I've been having a rather serious problem recently which perhaps one of you
could toss me some pointers to help solve.
In every kernel I've run >1.1.59, there seems to be a hole in Linux's memory
management infrastructure :)
i.e. on a 32 meg machine, I run X, it uses a big blob of RAM, I run clients,
they use ram, I exit X, the RAM stays allocated! I run gcc a few times, yet
more RAM down the tubes.
This process continues until the system becomes almost unusably slow.
It's not uncommon for me to be running with a working pool of only 4-6M free
on a 32 meg system with only a single VT, no X, and a bash running!!
Is there any way I can track down who's munching all these resources?
Thanks,
-Chris Patti