My Gentoo server is very unstable

mcbarlo · Tux's lil' helper Joined: 18 Jul 2005 Posts: 130

I have server/lan router on Gentoo. It was working great for one year.

Some days ago it starts randomly crash. I completly don't know why?! Server is working 2-3 days, after this time stops answer for evrything: keyboard freez, monitor is black only hard reset can help.

I think hardware is ok. I switched power supply, memory and UPS and it still crash. Memtest86 told that RAM is ok. In logs I didn't find anything interesting.

Maybe some application and/or kernel is cause of problem? I have 2.6.14-r5 hardened kernel. On server works always apache2, mysql, proftpd and bind in newest version from stable portage.

My hardware is:

P4-Celeron 2.26Ghz with ASRock motherboard (Intel chipset)
512MB RAM Kingstone
Riva TNT2 M64
IDE harddisk, some ethernet adapters (RTL8139)

Someone have similar problem? I don't know what can I do... I'm waiting for advices. Thnaks.

Zardam · n00b Joined: 20 May 2005 Posts: 17

Hello,

It really looks like an hardware problem. You don't have checked you cpu or mainboard, and there may be problem with them (like old capacitors or such things).

Maybe I'm wrong, but it's a possibility.

mcbarlo · Tux's lil' helper Joined: 18 Jul 2005 Posts: 130

It's possible, but changing motherbord take long time (when server will be offline). But if all other things don't help I must do that.

n0z3m · n00b Joined: 10 Jan 2005 Posts: 49 Location: Netherlands

Do a memtest (memtest86) , check your HD with the tools provided by the manufacturer, and install lmsensors, wierd behaviour is generally caused
by bad memory, and/or overheating..
_________________
Holy penguin pee!

Headrush · Posted: Wed Mar 08, 2006 1:59 pm Post subject:

Could be an overheating issue too.

If you can handle having the server non-operation for a while, boot off a LiveCD and run the same time as it caused failures before and see if it still fails. If it locks up you know it isn't software.

Secondly, open it up and see how dusty it is. Its amazing how fast computers can become overcome by dust. Get some spray air and clean it out. You could also blow a standard house fan into the side of open case and see if it fails. This should provide enough draft to avoid the hear issues if that was the problem. If you still get lockups, it indeed sounds like hardware failure.

n0z3m · n00b Joined: 10 Jan 2005 Posts: 49 Location: Netherlands

Forgot something :

Also check your MoBo for faulty capacitators, look if they are still flat on the top, if not, that's your problem
_________________
Holy penguin pee!

Headrush · Posted: Wed Mar 08, 2006 2:14 pm Post subject:

mcbarlo · Tux's lil' helper Joined: 18 Jul 2005 Posts: 130

mcbarlo · Tux's lil' helper Joined: 18 Jul 2005 Posts: 130

I have checked capacitors and all of them looks ok (all are flat). Do you have any suggestions?

n0z3m · n00b Joined: 10 Jan 2005 Posts: 49 Location: Netherlands

nothing in /var/log/messages ? what is the mobo typenumber?
_________________
Holy penguin pee!

brot · Guru Joined: 06 Apr 2004 Posts: 322

I think this is caused by the motherboard. I have had several Asrock Mainboads which caused problems.

Just wanted to add that, maybe it helps

brot

mcbarlo · Tux's lil' helper Joined: 18 Jul 2005 Posts: 130

mcbarlo · Tux's lil' helper Joined: 18 Jul 2005 Posts: 130

I swiched my mother board and CPU to another model. Now I have ASRock too, but another. This board and CPU works great with WinXP, and on my server is unstable.

I think mother board and processor is ok. What can crash? May be memory, but memtest told that everything is ok.

n0z3m · n00b Joined: 10 Jan 2005 Posts: 49 Location: Netherlands

Haven't got a clue with this information, if you run a server, I'd take a Asus of Abit Mobo (with latest bios), Kingston ram & seagate HD, that would rule out
almost every possible fault on that end....
_________________
Holy penguin pee!