I have a promise ultra66 card and two maxtor 40gb udma/66 drives.
dmesg snippet:
PDC20262: IDE controller on PCI bus 00 dev 88
PDC20262: not 100% native mode: will probe irqs later
PDC20262: (U)DMA Burst Bit ENABLED Primary PCI Mode Secondary PCI Mode.
ide2: BM-DMA at 0x6400-0x6407, BIOS settings: hde:DMA, hdf:DMA
ide3: BM-DMA at 0x6408-0x640f, BIOS settings: hdg:DMA, hdh:pio
hda: WDC AC32100H, ATA DISK drive
hde: Maxtor 94098U8, ATA DISK drive
hdg: Maxtor 94098U8, ATA DISK drive
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
ide2 at 0x6000-0x6007,0x6102 on irq 11
ide3 at 0x6200-0x6207,0x6302 on irq 11
hda: Disabling (U)DMA for WDC AC32100H
hda: DMA disabled
hda: WDC AC32100H, 2014MB w/128kB Cache, CHS=1023/64/63
hde: Maxtor 94098U8, 39082MB w/2048kB Cache, CHS=79406/16/63, UDMA(66)
hdg: Maxtor 94098U8, 39082MB w/2048kB Cache, CHS=79406/16/63, UDMA(66)
Floppy drive(s): fd0 is 1.44M
FDC 0 is a post-1991 82077
md driver 0.36.6 MAX_MD_DEV=4, MAX_REAL=8
raid0 personality registered
raid1 personality registered
I have patched my kernel with the ide patch. RAID support and both
the raid1 and raid0 personalities are compiled right into the kernel.
I'm using the raidtools that came with slackware 7.
my /etc/raid.conf looks like:
chrome:~# cat /etc/raid.conf
# Sample raid-1 configuration
raiddev /dev/md0
raid-level 1
nr-raid-disks 2
nr-spare-disks 0
device /dev/hde1
raid-disk 0
device /dev/hdg1
raid-disk 1
Ok, here's the problem. I can mkraid just fine, takes a while on my
40gb drives, but works. I:
mdadd /dev/md0 /dev/hde1 /dev/hdg1
mdrun -p1 /dev/md0
mount /dev/md0 /mnt/md0
All is well. I write files. I:
umount /mnt/md0
mdstop /dev/md0
shutdown -r now
All is fine. after booting:
mdadd /dev/md0 /dev/hde1 /dev/hdg1
mdrun -p1 /dev/md0
mount /dev/md0 /mnt/md0
Files are right where i left them. Here comes the problem.
umount /dev/md0
mdstop /dev/md0
<here /proc/mdstat shows md0 as inactive totally, no drives.>
I:
mdadd /dev/md0 /dev/hde1 /dev/hdg1
mdrun -p1 /dev/md0
All is fine (or so it seems). I:
mount /dev/md0 /mnt/md0
and it locks up. mount and kflushd both start sucking up a lot of
cpu and totally ignore kill -9 signals. I have to reboot to make
them stop. This makes my raid unrecoverable (these are preliminary
tests before actually putting my data on here), and right before it
reboots, it gives me zillions of messages like:
VFS: grow_buffers: size = 16384
VFS: grow_buffers: size = 16384
VFS: grow_buffers: size = 16384
VFS: grow_buffers: size = 16384
on the console and in the klogd file (prepended by '<4>'). every so
often in the klogd log it also outputs
384
which shows up like this in syslog:
Jul 24 14:16:19 chrome init: Switching to runlevel: 6
Jul 24 14:16:19 chrome kernel: 384
Jul 24 14:16:25 chrome last message repeated 27 times
Jul 24 14:16:25 chrome exiting on signal 15
Jul 24 14:17:14 chrome syslogd 1.3-3: restart.
ckraid takes 2+ hours on my two 40gb drives, even if they do operate
at full udma/66, not to mention i've been totally unable to recover
any data off of the drive after this happens, although i've yet to
try the dd then replace superblock trick... this seems just horribly
wrong for a seemingly standard setup with no extraordinary hardware
in the latest non-beta kernel...
As for the common recommendation of having backups, does anyone know
of an inexpensive way to backup 40gb? I bought these drives with the
hope of raiding them so that I would be protected from disk crashes,
but now I have a whole new bag of worries (superblock weirdness,
mis-syncing, massive filesystem corruption, etc) to keep me up at
night :P
-j
--
--------------------------------------------------
[EMAIL PROTECTED] - 0xCD91A427
9907 3747 3CE9 11C5 2B1C F141 D09F 488C CD91 A427
Note: key id 0x299450B6 is lost and inactive.
--------------------------------------------------
Copyright 2000 Jeffrey Paul.
The information contained in this message may be
privileged and confidential and protected from
disclosure. If the reader of this message is not
the intended recipient, or an employee or agent
responsible for delivering this message to the
intended recipient, you are hereby notified that
any dissemination, distribution or copying of this
communication is strictly prohibited. Thank you.
