Hello, all. We're running an IBM x235 with Suse 8.0, Kernel version smp-2.4.18-57. We've been having it crash about once a day, leaving messages like the following in the /var/log/messages file: Sep 28 00:09:42 brie /usr/sbin/cron[895]: (CRON) error (can't fork) Sep 28 00:25:44 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0xf0/0) Sep 28 00:31:38 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x1d0/0) Sep 28 00:31:38 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x70/0) Sep 28 00:31:38 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0xf0/0) Sep 28 00:31:38 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x1f0/0) Sep 28 00:31:38 brie last message repeated 2 times Since this looks like a memory management issue, we have 4GB of physical memory and a 2GB swap. Any ideas? I'll be more than happy to provide more information if anyone wants it. -Michael Smit -msmit@cpm.com
Does your board have one or more CPUs? If it only has one, don't use the smp kernel. There are more sugestions on the SDB, search for "smp" or "crash". For example: " * SMP Kernel Is Installed Despite Existing Single Processor Motherboard (14.06.2002)" -- Cheers, Carlos Robinson El 02.09.28 a las 16:18, Mike Smit escribió:
Date: Sat, 28 Sep 2002 16:18:55 -0500 From: Mike Smit <msmit@cpm.com> To: suse-linux-e@suse.com Cc: Erick Hallick <ehallick@cpm.com> Subject: [SLE] Suse 8.0 crash issues.
Hello, all.
We're running an IBM x235 with Suse 8.0, Kernel version smp-2.4.18-57. We've been having it crash about once a day, leaving messages like the following in the /var/log/messages file:
Sep 28 00:09:42 brie /usr/sbin/cron[895]: (CRON) error (can't fork) Sep 28 00:25:44 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0xf0/0) Sep 28 00:31:38 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x1d0/0) Sep 28 00:31:38 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x70/0) Sep 28 00:31:38 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0xf0/0) Sep 28 00:31:38 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x1f0/0) Sep 28 00:31:38 brie last message repeated 2 times
Since this looks like a memory management issue, we have 4GB of physical memory and a 2GB swap. Any ideas? I'll be more than happy to provide more information if anyone wants it.
-Michael Smit -msmit@cpm.com
Thank you for your response. We have two processors on this machine, both 2.0 GHz Intel Xeons. I will search the messages file today for "smp" and "crash" messages and get back to you. -Michael Smit -msmit@cpm.com -----Original Message----- From: Carlos E. R. [mailto:robin1.listas@tiscali.es] Sent: Saturday, September 28, 2002 8:37 PM To: suse-linux-e@suse.com Subject: Re: [SLE] Suse 8.0 crash issues. Does your board have one or more CPUs? If it only has one, don't use the smp kernel. There are more sugestions on the SDB, search for "smp" or "crash". For example: " * SMP Kernel Is Installed Despite Existing Single Processor Motherboard (14.06.2002)" -- Cheers, Carlos Robinson
The only entry in /var/log/messages that contains "SMP" and appears to refer to the problem at hand is in the following block: Sep 27 00:21:19 brie kernel: klogd 1.4.1, ---------- state change ---------- Sep 27 00:21:29 brie kernel: Inspecting /boot/System.map-2.4.18-64GB-SMP Sep 27 00:21:56 brie kernel: Loaded 13537 symbols from /boot/System.map-2.4.18-64GB-SMP. Sep 27 00:21:56 brie kernel: Symbols match kernel version 2.4.18. Sep 27 00:21:57 brie kernel: Loaded 573 symbols from 17 modules. Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x70/0) Sep 27 00:21:57 brie last message repeated 3 times Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0xf0/0) Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x70/0) Sep 27 00:21:57 brie last message repeated 2 times I'm not sure if this helps or not. The problem appears to be a memory leak of some sort, since we lose the machine like clockwork eighteen hours or so after we start it. Thanks for your help so far. -Michael Smit -msmit@cpm.com -----Original Message----- From: Carlos E. R. [mailto:robin1.listas@tiscali.es] Sent: Saturday, September 28, 2002 8:37 PM To: suse-linux-e@suse.com Subject: Re: [SLE] Suse 8.0 crash issues. Does your board have one or more CPUs? If it only has one, don't use the smp kernel. There are more sugestions on the SDB, search for "smp" or "crash". For example: " * SMP Kernel Is Installed Despite Existing Single Processor Motherboard (14.06.2002)" -- Cheers, Carlos Robinson
I forgot to ask. When my machine crashes (1 cpu only, kernel 2.4.16, suse 7.3) both Caps Lock and Scroll Lock LEDs blinks alternatively. That type of crash have been reported by various people, with the only common thing of using suse 7.3 or 8.0; but I don't know if your LEDs do that blinking. I had to append "disableapic" in lilo.conf, and use gdm instead of kdm as graphical login, and some adjustments in the kernel that I have forgotten. Now, the crashes are more rare, and not so bad, I'm able to "halt". Usually, it is memory swap failing, and some processes like cron stopping suddenly. I thought kernel 2.4.18 had that solved, maybe not. There are more suggestions at suse's sdb, try them. -- Cheers, Carlos Robinson El 02.09.30 a las 17:19, Mike Smit escribió:
Date: Mon, 30 Sep 2002 17:19:08 -0500 From: Mike Smit <msmit@cpm.com> To: suse-linux-e@suse.com Subject: RE: [SLE] Suse 8.0 crash issues.
The only entry in /var/log/messages that contains "SMP" and appears to refer to the problem at hand is in the following block:
Sep 27 00:21:19 brie kernel: klogd 1.4.1, ---------- state change ---------- Sep 27 00:21:29 brie kernel: Inspecting /boot/System.map-2.4.18-64GB-SMP Sep 27 00:21:56 brie kernel: Loaded 13537 symbols from /boot/System.map-2.4.18-64GB-SMP. Sep 27 00:21:56 brie kernel: Symbols match kernel version 2.4.18. Sep 27 00:21:57 brie kernel: Loaded 573 symbols from 17 modules. Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x70/0) Sep 27 00:21:57 brie last message repeated 3 times Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0xf0/0) Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x70/0) Sep 27 00:21:57 brie last message repeated 2 times
I'm not sure if this helps or not. The problem appears to be a memory leak of some sort, since we lose the machine like clockwork eighteen hours or so after we start it. Thanks for your help so far.
-Michael Smit -msmit@cpm.com
-----Original Message----- From: Carlos E. R. [mailto:robin1.listas@tiscali.es] Sent: Saturday, September 28, 2002 8:37 PM To: suse-linux-e@suse.com Subject: Re: [SLE] Suse 8.0 crash issues.
Does your board have one or more CPUs? If it only has one, don't use the smp kernel.
There are more sugestions on the SDB, search for "smp" or "crash". For example: " * SMP Kernel Is Installed Despite Existing Single Processor Motherboard (14.06.2002)"
-- Cheers, Carlos Robinson
One more thing: I get those same messages when running xcdroast: (booting) Sep 14 21:29:36 nimrodel kernel: NVRM: AGPGART: allocated 16 pages Sep 14 21:29:42 nimrodel kernel: Switching off penguin. Sep 14 22:31:02 nimrodel kernel: tion failed (gfp=0x20/0) Sep 14 22:31:02 nimrodel kernel: __alloc_pages: 3-order allocation failed (gfp=0x20/0) Sep 14 22:31:03 nimrodel last message repeated 286 times Sep 14 22:31:17 nimrodel kernel: tion failed (gfp=0x20/0) Sep 14 22:31:17 nimrodel kernel: __alloc_pages: 3-order allocation failed (gfp=0x20/0) Sep 14 22:31:17 nimrodel last message repeated 32 times Sep 14 22:31:17 nimrodel kernel: __alloc_pages: 2-order allocation failed (gfp=0x20/0) Sep 14 22:31:17 nimrodel kernel: __alloc_pages: 3-order allocation failed (gfp=0x20/0) Sep 14 22:31:17 nimrodel kernel: __alloc_pages: 3-order allocation failed (gfp=0x20/0) Now look at the times in the xcdroast log: Sat Sep 14 01:13:42 2002 XCDR 0.98alpha9: Exiting... Sat Sep 14 22:30:38 2002 XCDR 0.98alpha9: Loading config file /home/cer/.xcdroast/xcdr098.cfg Sat Sep 14 22:30:38 2002 XCDR 0.98alpha9: Starting... Sat Sep 14 22:31:02 2002 XCDR 0.98alpha9: Entering create cd Sat Sep 14 22:31:17 2002 XCDR 0.98alpha9: Entering create: master menu Kernel error messages start the very same second as I enter the "create cd" menu. Interesting, isn't? -- Cheers, Carlos Robinson El 02.09.30 a las 17:19, Mike Smit escribió:
Date: Mon, 30 Sep 2002 17:19:08 -0500 From: Mike Smit <msmit@cpm.com> To: suse-linux-e@suse.com Subject: RE: [SLE] Suse 8.0 crash issues.
The only entry in /var/log/messages that contains "SMP" and appears to refer to the problem at hand is in the following block:
Sep 27 00:21:19 brie kernel: klogd 1.4.1, ---------- state change ---------- Sep 27 00:21:29 brie kernel: Inspecting /boot/System.map-2.4.18-64GB-SMP Sep 27 00:21:56 brie kernel: Loaded 13537 symbols from /boot/System.map-2.4.18-64GB-SMP. Sep 27 00:21:56 brie kernel: Symbols match kernel version 2.4.18. Sep 27 00:21:57 brie kernel: Loaded 573 symbols from 17 modules. Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x70/0) Sep 27 00:21:57 brie last message repeated 3 times Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0xf0/0) Sep 27 00:21:57 brie kernel: __alloc_pages: 0-order allocation failed (gfp=0x70/0) Sep 27 00:21:57 brie last message repeated 2 times
I'm not sure if this helps or not. The problem appears to be a memory leak of some sort, since we lose the machine like clockwork eighteen hours or so after we start it. Thanks for your help so far.
-Michael Smit -msmit@cpm.com
-----Original Message----- From: Carlos E. R. [mailto:robin1.listas@tiscali.es] Sent: Saturday, September 28, 2002 8:37 PM To: suse-linux-e@suse.com Subject: Re: [SLE] Suse 8.0 crash issues.
Does your board have one or more CPUs? If it only has one, don't use the smp kernel.
There are more sugestions on the SDB, search for "smp" or "crash". For example: " * SMP Kernel Is Installed Despite Existing Single Processor Motherboard (14.06.2002)"
-- Cheers, Carlos Robinson
participants (2)
-
Carlos E. R. -
msmit＠cpm.com