That's not necessarily true. Forks start failing on one of my machines
once the number of processes goes above ~(140 + number_of_root_processes).
Since there are around 30 root processes on that machine, this number works
out to be ~170, which is well below 512.
I don't yet know what it is running out of. It is not a per-user limit,
because when one user pushes the total past 140, all non-root forks start
to fail. Root forks continue to work fine, and I've been able to start 50
new root processes while in the "user forks are failing" state (I stopped at
50 because I got tired of typing "sh<return>").
The machine has plenty of RAM and is not running out of swap:
total used free shared buffers cached
Mem: 517440 500076 17364 151476 104812 330812
-/+ buffers/cache: 64452 452988
Swap: 128516 0 128516
I've got other machines running basically the same software, and they
don't run into this problem. Kernel is 2.2.4.
--Tim Smith
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/