I/O error on a file on a backup drive

[Robert Moskowitz]

I am having a problem with rsyncing:

rsync -ah --stats --delete /home/ /media/HD103SI/backups/homebase/home
rsync: 
readlink_stat("/media/HD103SI/backups/homebase/home/abba/abba/Excel/Home/Power 
usage.xls") failed: Input/output error (5)
rsync: recv_generator: failed to stat 
"/media/HD103SI/backups/homebase/home/abba/abba/Excel/Home/Power 
usage.xls": Input/output error (5)
IO error encountered -- skipping file deletion

So I go to the directory and I try to ls P* and it hangs.

I tried to remove the bad file:

rm Power\ usage.xls
rm: cannot remove `Power usage.xls': Input/output error

How can I get this bad file out of the way so I can back up properly?

thanks

[Richard]

>  Date: Sunday, May 17, 2020 13:22:54 -0400
>  From: Robert Moskowitz <rgm(a)htt-consult.com&gt;
>
>  I am having a problem with rsyncing:
>  
>  rsync -ah --stats --delete /home/
>  /media/HD103SI/backups/homebase/home
>  rsync:
>  readlink_stat("/media/HD103SI/backups/homebase/home/abba/abba/Excel
>  /Home/Power usage.xls") failed: Input/output error (5)
>  rsync: recv_generator: failed to stat
>  "/media/HD103SI/backups/homebase/home/abba/abba/Excel/Home/Power
>  usage.xls": Input/output error (5)
>  IO error encountered -- skipping file deletion
>  
>  So I go to the directory and I try to ls P* and it hangs.
>  
>  I tried to remove the bad file:
>  
>  rm Power\ usage.xls
>  rm: cannot remove `Power usage.xls': Input/output error
>  
>  How can I get this bad file out of the way so I can back up
>  properly?
>   
I would probably fsck the drive to see if there are errors, and go
from there.

[George N. White III]

On Sun, 17 May 2020 at 14:54, Richard <lists-fedora(a)listmail.innovate.net&gt;
wrote:

>  > Date: Sunday, May 17, 2020 13:22:54 -0400
>  > From: Robert Moskowitz <rgm(a)htt-consult.com&gt;
>  >
>  > I am having a problem with rsyncing:
>  >
>  > rsync -ah --stats --delete /home/
>  > /media/HD103SI/backups/homebase/home
>  > rsync:
>  > readlink_stat("/media/HD103SI/backups/homebase/home/abba/abba/Excel
>  > /Home/Power usage.xls") failed: Input/output error (5)
>  > rsync: recv_generator: failed to stat
>  > "/media/HD103SI/backups/homebase/home/abba/abba/Excel/Home/Power
>  > usage.xls": Input/output error (5)
>  > IO error encountered -- skipping file deletion
>  >
>  > So I go to the directory and I try to ls P* and it hangs.
>  >
>  > I tried to remove the bad file:
>  >
>  > rm Power\ usage.xls
>  > rm: cannot remove `Power usage.xls': Input/output error
>  >
>  > How can I get this bad file out of the way so I can back up
>  > properly?
>  >
>
>  I would probably fsck the drive to see if there are errors, and go
>  from there.
>  
I worked in an IRIX64 shop for years.  IO errors could be due to
bad cables or termination on SCSI cables, but mostly disks that
were failing, so were telling us to replace the disk are restore
from backups.

According to the man page, "fsck.xfs" is just "exit 0".
I would first  check the disk's S.M.A.R.T status (smartmontools)
and then use "xfs_repair" (preferably on a ddrescue image -- note the
section of the xfs_repair man page on "Disk Errors").

-- 
George N. White III

[Cameron Simpson]

On 17May2020 15:11, George N. White III <gnwiii(a)gmail.com&gt; wrote:
> According to the man page, "fsck.xfs" is just "exit
> 0". 
It has has lead up logic to accept the correct command line arguments, 
but yes. The kernel "mount" for XFS does any repair. It is all very 
nice.

There is a separae "xfs_repair" command for very gnarly situations.

But Robert is using ext4.

Cheers,
Cameron Simpson <cs(a)cskk.id.au&gt;

[Robert Moskowitz]

On 5/17/20 1:52 PM, Richard wrote:
> > Date: Sunday, May 17, 2020 13:22:54 -0400
> > From: Robert Moskowitz <rgm(a)htt-consult.com&gt;
> >
> > I am having a problem with rsyncing:
> >
> > rsync -ah --stats --delete /home/
> > /media/HD103SI/backups/homebase/home
> > rsync:
> > readlink_stat("/media/HD103SI/backups/homebase/home/abba/abba/Excel
> > /Home/Power usage.xls") failed: Input/output error (5)
> > rsync: recv_generator: failed to stat
> > "/media/HD103SI/backups/homebase/home/abba/abba/Excel/Home/Power
> > usage.xls": Input/output error (5)
> > IO error encountered -- skipping file deletion
> >
> > So I go to the directory and I try to ls P* and it hangs.
> >
> > I tried to remove the bad file:
> >
> > rm Power\ usage.xls
> > rm: cannot remove `Power usage.xls': Input/output error
> >
> > How can I get this bad file out of the way so I can back up
> > properly?
> >
>  I would probably fsck the drive to see if there are errors, and go
>  from there. 
After umounting the drive from the server:

# fsck /dev/sdc
fsck from util-linux-ng 2.17.2
e2fsck 1.41.12 (17-May-2010)
fsck.ext2: Superblock invalid, trying backup blocks...
fsck.ext2: Bad magic number in super-block while trying to open /dev/sdc

The superblock could not be read or does not describe a correct ext2
filesystem.  If the device is valid and it really contains an ext2
filesystem (and not swap or ufs or something else), then the superblock
is corrupt, and you might try running e2fsck with an alternate superblock:
     e2fsck -b 8193 <device>

Yes, the server is running some old OS, but it is not accessible but on 
its subnet...

And

# parted /dev/sdc print
Model: SAMSUNG HD103SI (scsi)
Disk /dev/sdc: 1000GB
Sector size (logical/physical): 512B/512B
Partition Table: msdos

Number  Start   End     Size    Type     File system  Flags
  1      32.3kB  1000GB  1000GB  primary  ext4

[Jonathan Billings]

On May 17, 2020, at 14:14, Robert Moskowitz <rgm(a)htt-consult.com&gt; wrote:
>  
>  # fsck /dev/sdc
>  fsck from util-linux-ng 2.17.2
>  e2fsck 1.41.12 (17-May-2010)
>  fsck.ext2: Superblock invalid, trying backup blocks...
>  fsck.ext2: Bad magic number in super-block while trying to open /dev/sdc
>  
>  [...]
>  # parted /dev/sdc print
>  Model: SAMSUNG HD103SI (scsi)
>  Disk /dev/sdc: 1000GB
>  Sector size (logical/physical): 512B/512B
>  Partition Table: msdos
>  
>  Number  Start   End     Size    Type     File system  Flags
>   1      32.3kB  1000GB  1000GB  primary  ext4 
Try running fsck on the partition, not the whole disk device.  In other words, /dev/sdc1.


--
Jonathan Billings <billings(a)negate.org&gt;

[Robert Moskowitz]

On 5/17/20 3:32 PM, Jonathan Billings wrote:
>  On May 17, 2020, at 14:14, Robert Moskowitz
> <rgm(a)htt-consult.com&gt; wrote:
> > # fsck /dev/sdc
> > fsck from util-linux-ng 2.17.2
> > e2fsck 1.41.12 (17-May-2010)
> > fsck.ext2: Superblock invalid, trying backup blocks...
> > fsck.ext2: Bad magic number in super-block while trying to open /dev/sdc
> >
> > [...]
> > # parted /dev/sdc print
> > Model: SAMSUNG HD103SI (scsi)
> > Disk /dev/sdc: 1000GB
> > Sector size (logical/physical): 512B/512B
> > Partition Table: msdos
> >
> > Number  Start   End     Size    Type     File system  Flags
> >   1      32.3kB  1000GB  1000GB  primary  ext4
>  Try running fsck on the partition, not the whole disk device.  In other words, /dev/sdc1.

And now it is busy fixing things like:

File ... (inode #16521016, mod time Mon Jun 18 15:59:02 2007)
   has 14957 multiply-claimed block(s), shared with 5 file(s):
/repos/fedora/21/os/i386/Packages/p/PyQt4-doc-4.11.3-1.fc21.noarch.rpm 
(inode #20553615, mod time Wed Nov 12 21:53:30 2014)
/repos/fedora/21/os/i386/Packages/p/PyMca-data-4.7.3-2.fc21.noarch.rpm 
(inode #20553603, mod time Wed Aug 20 00:57:48 2014)
/repos/fedora/21/os/i386/Packages/p/pacemaker-1.1.12-2.fc21.i686.rpm 
(inode #20553639, mod time Wed Nov  5 11:46:53 2014)
/repos/fedora/21/os/i386/Packages/p/PySolFC-cardsets-2.0-9.fc21.noarch.rpm 
(inode #20553624, mod time Fri Jul 11 04:38:33 2014)
/repos/fedora/21/os/i386/Packages/p/PersonalCopy-Lite-soundfont-4.1-12.fc21.noarch.rpm 
(inode #20553591, mod time Fri Jul 11 05:08:14 2014)
Clone multiply-claimed blocks<y>? yes

File /backups/dumphere/Pics/Q3-2001/P1010003.JPG (inode #16521018, mod 
time Mon Jun 18 15:58:57 2007)
   has 53 multiply-claimed block(s), shared with 1 file(s):
     /Archive/Videos/Daf/09-rosh_hashana/rosh_hashana_21_video.zip 
(inode #58853708, mod time Sun Jun  1 21:59:32 2014)

Fixed - Re: I/O error on a file on a backup drive

[Robert Moskowitz]

I hope.

I basically jammed done the y key until it finished.  I am going to get 
a new drive...

So many errors.  I wonder how?  Perhaps 2 rsyncs running at the same time?



On 5/17/20 3:51 PM, Robert Moskowitz wrote:
>  On 5/17/20 3:32 PM, Jonathan Billings wrote:
> > On May 17, 2020, at 14:14, Robert Moskowitz <rgm(a)htt-consult.com&gt; wrote:
> >> # fsck /dev/sdc
> >> fsck from util-linux-ng 2.17.2
> >> e2fsck 1.41.12 (17-May-2010)
> >> fsck.ext2: Superblock invalid, trying backup blocks...
> >> fsck.ext2: Bad magic number in super-block while trying to open 
> >> /dev/sdc
> >>
> >> [...]
> >> # parted /dev/sdc print
> >> Model: SAMSUNG HD103SI (scsi)
> >> Disk /dev/sdc: 1000GB
> >> Sector size (logical/physical): 512B/512B
> >> Partition Table: msdos
> >>
> >> Number  Start   End     Size    Type     File system  Flags
> >>   1      32.3kB  1000GB  1000GB  primary  ext4
> > Try running fsck on the partition, not the whole disk device. In 
> > other words, /dev/sdc1.
>
>
>  And now it is busy fixing things like:
>
>  File ... (inode #16521016, mod time Mon Jun 18 15:59:02 2007)
>    has 14957 multiply-claimed block(s), shared with 5 file(s):
>  /repos/fedora/21/os/i386/Packages/p/PyQt4-doc-4.11.3-1.fc21.noarch.rpm 
>  (inode #20553615, mod time Wed Nov 12 21:53:30 2014)
>  /repos/fedora/21/os/i386/Packages/p/PyMca-data-4.7.3-2.fc21.noarch.rpm 
>  (inode #20553603, mod time Wed Aug 20 00:57:48 2014)
>  /repos/fedora/21/os/i386/Packages/p/pacemaker-1.1.12-2.fc21.i686.rpm 
>  (inode #20553639, mod time Wed Nov  5 11:46:53 2014)
>  /repos/fedora/21/os/i386/Packages/p/PySolFC-cardsets-2.0-9.fc21.noarch.rpm 
>  (inode #20553624, mod time Fri Jul 11 04:38:33 2014)
>  /repos/fedora/21/os/i386/Packages/p/PersonalCopy-Lite-soundfont-4.1-12.fc21.noarch.rpm 
>  (inode #20553591, mod time Fri Jul 11 05:08:14 2014)
>  Clone multiply-claimed blocks<y>? yes
>
>  File /backups/dumphere/Pics/Q3-2001/P1010003.JPG (inode #16521018, mod 
>  time Mon Jun 18 15:58:57 2007)
>    has 53 multiply-claimed block(s), shared with 1 file(s):
>      /Archive/Videos/Daf/09-rosh_hashana/rosh_hashana_21_video.zip 
>  (inode #58853708, mod time Sun Jun  1 21:59:32 2014)
>
>  _______________________________________________
>  users mailing list -- users(a)lists.fedoraproject.org
>  To unsubscribe send an email to users-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct: 
>  https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives: 
>  https://lists.fedoraproject.org/archives/list/[e-mail]

Fixed - Re: I/O error on a file on a backup drive

[George N. White III]

On Sun, 17 May 2020 at 17:30, Robert Moskowitz <rgm(a)htt-consult.com&gt; wrote:

>  I hope.
>
>  I basically jammed done the y key until it finished.  I am going to get
>  a new drive...
>
>  So many errors.  I wonder how?  Perhaps 2 rsyncs running at the same time?
>  
It is almost certainly a hardware problem: hard disk, cables, power supply,
etc.   The S.M.A.R.T diagnostics and self-tests  don't always catch failing
disks, but if they tell you there is a problem then you can reasonably
certain the other suspects are innocent.


>  On 5/17/20 3:51 PM, Robert Moskowitz wrote:
>  >
>  >
>  > On 5/17/20 3:32 PM, Jonathan Billings wrote:
>  >> On M ay 17, 2020, at 14:14, Robert Moskowitz <rgm(a)htt-consult.com&gt;
>  wrote:
>  >>> # fsck /dev/sdc
>  >>> fsck from util-linux-ng 2.17.2
>  >>> e2fsck 1.41.12 (17-May-2010)
>  >>> fsck.ext2: Superblock invalid, trying backup blocks...
>  >>> fsck.ext2: Bad magic number in super-block while trying to open
>  >>> /dev/sdc
>  >>>
>  >>> [...]
>  >>> # parted /dev/sdc print
>  >>> Model: SAMSUNG HD103SI (scsi)
>  >>> Disk /dev/sdc: 1000GB
>  >>> Sector size (logical/physical): 512B/512B
>  >>> Partition Table: msdos
>  >>>
>  >>> Number  Start   End     Size    Type     File system  Flags
>  >>>   1      32.3kB  1000GB  1000GB  primary  ext4
>  >> Try running fsck on the partition, not the whole disk device. In
>  >> other words, /dev/sdc1.
>  >
>  >
>  > And now it is busy fixing things like:
>  >
>  > File ... (inode #16521016, mod time Mon Jun 18 15:59:02 2007)
>  >   has 14957 multiply-claimed block(s), shared with 5 file(s):
>  > /repos/fedora/21/os/i386/Packages/p/PyQt4-doc-4.11.3-1.fc21.noarch.rpm
>  > (inode #20553615, mod time Wed Nov 12 21:53:30 2014)
>  > /repos/fedora/21/os/i386/Packages/p/PyMca-data-4.7.3-2.fc21.noarch.rpm
>  > (inode #20553603, mod time Wed Aug 20 00:57:48 2014)
>  > /repos/fedora/21/os/i386/Packages/p/pacemaker-1.1.12-2.fc21.i686.rpm
>  > (inode #20553639, mod time Wed Nov  5 11:46:53 2014)
>  >
>  /repos/fedora/21/os/i386/Packages/p/PySolFC-cardsets-2.0-9.fc21.noarch.rpm
>  > (inode #20553624, mod time Fri Jul 11 04:38:33 2014)
>  >
>  /repos/fedora/21/os/i386/Packages/p/PersonalCopy-Lite-soundfont-4.1-12.fc21.noarch.rpm
>
>  > (inode #20553591, mod time Fri Jul 11 05:08:14 2014)
>  > Clone multiply-claimed blocks<y>? yes
>  >
>  > File /backups/dumphere/Pics/Q3-2001/P1010003.JPG (inode #16521018, mod
>  > time Mon Jun 18 15:58:57 2007)
>  >   has 53 multiply-claimed block(s), shared with 1 file(s):
>  >     /Archive/Videos/Daf/09-rosh_hashana/rosh_hashana_21_video.zip
>  > (inode #58853708, mod time Sun Jun  1 21:59:32 2014)
>  >
>  > _______________________________________________
>  > users mailing list -- users(a)lists.fedoraproject.org
>  > To unsubscribe send an email to users-leave(a)lists.fedoraproject.org
>  > Fedora Code of Conduct:
>  > https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  > List Archives:
>  >
>  https://lists.fedoraproject.org/archives/list/[e-mail]
>  _______________________________________________
>  users mailing list -- users(a)lists.fedoraproject.org
>  To unsubscribe send an email to users-leave(a)lists.fedoraproject.org
>  Fedora Code of Conduct:
>  https://docs.fedoraproject.org/en-US/project/code-of-conduct/
>  List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
>  List Archives:
>  https://lists.fedoraproject.org/archives/list/[e-mail]
>  

-- 
George N. White III

Fixed - Re: I/O error on a file on a backup drive

[Samuel Sieb]

On 5/17/20 1:29 PM, Robert Moskowitz wrote:
>  I basically jammed done the y key until it finished.  I am going to
> get 
>  a new drive...
>  
>  So many errors.  I wonder how?  Perhaps 2 rsyncs running at the same time? 
No.  You could possibly get messed up files, although rsync is pretty 
safe.  But you should never be able to cause filesystem corruption from 
user level actions.

Fixed - Re: I/O error on a file on a backup drive

[Robert Moskowitz]

On 5/17/20 6:37 PM, Samuel Sieb wrote:
>  On 5/17/20 1:29 PM, Robert Moskowitz wrote:
> > I basically jammed done the y key until it finished.  I am going to 
> > get a new drive...
> >
> > So many errors.  I wonder how?  Perhaps 2 rsyncs running at the same 
> > time?
>
>  No.  You could possibly get messed up files, although rsync is pretty 
>  safe.  But you should never be able to cause filesystem corruption 
>  from user level actions. 
Well I am shopping for a USB ext HD...

At least 2TB, <$100, USB 2, Good life time ....   2,5" or 3.5"?

Well I will call the local Micro Center tomorrow and see what they have 
in stock and do a drive up pickup.

Fixed - Re: I/O error on a file on a backup drive

[Cameron Simpson]

On 17May2020 19:06, Robert Moskowitz <rgm(a)htt-consult.com&gt; wrote:
> Well I am shopping for a USB ext HD...
>
> At least 2TB, <$100, USB 2, Good life time ....   2,5" or 3.5"?
>
> Well I will call the local Micro Center tomorrow and see what they have 
> in stock and do a drive up pickup. 
I tend to buy Western Digital Elements or MyPassport. 2.5", USB3, bus 
powered.

I use them for backups and data transport, not hammering.

Cheers,
Cameron Simpson <cs(a)cskk.id.au&gt;

Fixed - Re: I/O error on a file on a backup drive

[Robert Moskowitz]

On 5/17/20 7:12 PM, Cameron Simpson wrote:
>  On 17May2020 19:06, Robert Moskowitz <rgm(a)htt-consult.com&gt;
> wrote:
> > Well I am shopping for a USB ext HD...
> >
> > At least 2TB, <$100, USB 2, Good life time ....   2,5" or 3.5"?
> >
> > Well I will call the local Micro Center tomorrow and see what they 
> > have in stock and do a drive up pickup.
>
>  I tend to buy Western Digital Elements or MyPassport. 2.5", USB3, bus 
>  powered.
>
>  I use them for backups and data transport, not hammering. 
The drive is permanently mounted, but only accessed at night with the 
rsyncs.  Or when I have to rescue an old version.

It is an old HP PC, USB2, and I am a little leery of using USB power.

Fixed - Re: I/O error on a file on a backup drive

[Cameron Simpson]

On 17May2020 19:38, Robert Moskowitz <rgm(a)htt-consult.com&gt; wrote:
> On 5/17/20 7:12 PM, Cameron Simpson wrote:
> >On 17May2020 19:06, Robert Moskowitz <rgm(a)htt-consult.com&gt; wrote:
> >>Well I am shopping for a USB ext HD...
> >>At least 2TB, <$100, USB 2, Good life time ....   2,5" or 3.5"?
> >>
> >>Well I will call the local Micro Center tomorrow and see what they 
> >>have in stock and do a drive up pickup.
> >
> >I tend to buy Western Digital Elements or MyPassport. 2.5", USB3, 
> >bus powered.
> >
> >I use them for backups and data transport, not hammering.
>
> The drive is permanently mounted, but only accessed at night with the 
> rsyncs.  Or when I have to rescue an old version.
>
> It is an old HP PC, USB2, and I am a little leery of using USB power. 
Aye. One of my backup pairs is usually mounted, like yours.

I recently bought a 7 port USB3 powered hub just to accomodate multiple 
external bus powered drives, for exactly the issues you're concerned 
about. A huge win. So the hub is always plugged into the server on one 
of the USB3 ports, and the drives come and go.

Cheers,
Cameron Simpson <cs(a)cskk.id.au&gt;

Fixed - Re: I/O error on a file on a backup drive

[Ed Greshko]

On 2020-05-18 07:06, Robert Moskowitz wrote:
>  On 5/17/20 6:37 PM, Samuel Sieb wrote:
> > On 5/17/20 1:29 PM, Robert Moskowitz wrote:
> >> I basically jammed done the y key until it finished.  I am going to get a new
> drive...
> >>
> >> So many errors.  I wonder how?  Perhaps 2 rsyncs running at the same time?
> >
> > No.  You could possibly get messed up files, although rsync is pretty safe.  But you
> should never be able to cause filesystem corruption from user level actions.
>
>  Well I am shopping for a USB ext HD...
>
>  At least 2TB, <$100, USB 2, Good life time ....   2,5" or 3.5"?
>
>  Well I will call the local Micro Center tomorrow and see what they have in stock and do a
> drive up pickup.
>  
FWIW, when it comes to backups I prefer not to have a single point of failure.  Yes, $100
is inexpensive but I
still rather, and have, shelled out about $450 for a 3TB 2-Bay NAS.  The drives themselves
are about $100.
And if one dies, you don't lose anything, and you just pop in a replacement.  Plus a
NAS, for me, is nice to
have and be able to share filesystems via NFS to various systems in the network.

The other advantage for a NAS is that many of them have addon packages like a bittorrent
client, DNS server,
DHCP server, OpenVPN server and client, etc.  This can be helpful for folks such as myself
who want those
extras but rather not be bothered with installing and configuring from scratch.  :-)

-- 
The key to getting good answers is to ask good questions.

Fixed - Re: I/O error on a file on a backup drive

[Robert Moskowitz]

On 5/17/20 7:33 PM, Ed Greshko wrote:
>  On 2020-05-18 07:06, Robert Moskowitz wrote:
> >
> > On 5/17/20 6:37 PM, Samuel Sieb wrote:
> >> On 5/17/20 1:29 PM, Robert Moskowitz wrote:
> >>> I basically jammed done the y key until it finished.  I am going to get a new
> drive...
> >>>
> >>> So many errors.  I wonder how?  Perhaps 2 rsyncs running at the same time?
> >> No.  You could possibly get messed up files, although rsync is pretty safe.  But
> you should never be able to cause filesystem corruption from user level actions.
> > Well I am shopping for a USB ext HD...
> >
> > At least 2TB, <$100, USB 2, Good life time ....   2,5" or 3.5"?
> >
> > Well I will call the local Micro Center tomorrow and see what they have in stock and
> do a drive up pickup.
> >
>  FWIW, when it comes to backups I prefer not to have a single point of failure.  Yes, $100
> is inexpensive but I
>  still rather, and have, shelled out about $450 for a 3TB 2-Bay NAS.  The drives
> themselves are about $100.
>  And if one dies, you don't lose anything, and you just pop in a replacement.  Plus a
> NAS, for me, is nice to
>  have and be able to share filesystems via NFS to various systems in the network.
>
>  The other advantage for a NAS is that many of them have addon packages like a bittorrent
> client, DNS server,
>  DHCP server, OpenVPN server and client, etc.  This can be helpful for folks such as
> myself who want those
>  extras but rather not be bothered with installing and configuring from scratch.  :-)
>  
That may be my next move, to get away from maintaining this server. But 
right now I have to have a trustable backup drive.

Fixed - Re: I/O error on a file on a backup drive

[Cameron Simpson]

On 17May2020 16:29, Robert Moskowitz <rgm(a)htt-consult.com&gt; wrote:
> I basically jammed done the y key until it finished.  I am going to
> get 
> a new drive... 
FYI, there's a BSD derived command called "yes" whose purpose in life is 
to print "yes" (or a string of your choice) forever, precisely for 
piping into other programmes.

I've always been a little afraid to use it, but it seems eminently 
suitable.

OTOH, so would fsck's -y option, if I understand your post.

Remember of course that fsck isn't fixing bad data, it is making the 
filesystem data structure self consistent. You will still have files 
with bad data, just not integrity errors at the fs level.

Also, get 2. I have 2 sets of backup drives for the home server.

Cheers,
Cameron Simpson <cs(a)cskk.id.au&gt;

[Gordon Messmer]

On 5/17/20 10:22 AM, Robert Moskowitz wrote:
>  How can I get this bad file out of the way so I can back up properly?
>  

IIRC, an IO error can result from either a corrupt filesystem or a media 
error.  The kernel "dmesg" can tell you more about the cause of the 
error, typically.  If there's filesystem corruption, you may be able to 
correct the problem with fsck and continue using the drive.  If there's 
a media error, you're probably better off replacing it.

[Robert Moskowitz]

On 5/17/20 2:13 PM, Gordon Messmer wrote:
>  On 5/17/20 10:22 AM, Robert Moskowitz wrote:
> > How can I get this bad file out of the way so I can back up properly? 
>
>
>  IIRC, an IO error can result from either a corrupt filesystem or a 
>  media error.  The kernel "dmesg" can tell you more about the cause of 
>  the error, typically.  If there's filesystem corruption, you may be 
>  able to correct the problem with fsck and continue using the drive.  
>  If there's a media error, you're probably better off replacing it. 
dmesg >dm.txt makes a 43K file.  I grepped on sdc and got things like:

sd 6:0:0:0: [sdc] Assuming drive cache: write through
sd 6:0:0:0: [sdc] Attached SCSI disk
sd 6:0:0:0: [sdc]  Result: hostbyte=DID_NO_CONNECT driverbyte=DRIVER_OK
sd 6:0:0:0: [sdc] CDB: Read(10): 28 00 3a 07 d5 cf 00 00 f0 00
end_request: I/O error, dev sdc, sector 973592015
sd 6:0:0:0: [sdc]  Result: hostbyte=DID_NO_CONNECT driverbyte=DRIVER_OK
sd 6:0:0:0: [sdc] CDB:
EXT4-fs (sdc1): error loading journal
end_request: I/O error, dev sdc, sector 973592255

sd 7:0:0:0: [sdc] Attached SCSI disk
EXT4-fs (sdc1): warning: mounting fs with errors, running e2fsck is 
recommended
EXT4-fs (sdc1): recovery complete
EXT4-fs (sdc1): mounted filesystem with ordered data mode. Opts:
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26386452
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 18481598
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 18482716
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 16520138
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 18481193
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 18745520
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 16520134
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 20709965
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156
EXT4-fs error (device sdc1): ext4_lookup: deleted inode referenced: 26098156


so what should I include in the e2fsck command?

And should probably order a new backup drive...

[Thomas Schmitt]

Hi,

Robert Moskowitz wrote:
>  sd 6:0:0:0: [sdc]  Result: hostbyte=DID_NO_CONNECT
> driverbyte=DRIVER_OK 
This is a command transport failure of the lowlevel driver. I know such
effects from dying USB boxes with optical drives (having SATA controllers).

 
https://www.tldp.org/HOWTO/archived/SCSI-Programming-HOWTO/SCSI-Programmi...
  "21.4 Host codes
   The following host codes are defined in drivers/scsi/scsi.h.
   They are set by the kernel driver.
   [...]
   0x01  | DID_NO_CONNECT | Couldn't connect before timeout period
   0x02  | DID_BUS_BUSY   | BUS stayed busy through time out period
   0x03  | DID_TIME_OUT   | TIMED OUT for other reason
   [...]
   21.5 Driver codes
   The midlevel driver categorizes the returned status from the lowlevel
   driver based on the sense key from the device. [...]
   [...]
   0x00  | DRIVER_OK      | No error
   "

There is no "sense key from the device", because the SCSI command could
not be transferred, according to hostbyte.

Despite the archived state of the SCSI HOWTO, this aspect is still very
present in the kernel:
  https://elixir.bootlin.com/linux/v5.6/ident/DID_NO_CONNECT
Clicking on some of the links there leads to bail-out code of drivers.


Have a nice day :)

Thomas