Re: debian crashes server (can't use debian ? :-( )
> most likely there is a hardware problem. Linux doesn't just shutdown
> without a reason :)
Yep, my "workmachine" has 2 debian's and a win os running and i never
had problems with my debian systems before.
> win98 is a piece of shit OS, and doesn't stress the hardware nearly
> as much as a linux installation does(or perhaps even a NT4 install or
> win2000)
i know. anyway, i'm not going to run 98 on it. if my reinstall of debian
fails i'll try another linux distro to see if that makes any difference ( i
doubt it but i'm going to try it anyway).
Crux seemed nice, vector linux too. But i really don't think this will make
a difference.
(*** i just started my server while typing this message *** )
I just started the server: it booted *nicely* (shrug) into windows 98.
I opened windows explorer et voila. gone. same kind of shutdown i
experienced in linux. haha. so tomorrow, it's back to the computer
store since it ISNT linux related. i never tought it was and now i have
proof. Of course if they weren't able to detect it, i don't have much faith
in there skills either :-(
I'll try to look further myself.
Anyway, again while doing something that accesses the HD altough i
replaced it. hmmm. my the onboard ide controller is faulty?
> > This server had Suse 7.2 before i installed Debian Woody on it and i
> > remember that i had the "occasional" crash too altough it only happened
> > about two or three times.
>
> what kind of crash? your being very vague.
The crashes that i'm experiencing is mostly noticed when my system boots.
>From the moment the kernel starts accessing the harddisk it fails.
Sometimes it completes the boot and stays up and running for several
days. Sometime it doesn't. It just shuts down as if you would unplug
the power.
I noticed last week that when i booted the system with Knoppix that it
always crashes when the boot sequence is about to access the drive.
Sometimes i runs further, sometimes it doesn't.
As i said, a couple of times it could restart and then i checked the logs,
nothing to be found in there to indicate what could be wrong. I checked
nearly every log in /var/log.
So i have no messages or log files to go on.
> you need to determine the cause of the problem. you can start by
> telling the list exactly what hardware you have(If you have before,sorry
> I didn't see it).
Yeah, i have posted before on this problem ( 2 Nov 2002 ) under
"server with debian 3.0 randomly shuts down when booting"
and you had also replied to it :-)
And then some time later i said that it was solved. NOT :-)
I changed the powersupply to a 450 W but that didn't help
Software:
* debian woody testing, base install with shorewall, nothing else.
* 2.4.18 bf24 kernel, non customized
* modules: none (everything seemed to work right out of the box
with kernel 2.4.18bf24. i could be wrong though)
* not running x server, plain base install
Hardware:
* AHA 2940 PCI SCSI adapter
* ISA SB AWE64
* Guillemot maxtor gamer 32 MB (was in my desktop a while ago
but i wanted a faster card so i "upgraded" my server :-)
* 2 PCI Realtek rtl8139 network cards
* Maxtor 6 GB IDE harddisk (ext3 partitions)
* 160 MB EDO Ram
* Motherboard: DFI 686IPK (Intel 440FX chipset. Supports Intel
Pentium II CPU up to 266 MHz. 4xPCI, 4xISA (1 shared) ,
4 x SIMM sockets.)
* Award 4.51 PG Bios
* Plextor PX20 TS SCSI cdrom
* Pentium II 266 Mhz
* 450 W power supply
* sony disk drive
Since the crashes i changed my harddisk, my powersupply and
ran tests by removing the scsi card and scsi cdrom, removing
the isa soundblaster and so on.
Only thing i haven't changed yet is my guillemot graphics card.
Could a graphics card result in crashes while it's only used
when displaying console messages? Maybe it overheats?
> Is there any questionable hardware in
> the system(e.g. like a stick of ram you found under the couch).
No
> Are you overclocking?
No.
> How many fans in the system?
2 on the power supply and 1 on the processor
>What kind of power supply?
450 w
>Is the machine connected to a UPS?
Nope. Thats another test i did. I plugged my power cable into
another powercircuit to be sure it wasn't the circuit that was causing
the randome crashes.
> What is the temperature of the room the machine is running in?
Room temperature: between 14°C (57 °F) - 24 °C (75 °F)
> one of my co workers was having a machine crash quite often, it
> was an older box, all it took was some heat sink grease between the
> heat sink and cpu and it fixed it. I had another machine, dual p3 with
> a gig of ram, which I cannibalized from 2 other systems(which ran fine
> for 2 years straight). It crashed non stop. Even after running memtest86
> for more then 2 weeks 24/7 it detected no errors. I replaced the ram
> and the crashes immediately stopped. Further memory testing(1 stick at
> a time) revealed problems on the ram.
Interesting. I will try to test the memory too.
You mention memtest386. Are there any other tools to test the system?
> I've had crashes by having too much load on the power supply(e.g. 6
> hard disks on a 300 or 350 watt power supply). Upgrading the power
> supply(450 watt) fixed it.
It's already at 450 so i don't think that's the problem.
Anyway, thanks for the pointers.
(again, since you answered my very first posts about this problem too :-)
Benedict
Reply to: