Does anyone else have one of these on 24/7 running RHEL 5.4 or Centos 5.4 ?? That box mysteriously stops ( no crash data, it just stops) hours or days after boot, with or without X. Cntl-Alt-Delete reboots it without power cycling. The usual diagnostics (memtest and fsck) return no errors. Works great until it freezes. No evidence of kernel oops anywhere. If others report the same issues we will scrap the board. Otherwise I will continue to investigate what is happening, since it is no fun to pull the board out and replace with an atom processor and it is not critical. The hard disk in it came from an Athlon M board that died an untimely death this week. Service cpufreq reports "unavailable" so I am assuming that it is turned off. Could the cpu freqquency control be doing something without tellling us? We havea power monitor on it so I can discern if it dies with max cpu activity or not. Also I am leaving it run "top" without screensaver so I can see if a particular load or program is borking it. Am I missing something obvious here? TIA. benm
On 22/11/09 18:23, Ben Mohilef wrote:> That box mysteriously stops ( no crash data, it just stops) hours or days after boot, with or without X. > Cntl-Alt-Delete reboots it without power cycling. The usual diagnostics (memtest and fsck) return no > errors. Works great until it freezes. No evidence of kernel oops anywhere.$ cat /proc/cpuinfo | grep model model : 9 model name : VIA Nehemiah $ uptime 01:25:57 up 40 days, 5:26, 11 users, load average: 1.15, 1.03, 1.01 $ uname -r 2.6.18-164.el5 I dont seem to have any problems ( use this machine as a cheap h/w random number generator, so its always under load ~ 1 ) - KB
> I dont seem to have any problems ( use this machine as a cheap h/w > random number generator, so its always under load ~ 1 ) >Same board, same model. Thanks for the information. This looked and felt like an acpi induced problem. I discovered that acpid was turned on for some reason, so I turned off the acpid daemon and it has behaved well for the last day. Are you booting yours with acpi=off as well ?? Thanks again. regards, benm
----- "Karanbir Singh" <mail-lists at karan.org> wrote:> On 23/11/09 17:50, Ben Mohilef wrote: > > This looked and felt like an acpi induced problem. I discovered that > acpid was turned on for some > > reason, so I turned off the acpid daemon and it has behaved well for > the last day. Are you booting > > yours with acpi=off as well ?? > > nope, just the straight kernel and initrd. nothing special in there. > > this is actually quite a boring standard install with almost nothing > changed on there from what is left behind by a minimal 5.1 install ( > and > its been updated nightly since then ) > > - KBIs it possible you have an application that runs at odd times that is compiled for the wrong arch? IIRC the C3 is i586 compatible (almost i686 but missing something stupid like 'cmov' instruction). I've got a handful of C3 based systems in my personal lab and run into 'oddities' every now and then because of this... Tim Nelson Systems/Network Support Rockbochs Inc. (218)727-4332 x105
On 23/11/09 17:50, Ben Mohilef wrote:> This looked and felt like an acpi induced problem. I discovered that acpid was turned on for some > reason, so I turned off the acpid daemon and it has behaved well for the last day. Are you booting > yours with acpi=off as well ??nope, just the straight kernel and initrd. nothing special in there. this is actually quite a boring standard install with almost nothing changed on there from what is left behind by a minimal 5.1 install ( and its been updated nightly since then ) - KB