Hi,

I have an old home server which ran out of disk space so I added a big disk over USB which I use for backup (mounted on /backup). Problem is that it has happened a few times that the USB controller has crapped out resulting in the disk being dismounted and the backup filling out /. I then have to reboot to be able to access the disk again (fsck/mount).

Is this due to my ancient HW, crappy USB controller card, or is USB mounted disks just less stable? 

/jkm

/var/log/messages:
Dec 17 03:12:17 maya /bsd: ehci0: unrecoverable error, controller halted
Dec 17 03:12:17 maya /bsd: ehci0: blocking intrs 0x10
Dec 17 03:12:17 maya /bsd: sd0 detached
Dec 17 03:12:17 maya /bsd: scsibus0 detached
Dec 17 03:12:17 maya /bsd: umass0 detached
Dec 17 03:12:23 maya /bsd: ehci_idone: ex=0xd0f7b700 is done!
Dec 17 03:12:56 maya last message repeated 6 times
Dec 17 03:12:56 maya /bsd: uhub0: device problem, disabling port 1

joakim@maya$ sudo fsck /dev/sd0a
Password:
Can't open /dev/rsd0a: Device not configured


OpenBSD 4.6-stable (GENERIC) #0: Sun Nov 29 21:19:21 CET 2009
    root@maya.aronius.com:/usr/src/sys/arch/i386/compile/GENERIC
cpu0: Intel Pentium III ("GenuineIntel" 686-class) 799 MHz
cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,MMX,FXSR,SSE
real mem  = 267743232 (255MB)
avail mem = 250073088 (238MB)
mainbus0 at root
bios0 at mainbus0: AT/286+ BIOS, date 02/22/00, BIOS32 rev. 0 @ 0xfda74, SMBIOS rev. 2.3 @ 0xf0ea0 (58 entries)
bios0: vendor Intel Corp. version "VC82010A.10A.0024.P03.0002221044" date 02/22/2000
bios0: Dell Computer Corporation XPS800
apm0 at bios0: Power Management spec V1.2
apm0: AC on, battery charge unknown
acpi at bios0 function 0x0 not configured
pcibios0 at bios0: rev 2.1 @ 0xf0000/0x10000
pcibios0: PCI IRQ Routing Table rev 1.0 @ 0xf3230/208 (11 entries)
pcibios0: PCI Interrupt Router at 000:31:0 ("Intel 82371FB ISA" rev 0x00)
pcibios0: PCI bus #2 is the last bus
bios0: ROM list: 0xc0000/0x8000 0xc8000/0x1000
cpu0 at ...
[ message continues ]

> I added a big disk over USB which I use for
 > backup (mounted on /backup).

Well don't do that!


Even if the externally powered USB disk never fails, still there is the 
chance that the two independently power supplied boxes get out of sync, 
for instance in the event of a power cycle...
You power cycle both of them: do they come back again in time to pick 
each other?
(I have a machine with a USB disk where this NEVER happens, when the 
bootstrapping OS tries to mount the USB disk, the disk is always 
spinning up or whatever but it's not ready yet so the mount fails)

-- 
Mauro Rezzonico <mauro@ch23.org>, Como, Italia
"Maybe this world is another planet's hell" - H.Huxley

[ message continues ]

And/Or wrap the backup script with something that checks for the mount.

Horrible shell script example:

FSMOUNT=`mount | grep -c backup`
if [ $FSMOUNT -ne "1" ]; then
        echo "Backup partition not mounted, aborting"
        exit 1
fi
if [ $FSMOUNT -eq "1" ]; then
        echo "Backup partition found, continuing"
fi

Amusing results of disk unmounting itself during the backup left to the reader.

-- 
Jon

[ message continues ]

Thanks, I will do both and throw in a mail notification when the disk is unavailable. (mounting under / was pretty stupid :)

Cheers,
/Joakim

[ message continues ]

can you please enlighten me on why that's a bad thing?

-- 
O< ascii ribbon campaign - stop html mail - www.asciiribbon.org

[ message continues ]

Filling up / can be more annoying than filling up /usr.

It's better to make sure your mounts work and not try to work around
broken systems, though.

[ message continues ]

This may or may not be the same problem, but...

I had a usb flash adaptor that would randomly error out with strange errors that would be different from time to time.  I threw it away and my next usb flash adaptor did the same thing.  Eventually I looked at it closer and realized that the usb cable was loose in the flash adaptor's usb plug.  I started using a higher quality usb cable and the problem went away.


-- 
The past cannot be changed.  The future cannot be guaranteed.

[ message continues ]

Hi,

An update for the internet archive..
The USB controller died on me (had to remove the PCI card to boot the machine) so the earlier problems were probably related to that.

Cheers,
/Joakim

/var/log/messages:
Jan  5 04:21:49 maya /bsd: ehci0: unrecoverable error, controller halted
Jan  5 04:21:49 maya /bsd: ehci0: blocking intrs 0x10
Jan  5 04:21:49 maya /bsd: uhci0: host controller process error
Jan  5 04:21:49 maya /bsd: uhci0: host controller halted
Jan  5 04:21:49 maya /bsd: sd0 detached
Jan  5 04:21:49 maya /bsd: scsibus0 detached
Jan  5 04:21:49 maya /bsd: umass0 detached
Jan  5 04:21:49 maya /bsd: uhci1: host system error
Jan  5 04:21:49 maya /bsd: uhci1: host controller process error
Jan  5 04:21:49 maya /bsd: uhci1: host controller halted
Jan  5 04:21:49 maya /bsd: uhci_freex: xfer=0xd0f36500 not busy, 0x4f4e5155
Jan  5 04:21:49 maya /bsd: uhci_freex: xfer=0xd0f36700 not busy, 0x4f4e5155
Jan  5 04:22:00 maya /bsd: ehci_sync_hc: tsleep() = 35
Jan  5 04:22:15 maya last message repeated 2 times
Jan  5 04:22:17 maya /bsd: ehci_idone: ex=0xd0f7b700 is done!
Jan  5 04:22:20 maya last message repeated 3 times
Jan  5 04:22:20 maya /bsd: ehci_sync_hc: tsleep() = 35
Jan  5 04:22:20 maya /bsd: ehci_idone: ex=0xd0f7b700 is done!
Jan  5 04:22:30 maya /bsd: ehci_sync_hc: tsleep() = 35
Jan  5 04:22:32 maya /bsd: ehci_idone: ex=0xd0f7b700 is done!
Jan  5 04:22:35 maya last message repeated 3 times
Jan  5 04:22:35 maya /bsd: ehci_sync_hc: tsleep() = 35
Jan  5 04:22:35 maya /bsd: ehci_idone: ex=0xd0f7b700 is done!
Jan  5 04:22:36 maya /bsd: ehci0: port reset timeout
Jan  5 04:22:36 maya /bsd: ehci_freex: xfer=0xd0f7b700 not busy, 0x4f4e5155
Jan  5 04:22:46 maya /bsd: ehci_sync_hc: tsleep() = 35
Jan  5 04:22:47 maya /bsd: ehci_idone: ex=0xd0f7b800 is done!
Jan  5 04:22:50 maya last message repeated 3 times
Jan  5 04:22:51 maya /bsd: ehci_sync_hc: tsleep() = 35
Jan  5 04:22:51 maya /bsd: ehci_idone: ex=0xd0f7b800 is done!
Jan  5 04:23:01 maya /bsd: ehci_sync_hc: tsleep() = ...
[ message continues ]