Since the last two couple of kernel patches I've had problems with various services unexpectedly dying (including sshd, cupsd and ypbind). Here's what I see in the logs: Feb 27 05:11:07 tweety -- MARK -- Feb 27 05:16:02 tweety kernel: __alloc_pages: 0-order allocation failed (gfp=0xf 0/0) Feb 27 05:16:18 tweety kernel: __alloc_pages: 0-order allocation failed (gfp=0xf 0/0) Feb 27 05:18:14 tweety kernel: __alloc_pages: 0-order allocation failed (gfp=0xf 0/0) Feb 27 05:18:15 tweety kernel: __alloc_pages: 0-order allocation failed (gfp=0x1 d2/0) Feb 27 05:18:15 tweety kernel: VM: killing process sshd I have k_smp, 2.4.20, 105 SuSE Linux 8.2 (i586), VERSION = 8.2 Linux version 2.4.20-64GB-SMP (root@SMP_X86.suse.de) (gcc version 3.3 20030226 ( prerelease) (SuSE Linux)) #1 SMP Sat Feb 7 02:07:52 UTC 2004 This all started on the 20th February and there have been thirteen instances since then. Any ideas? -- Simon Oliver
After a little investigation it seems that my system is running out of memory! It looks like a user is trying to allocate more memory that the system has (real+swap). I know that this can be controlled by setting resource limits. But please can someone explain the difference between the following, and where one would use them. datasize stacksize memoryuse vmemoryuse memorylocked I've read the man pages for both bash/ulimit and tcsh/limit but the only limit that seems to make any difference is vmemoryuse. -- Simon Oliver
participants (1)
-
Simon Oliver