>>> free -m >>> total used free shared buffers cached >>> Mem: 249 246 2 72 19 161 >>> -/+ buffers/cache: 66 183 >>> Swap: 125 37 87 > It is apache-ssl which uses ..when i stop it i get a free of 67MB > when i restart i get only 2MB of free memory.. > since its a server i need to have ssl on.. > i am scared to reboot:( cause of this seg fault.. > since its a server .. > its happeening from few acculmulated uptime.. > > Vijaya Oh, sorry, now on second look I see I've made a wrong conclusion. The memory is still there, it's just in the cache - the second line tells the true values (used 66M, free 183M, after subtracting/adding buffers/cache). This lonely and alarming "2" made me rash. ;-) So it's not the memory, although everything points in that direction... the VM wreaking havoc killing processes, the binaries with the smallest memory footprint segfaulting... Oooh ha. I've found something more promising. :-) Perhaps not the memory, but the maximum number of open files is killing your system... I just stumbled across the "VFS:" message in your dmesg dump. Please check the output of /proc/sys/fs/file-nr, and if the first (allocated file handles) and last (maximum number of file handles) numbers are close or equal, increase the maximum via echo NEW_MAXIMUM > /proc/sys/fs/file-max Mine is showing jcn@hejre:/etc/apache$ cat /proc/sys/fs/file-nr 1375 0 51004 so there's lots of room, but your VFS message suggests that that may be different on your server. I hope we're getting closer... I'm seriously running out of ideas... ;-) Jan -- Jan C. Nordholz <jckn At gmx net>
Attachment:
signature.asc
Description: Digital signature