Strange remote crash

Some days ago, my remote machine crashed. I think that some process grew up using all the available ram. If I am not wrong, these logs confirm that. Dec 26 17:35:32 praisenet kernel: __alloc_pages: 0-order allocation failed (gfp=0x1d2/0) Dec 26 17:35:32 praisenet kernel: VM: killing process nmbd I got many line like those. My question is: how can I find out which process made my machine crash? I had installed courier imap some time before the crash and I am suspecting it to be the cause of the crash, anybody had a similar experience? I know I can limit resources for remote users via pam, but I am not aware of a similar way to limit system daemons too. How could I do that? Praise