[p4] Perforce server lockup problem

Arnt Gulbrandsen arnt at gulbrandsen.priv.no
Mon Mar 14 02:08:12 PST 2005

Kenneth Olwing writes:
> About a year ago we tried upgrading our trusty old Linux to RH 9 
> Enterprise. It mostly worked, but also exhibited freezes (the 'press 
> the powerbutton' kind). After a day I simply called it a failure and 
> went back to the old server pending investigation.
> It ended up being a memory overcommit situation that killed Linux.

Why did you allocate so much memory? Is/was your server seriously short 
of RAM, or was there some other serious problem?

The overcommit froze linux, but did not cause the problem - it simply is 
a particularly unpleasant consequence. When you do this

> echo 2 > /proc/sys/vm/overcommit_memory

you're telling the kernel to pick some other unpleasant response to the 

> Now, changed working style nicely kills the process if it runs away.

Usually. However, the kernel may instead choose to kill p4d and let your 
process run for another minute. The kernel tries to guess which process 
it's best to kill, but its guess isn't always correct.

I strongly suggest that you investigate what's eating your memory.


More information about the perforce-user mailing list