8600 hangs !

[mucka]

Hello Al , any help please we recently had a problem were our main 8600 froze .It is a fully polulated box with many stacks/servers/main frames , and other 8600,s hanging off it.We couldnt telnet or device manage it although we coud ping it , unfortunately my collegue didnt have a console cable so he had to power down & up.It came back up ok but no clues in the log file because of the power down.
No servers or devices could communicate through it.
We are running release 4.0.2.1 , has anyone any idea ? or are there any known issues with this version ?
Any help would be great
Thanks
Mucka

 

[anthonyanderberg]

I haven't seen your specific problem, but I did have several bugs with the 4.0.2.1 release. Most of my problems disapeared after I upgraded to 4.0.4, and 4.0.5 fixed the last one.

Any idea what the CPU utilization looked like when you couldn't connect? Was it not passing traffic or was it just management traffic affected? Do you have R-series or E-Series blades?

 

[mucka]

Thanks for the reply Anthonyanderberg my collegue says the utilisation was not unduly high .It was not passing traffic through to servers etc. We use e series on this box.
What particular bugs did you experience with 4.0.2.1. ?
Thanks again
Mucka

 

[jwmolenaar]

Is there any .000 file on /flash/ or /pcmcia/ ?
If yes, take a look at this file.

 

[mucka]

Yes , thanks for the info the .000 file gave the fol error message at 4am sat morning: error code xxxx can,t malloc pkt memory 32 <np> IP-RIP eroor iprx: out of memory.(Whatever that means!)

This repeated umpteen times ,and we do not have rip enabled !

foll this message there was nothing till 11:50 am when we rebooted,so I assume it was locked up from this time.

Any Ideas ??
thanks again
Mucka

 

[anthonyanderberg]

Malloc errors are never good... it sounds like some process ate up all of your RAM. It would be worth
opening a case with Nortel to make sure there isn't
something odd in your config that is going to cause
the problem to come back.
My main problem with earlier versions of the 4.0.x code
was a nasty ARP table corruption bug we got fixed in
4.0.4.
The other was an OSPF routing table bug that
had to do with multicast protocol issues, that was
fixed in released in 4.0.5.
Neither of the issues caused it to stop forwarding
packets or spit out malloc errors though.
I have R-series blades, and could provide case IDs
and such if anyone has similar issues.

 

[mucka]

Thanks once again guys , around the same time we had a contivity box crash that connects directly into the passport and it took the hard drive out , coincidence ??
I have since been checking the "show log file tail" for clues but no "malloc errors". are there any checks/tests I can do to monitor the situation ?
Big Thanks
mucka

 

[anthonyanderberg]

Are the two systems on the same power circuits?
Brownouts and surges can manifest themselves in all kinds of unfortunate ways. If it happens again let Nortel know, they've got more detailed logging that can be turned on - it can take a lot of CPU cycles so its not normally available.

 

[mucka]

As it happens both 8600 & contivity where on the same power circuits but now we have moved contivity elswhere.
We do not seem to have any issues since so hopefully it was a one off .
Thanks for all reply,s and advice
Mucka