Cannot determine why my 8.1 box keeps hanging.
Hello list, Been a loyal SuSE user for a couple years now and recently bought a used box with the intent to run it as a database server. The seller claims that it ran Linux flawlessly before he sold it to me. Well, after three complete reinstalls and one hard drive failure, I just cannot get it to stay up. The frustrating thing is that there are never any error messages when it goes down. Actually, it doesn't really go down, it just hangs. No network and no X. Reset doesn't work, holding in the power button for 10 seconds does though to restart it. What I've noticed is that it seems to happen more often than not when I am using my laptop to run a remote desktop on this box using VNC. About once or twice a day everything will stop and I'll have to go do the power button thing. Nothing interesting in /var/log/messages (or any /var/log file for that matter). Is there anything I can do to get more information? Here's what I know: Major software installed: cups, samba, nfs, mysql, postgresql, oracle (though I quit running it due to crashing thinking it was the problem), tomcat, and zope. Almost complete set of YOU patches installed except new kernel and glibc which may completely hose oracle. Previous installation with all YOU patches failed too. 512 MB memory, 1GB swap, system doesn't seem overworked. I stopped nearly all processes including cron and it still died once. I had a hard drive failure a while back so I installed on a new drive, then I put the old drive in as a slave (IDE) and zeroed it out. I don't have anything on it, but it is still on the IDE cable, could it be causing hardware problems? I also noticed that the video card fan is not always working. The box has three other fans though. Might try to replace the video card to see if that helps. Any other thoughts? Let me know what other information will help in diagnosing the problem... Thanks, Josh
-----Original Message----- From: Josh Trutwin <josh@trutwins.homeip.net> To: Mailing Suse Linux <suse-linux-e@suse.com> Date: Thu, 14 Aug 2003 17:59:26 -0500 Subject: [SLE] Cannot determine why my 8.1 box keeps hanging.
Hello list,
Been a loyal SuSE user for a couple years now and recently bought a used box with the intent to run it as a database server. The seller claims that it ran Linux flawlessly before he sold it to me.
Well, after three complete reinstalls and one hard drive failure, I just cannot get it to stay up. The frustrating thing is that there are never any error messages when it goes down. Actually, it doesn't really go down, it just hangs. No network and no X. Reset doesn't work, holding in the power button for 10 seconds does though to restart it.
What I've noticed is that it seems to happen more often than not when I am using my laptop to run a remote desktop on this box using VNC. About once or twice a day everything will stop and I'll have to go do the power button thing. Nothing interesting in /var/log/messages (or any /var/log file for that matter). Is there anything I can do to get more information?
Here's what I know:
Major software installed: cups, samba, nfs, mysql, postgresql, oracle (though I quit running it due to crashing thinking it was the problem), tomcat, and zope.
Almost complete set of YOU patches installed except new kernel and glibc which may completely hose oracle. Previous installation with all YOU patches failed too.
512 MB memory, 1GB swap, system doesn't seem overworked. I stopped nearly all processes including cron and it still died once.
I had a hard drive failure a while back so I installed on a new drive, then I put the old drive in as a slave (IDE) and zeroed it out. I don't have anything on it, but it is still on the IDE cable, could it be causing hardware problems?
I also noticed that the video card fan is not always working. The box has three other fans though. Might try to replace the video card to see if that helps.
Any other thoughts? Let me know what other information will help in diagnosing the problem...
Thanks,
Josh
Josh, Kinda sounds like a hardware problem, maybe the power supply is going bad does not have enough power. Starting to see the same thing here and ans suspect it's the power supply. Ken
On Thursday 14 August 2003 14:59, Josh Trutwin wrote:
Hello list,
Been a loyal SuSE user for a couple years now and recently bought a used box with the intent to run it as a database server. The seller claims that it ran Linux flawlessly before he sold it to me.
Well, after three complete reinstalls and one hard drive failure, I just cannot get it to stay up. The frustrating thing is that there are never any error messages when it goes down. Actually, it doesn't really go down, it just hangs. No network and no X. Reset doesn't work, holding in the power button for 10 seconds does though to restart it.
Here's what I know:
512 MB memory, 1GB swap, system doesn't seem overworked. I stopped nearly all processes including cron and it still died once.
I had a hard drive failure a while back so I installed on a new drive, then I put the old drive in as a slave (IDE) and zeroed it out. I don't have anything on it, but it is still on the IDE cable, could it be causing hardware problems?
Yes. Unconnect the old drive for a while. I've had problems with drives that periodically lock up the entire machine even when that drive was not actually being used.
I also noticed that the video card fan is not always working. The box has three other fans though. Might try to replace the video card to see if that helps.
Maybe it only works when it needs to, and maybe its just full of dirt and won't turn. Give it a nudge and see if it will start. A new vid card may be in order - get one without a fan. Fans go bad way too often.
Any other thoughts? Let me know what other information will help in diagnosing the problem...
Run an exhaustive memory test on this machine for several hours. You may have ram with intermitant failures. See if you can get lmsensors package working to measure the temp and voltages of the system. If you have a spare power supply hook it up (you don't have to go thru a full install, just cable it in with the cover off for a day or two to see if it still fails). I'm presumeing an ATX power supply and motherboard connection here... Notice the larger capacitors on the mobo near where the atx power supply connects. If any of them have bulgeing ends, and perhaps leakage, that means your mobo is dieing. This is fairly common on slot 1 mobos of a certain vintage, namely just about the time the P3-500mhz cpus came out. Under rated power section on the mobo. Do-Over. -- _____________________________________ John Andersen
Yes. Unconnect the old drive for a while. I've had problems with drives that periodically lock up the entire machine even when that drive was not actually being used.
Thanks for the suggestions, last night I took the box apart and reseeded everything (there is only a NIC and a video card though) and re-plugged in all cables. Took out the bad drive and rebooted. It worked through the night, this morning I ran VNC remote desktop for about 15 minutes and it choked. It must be the video card. I'm assuming that the video card is involved even when running remote desktops.
Maybe it only works when it needs to, and maybe its just full of dirt and won't turn. Give it a nudge and see if it will start. A new vid card may be in order - get one without a fan. Fans go bad way too often.
Run an exhaustive memory test on this machine for several hours. You may have ram with intermitant failures.
See if you can get lmsensors package working to measure the temp and voltages of the system. If you have a spare power supply hook it up (you don't have to go thru a full install, just cable it in with the cover off for a day or two to see if it still fails).
I'm presumeing an ATX power supply and motherboard connection here... Notice the larger capacitors on the mobo near where the atx power supply connects. If any of them have bulgeing ends, and perhaps leakage, that means your mobo is dieing. This is fairly common on slot 1 mobos of a certain vintage, namely just about the time the P3-500mhz cpus came out. Under rated power section on the mobo. Do-Over.
It's a PIII 1Ghz with ATX. It's only powering a HD, floppy, CDROM, NIC, AGP and two extra fans. Power supply problems wouldn't surprise me though. I notice that when I plug the power cord in, it make a few faint crackling sounds. I think the first thing I'll try is just getting a low-end video card. Thanks, Josh
The 03.08.15 at 09:52, Josh Trutwin wrote:
It's a PIII 1Ghz with ATX. It's only powering a HD, floppy, CDROM, NIC, AGP and two extra fans. Power supply problems wouldn't surprise me though. I notice that when I plug the power cord in, it make a few faint crackling sounds.
I would try checking the psu cables first, replace the mains cable, and finally, the psu itself. I don't like noises, except in motors. -- Cheers, Carlos Robinson
Josh Trutwin wrote:
Yes. Unconnect the old drive for a while. I've had problems with drives that periodically lock up the entire machine even when that drive was not actually being used.
Thanks for the suggestions, last night I took the box apart and reseeded everything (there is only a NIC and a video card though) and re-plugged in all cables. Took out the bad drive and rebooted. It worked through the night, this morning I ran VNC remote desktop for about 15 minutes and it choked. It must be the video card. I'm assuming that the video card is involved even when running remote desktops.
Depends how you use it. On display 0 it probely would, but the other display it doesn't. Stefan
On Sunday 17 August 2003 08:12 am, S. Bulterman wrote:
Yes. Unconnect the old drive for a while. I've had problems with drives that periodically lock up the entire machine even when that drive was not actually being used.
I have seen a similar problem on many boxes when the pins in the power cable going to the hd become spread a little. That can drive you nuts! Pull the plug and look at the pins, one may look a bit wider than the others. A simple fix is to use a very thin screwdriver to compress the pin slightly. What usually happens is you have to wiggle the connector to get it off a drive and that wiggle can spread the pin slightly. i found the problem on one machine by carefully moving the power cables as the machine was running. when the connector failed, the hd would click. BE VERY JUDICIOUS IN DOING THIS AS YOU CAN CORRUPT YOUR DATA ON THE HD AS IT STOPS AND STARTS!!!! Richard
The 03.08.14 at 17:59, Josh Trutwin wrote:
Well, after three complete reinstalls and one hard drive failure, I just cannot get it to stay up. The frustrating thing is that there are never any error messages when it goes down. Actually, it doesn't really go down, it just hangs. No network and no X. Reset doesn't work, holding in the power button for 10 seconds does though to restart it.
It is 8.1? The kernel on 8.1 is faulty, if you have at least one reiserfs partition. You should enable "barrier=none" option in fstab for those partitions, or use the patched kernel they provided. The symptom was a complete absolute sudden crash: every ting stopped, nothing in the logs, no network access, keyboard locked, num lock and cap lock led blinking --> ie, kernel crash signaling. Often the HD led would remain on. One more possibility apart from the other things you were told already :-) -- Cheers, Carlos Robinson
participants (6)
-
Carlos E. R.
-
John Andersen
-
Josh Trutwin
-
Ken Schneider
-
Richard
-
S. Bulterman