cpu 100% and network unreachable on SCO 5.0.5

[andtony]

Hi,

I have just enough UNIX knowledge to do a few things on our SCO 5.0.5 server, but I'm stuck with a tough problem :

Some times, for no particular reason, the server is very very slow. I run sar and see that the %sys column is around 99% ... 3 or 4 minutes later, all clients are disconnected from the server, which we can't ping anymore.

I ran this command : ps -A -o "pid pcpu ruser args" to see which process was eating the cpu, and I saw that the process called "vhand" was eating about 90% of the cpu !!

Is it a well-known problem ? The server is running for years and we didn't change anything unless the activation of SMB Client for VisionFS (and it was a year ago !)

Any information would be helpful ... Thanks in advance !

 

[patw]

vhand is the memory allocation daemon - something else is reducing available memory and vhand is trying to steal back some pages. Check this SCO URL out:

http://osr507doc.sco.com/en/PERFORM/ident_mem_bound.html

It's for 5.0.7, but the various tests are the same for 5.0.5. A possibility - hardware failure in either the system board or the installed RAM cards is suddenly causing the system to be starved for memory.

Another possibility is a recent program change is buggy and causing a memory leak until it starts swapping pages to disk like crazy.

----------------------------------------------------------------
Pat Welch, UBB Computer Services, a WCS Affiliate
SCO Authorized Partner
Microlite BackupEdge Certified Reseller
Unix/Linux/Windows/Hardware Sales/Support
(209) 745-1401 Cell: (209) 251-9120
E-mail: patubb@inreach.com

 

[andtony]

Thanks for the answer.

We made a little script which told us that the system is freezing everyday at about 6.55pm

We're looking for a software issue, but it's hard to find it. No computer on the network is sending anything at this time.

 

[PHV]

You may check all the cron jobs.

Hope This Helps, PH.
FAQ219-2884
FAQ181-2886