Linus Torvalds wrote (ao):
> On Tue, 21 Mar 2006, Jeff Garzik wrote:
> > In any case, one could be lazy, and simply bisect the main tree
> > (and/or simply verify that the problem is gone in
> > 2.6.16-git<today>).
>
> Yes, just testing the current git tree (and if you're not a git user,
> just waiting for the next nightly snapshot) sounds like the
> appropriate thing to do.
The 2.6.16-git3 snapshot is stable for me like -rc6-mm1 and -rc6-mm2
are :-)
To recap:
Running eight Maxtor disks connected to a
MV88SX6081 8-port SATA II PCI-X Controller (rev 09)
mdadm -C -l5 -n8 /dev/md0 /dev/sda1 /dev/sdb1 /dev/sdc1 \
/dev/sdd1 /dev/sde1 /dev/sdf1 /dev/sdg1 /dev/sdh1
mke2fs -j -m1 /dev/md0
mount -o data=writeback,nobh /dev/md0 /mnt
for i in `seq 4`
do dd if=/dev/zero of=bigfile.$i bs=1024k count=10000
dd if=bigfile.$i of=/dev/null bs=1024k count=10000
done
time md5sum bigfile.*
time rm bigfile.*
or
for i in `seq 4`
do ( dd if=/dev/zero of=bigfile.$i bs=1024k count=10000 ; \
time md5sum bigfile.$i ) &
done
Kernel 2.6.16-rc6 and 2.6.16 always crash during the md5sum (and leave
no output).
2.6.16-rc6-mm1, 2.6.16-rc6-mm2 and 2.6.16-git3 are stable without a
crash or data corruption.
I'm aware that the test is very simple and bugs might still hide. I'll
go and find some more serious stress tests.
Should I do more testing/bisecting/etc?
Btw, I do still get these (any kernel), but with no visible effect:
[ 2306.952183] ata6: translated ATA stat/err 0xd0/00 to SCSI SK/ASC/ASCQ 0xb/47/00
[ 2306.952246] ata6: status=0xd0 { Busy }
[ 2891.892225] ata5: translated ATA stat/err 0xd0/00 to SCSI SK/ASC/ASCQ 0xb/47/00
[ 2891.892277] ata5: status=0xd0 { Busy }
[ 4550.013582] ata6: translated ATA stat/err 0xd0/00 to SCSI SK/ASC/ASCQ 0xb/47/00
[ 4550.013637] ata6: status=0xd0 { Busy }
[ 4864.850340] ata9: translated ATA stat/err 0xd0/00 to SCSI SK/ASC/ASCQ 0xb/47/00
[ 4864.850393] ata9: status=0xd0 { Busy }
[ 4968.681651] ata9: translated ATA stat/err 0xd0/00 to SCSI SK/ASC/ASCQ 0xb/47/00
[ 4968.681711] ata9: status=0xd0 { Busy }
Thanks!
Sander
--
Humilis IT Services and Solutions
http://www.humilis.net
-
: 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
