More VIA/SATA issues
Ok, I am at a loss. I am having increasingly frequent disk timeouts which have caused me to have to rebuild my system, and now I can't even boot the re-installed system. Its a VIA K8 Albatron MB with a maxtor 200G drive. The major symptom is that I have a clicking sound from the drive that seems to be the heads parking whenever it locks up and the sound then cycles every 30 seconds after that. I've tried 2.6.5 kernel with only worse results. I've tried turning off DMA, ACPI, APM, POLL, etc individually. With the failsafe boot it will work indefinitely without lock-up (or left overnight in install state) so it doesn't seem to be a hardware issue. Turning off APM and ACPI together seems to make it much better. I thought this was the DMA issue with VIA, but using 2.6.5 with/without DMA turned off still produces the same result. It does give a dma_timeout error on boot if dma is turned on in the bios, but turning dma off in bios (and kernel) does not solve the problem, it only gets rid of the message. And this seems to occur whether the sda scsi emulated driver is used or the hda driver is used, IE changing the initrd lines to point to a different via_sata and libata. Suggestions? This drive is supposed to support hypertransport, so I have a problem with just dumping it for an IDE drive, but I'm out of ideas at this point. Thanks.
tisdag 20 april 2004 16:58 skrev Kevin Trotman:
Ok, I am at a loss. I am having increasingly frequent disk timeouts which have caused me to have to rebuild my system, and now I can't even boot the re-installed system. Its a VIA K8 Albatron MB with a maxtor 200G drive. The major symptom is that I have a clicking sound from the drive that seems to be the heads parking whenever it locks up and the sound then cycles every 30 seconds after that. I've tried 2.6.5 kernel with only worse results. I've tried turning off DMA, ACPI, APM, POLL, etc individually. With the failsafe boot it will work indefinitely without lock-up (or left overnight in install state) so it doesn't seem to be a hardware issue. Turning off APM and ACPI together seems to make it much better. I thought this was the DMA issue with VIA, but using 2.6.5 with/without DMA turned off still produces the same result. It does give a dma_timeout error on boot if dma is turned on in the bios, but turning dma off in bios (and kernel) does not solve the problem, it only gets rid of the message. And this seems to occur whether the sda scsi emulated driver is used or the hda driver is used, IE changing the initrd lines to point to a different via_sata and libata.
Suggestions? This drive is supposed to support hypertransport, so I have a problem with just dumping it for an IDE drive, but I'm out of ideas at this point.
I got the same activity here, and this is a MSI Neo mainboard. What I did, was turn off DMA, then shut down the computer for a while and turn it on again. It hasn't occurred now for a few hours ... but this is concerning, to say the least.
Thanks.
On 20 Apr 2004 at 9:58, Kevin Trotman wrote:
Suggestions? This drive is supposed to support hypertransport, so I have a problem with just dumping it for an IDE drive, but I'm out of ideas at this point.
The last time I heard such clicks similar to yours my HDD crashed a few days later... In my system with a MSI K8T Neo I have a Seagate 120GB SATA Drive, connected to the VIA chipset, no problems so far (after updating BIOS)...
On Tue, Apr 20, 2004 at 09:58:20AM -0500, Kevin Trotman wrote:
Suggestions? This drive is supposed to support hypertransport, so I have a problem with just dumping it for an IDE drive, but I'm out of ideas at this point.
Sounds like a dying harddisk to me. You can use smartctl to check the number of hard errors. Make sure you have good backups. -Andi
participants (4)
-
Andi Kleen -
h.andy＠gmx.de -
Kevin Trotman -
Örn Hansen