Dear list, the error, actually, is: kernel: __alloc_pages: 0-order allocation failed (gfp=0xf0/0) I have lots of these errors and eventually my 7.3 system is unable to work. Some things are working but I'm unable to "shutdown -r" the system. The machine "partially" freezes. That is, I'm unable to shut it down with ctrl-alt-delete, too; alternate consoles are not working, but I can login with ssh. Really strange stuff. Eventually I had to press the reset button and, of course, I have to manually fsck all the disks. :-( At first I suspected cron, because the errors came at nigth, around 00:57-00:59, but today they started at 2:07:22 and ended at 2:18:32, with "gfp={0xf0,0x1f0,0x70}/0" randomly alternating. Only a few gfp=0x1d0/0 and gfp=0x20/0 are scattered in between. Lots of "last message repeated x times", too. Strangely I had: Feb 3 01:07:41 my_computer kdm[992]: Abnormal helper termination, code 1, signal 0 Feb 3 01:07:40 my_computer nscd: 903: cannot stat() file `/etc/passwd': Cannot allocate memory Feb 3 01:07:46 my_computer kdm[992]: fatal IO error 12 (Cannot allocate memory) Feb 3 01:07:51 my_computer kdm[962]: Cannot run config reader: pipe() failed mixed between the allocation failed errors, as if they were delayed.. ??? Other errors among them are: Feb 3 02:08:31 my_computer atalkd[991]: select: Cannot allocate memory Feb 3 01:16:54 my_computer nscd: 903: cannot stat() file `/etc/passwd': Cannot allocate memory Feb 3 02:17:04 my_computer sshd[1140]: error: select: Cannot allocate memory and a lot of: Feb 3 02:08:37 my_computer syslogd: select: Cannot allocate memory Now my problem is: how can I find who is calling __alloc_pages?? No clues in the var/log files. Looking at the errors above, I suspect kde, but why the almost 1 hour delay?? Ah! Yes, I ran memtest and I had no errors. Someone in this list is able to help me? Linux 2.4.10-64GB-SMP i686 system. Thank you, Dott. Ermanno Polli INFN-LNF