[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: disk problems: which ATA?



On Sun, 2011-07-03 at 12:07 -0700, JD wrote:
> On 07/03/2011 11:25 AM, Ross Boylan wrote: 
> > How can I tell which ata device is which hard drive?  It's come up
> > several times for me, most recently with
> > ata2.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
> > 
> > It appears from other info that sdb is the problem:
> > <terminal>
> > # smartctl -H /dev/sdb
> > Sun Jul  3 10:26:29 PDT 2011
> > smartctl version 5.38 [i686-pc-linux-gnu] Copyright (C) 2002-8 Bruce
> > Allen
> > Home page is http://smartmontools.sourceforge.net/
> > 
> > Short INQUIRY response, skip product id
> > A mandatory SMART command failed: exiting. To continue, add one or more
> > '-T permissive' options.
> > </terminal>
> > but I'd like to know how to do this in general.
> > 
> > I'd also welcome advice about the disk problems, but I was hoping to
> > raise the odds of an answer by keeping it simple :)
> > 
> > sdb has had hardware problems for awhile; it wouldn't be surprising if
> > it's failed.
> > 
> > Running lenny with linux 2.6.26-2-686.
> > 
> > Thanks.
> > Ross
> > 
> > 
> 
> My basic advice in this situation is:
> 1. unmount the failing drive if it is mounted, and mount it's
> partitions read only.
It's a bit tricky, since my mounts are mostly on LVM logical volumes.
The main volume group I use includes the failing disk and some good
disks.  Unfortunately, LVM (specifically pvs or vgdisplay) won't give me
info on the VG, even though I seem to have continuing access to many of
its LVs.  I *think* I've already moved most LVs to be backed by other
disks.
> 2. back up it's mounted partitions to a good drive using tar:
>     Assuming the mount points are
>     /sdb1  /sdb2    /sdb3
>     cd /
>     for i in 1 2 3; do
>         tar cf-   sdb$i  |  bzip2 -c -9 > /somewhere/backup/sdb$i
>     done
The latest problem emerged while doing my monthly (full) backups.  I
think the backups are good but the partition with the iso image failed.
> 3. If the drive is still under warranty, then run secure erase.
>     read up on hdparm (man hdparm)
>     If you go this route, I can send you a script
>     for proper use of hdparm to securely erase a drive.
>     After erasing, then get and RMA and send it away for
>     replacement.
> 
It was when it started to fail.  I was hoping to do some kind of LVM
migration to get the info off, and so things have dragged on.
> 4. If you purchase a new drive, you will then partition it
>     and tar back the backups to it.
Thanks for the advice.
> 
> -- To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org with
> a subject of "unsubscribe". Trouble? Contact
> listmaster@lists.debian.org Archive:
> 4E10BDE7.4040104@gmail.com">http://lists.debian.org/4E10BDE7.4040104@gmail.com


Reply to: