Jim MacBaine wrote:
On Jan 13, 2008 1:07 PM, Mikael Pettersson <mikpe@xxxxxxxx> wrote
The fact that the problems occur on different disks on
different controllers driven by different drivers indicates
that it's not a disk, controller, or driver problem.
I strongly suspect an underdimensioned or failing PSU.
Thanks a lot for your clues.
I bought a new PSU on Monday and didn't get any new disk failures for
days. But last night the same time-outs occurred again on two disks.
I guess I will try to replace the motherboard including the two SATA
controllers next.
I don't know if your problems are similar to mine or not. But I have
been having extensive problems for quite some time now. Do you get these
timeouts when using optical drives? That's what seems to trigger it in
my case: If I'm using the optical drives, I'll often see the errors with
them first, and then the whole ATA subsystem seems to go down. Then I
get journal commit errors, general read errors, etc, until the system
basically locks up. Worst case, it all happens very suddenly, and
there's not even anything in the logs. Just a couple messages to the
terminal, usually a journal commit error.
In my case, the opticall drives are a brand new Pioneer DVD-RW on SATA
and an old Plextor on PATA. I mostly see the errors with the latter but
have also seen them with the former. I'd thought I'd fixed it by adding
pnpacpi=off and pci=nomsi,nommconf to the kernel boot options, as well
as libata noacpi=1 to modules.conf, but now I've just had the problem
again. I'm now thinking I should try eliminating the Plextor drive. It
may be that it's the PATA drive that is causing all the trouble. I'll
report if so.
FYI, here are the relevant modules being loaded:
[root@rghquad rgheck]# lsmod | grep ata
pata_amd 20293 0
pata_pdc2027x 17477 0
sata_nv 25157 8
ata_generic 14405 0
libata 114673 4 pata_amd,pata_pdc2027x,sata_nv,ata_generic
scsi_mod 145657 5 sr_mod,sg,usb_storage,libata,sd_mod
The IDE interface is an nVidia MCP55, apparently, on an ASUS P5N32-E mb.
I doubt very much it's a PS issue in my case. There's not that much in
the box.
Richard
-
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html