You are not logged in.
Yeah, I'm using an Intel D945GCLF as my home server and, well. For outside usage as well at some points.
Lately when I've logged on to SSH and for example tried to execute ´ls -la´ it just returns "Bus error".
I restart the server and it works fine for a couple of more days, then the same thing again.
I've looked through the logs several times, no indication on any type of hardware issues or anything remotely related to one either.
Is there anyway I could diagnose my hardware from Linux?
Oh by the way, I overclocked (or rather underclocked) my GPU to 66Mhz to decrease power usage and thermals, might this be the issue?
Last edited by root_op (2010-09-27 08:26:30)
Offline
Hello
First, I think we can rule out the GPU from this issue.
I think this might be a ATA bus error. Can you tell me if you are using a SATa or PATA hard disk ?
Next time it happens, please try as root :
- `ls /usr` (should fail)
- search for relevant errors at the end of `dmesg`
- `ls /sys` (or any virtual filesystem, like /dev or /proc), it should work if the issue is with the ATA bus
- `hdparm -t /dev/sda` (replace sda with whatever name your harddisk has) to try accessing your hardisk directly (not through a filesystem)
Cheers
Offline
Thanks for the detailed reply, I sure will. Although there's one wierd thing about this.
´ls´ works fine, but ´ls -la´ doesn't. If it was due to the ATA buss, wouldn't ´ls´ fail as well?
I'm using SATA drivers.
Thank you.
Offline
This has nothing to do with the hard disk bus. An error there would manifest as an unresponsive/laggy system and errors in dmesg.
This is output on the command line from a SIGBUS signal, like "Segmentation fault" results from a SIGSEGV. These are generated by the kernel because of some problem.
http://en.wikipedia.org/wiki/SIGBUS
Is there a BIOS update available? The kernel might take cues from the BIOS on how to build the memory map (the -e820 lines in dmesg). Other than that I would try reinstalling coreutils. A corrupted binary may create some kind of memory address problem, I guess.
Last edited by fphillips (2010-10-01 03:31:09)
Offline
I have not seen any "-e820" lines in my dmesg, although this has worked before _exactly_ like it is now. That is, software versions. Hardware and so on.
I'll look over if there's a newer BIOS available, although I doubt that's the issue.
Thanks for the reply!
Offline
I'm sorry to say that it was infact a hardware failure. Today my LAN port gave up on me, I plugged it in with no signal on neither Ethernet, USB, PS/2 or VGA.
Case closed.
Offline