead parts.

See if you have any high-power devices nearby, and if not try replacing
the RAM and see if that solves the problem.

     -Bill

*****************************
Waveform Technology
UNIX Systems Administrator




*****************************
Waveform Technology
UNIX Systems Administrator