Discussion:
: Data recovery emergency on a downed server... Help please!!
(too old to reply)
Madison Kelly
2003-12-29 05:23:05 UTC
Permalink
Hi everyone,

Well, the fact that I am posting this after 12:00am Sunday night (and on
my birthday, to-boot!) should indicate just how dire the situation is...
I have had a drive failure on a server and I can no longer mount the
data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
the replacement server (Fedora Core 1) claims that the drive's
superblock is toast.

Now, before anyone rips me a new one, I -DO- have a tape drive but that,
too, has rather depressingly failed. When I try to recover the files
from the IDE Travan 8GB drive it successdully recovers about 50 files
and then fails with this...

-= Begin Error =-
[root-9CO/***@public.gmane.org tape]# tar zxvf /dev/ht0
.
[bunch-of-files]
.
home/OldC/Program Files/Common Files/Symantec
Shared/VirusDefs/19970902.001/navex15.exp
home/OldC/Program Files/Common Files/Symantec
Shared/VirusDefs/19970902.001/navex15.sys
home/OldC/Program Files/Common Files/Symantec
Shared/VirusDefs/19970902.001/NAVEX15.VXD

gzip: stdin: unexpected end of file
tar: Unexpected EOF in archive
tar: Unexpected EOF in archive
tar: Error is not recoverable: exiting now
[root-9CO/***@public.gmane.org tape]#
-= End Error =-

This is true on three of the tapes I have tried to recover from. The data
was backed up via this simple script:

-= Begin Script =-
mt -f /dev/ht0 erase
mt -f /dev/ht0 rewind
tar zcvf /dev/ht0 /home/*
-= End Script =-

I have been Googling for a while now with no success. If anyone has any
clues, by all means -please- share them... Also, if anyone here can
recommend (as a fall back) a data recovery house that won't balk at an
ext3 partition, please forward contact info.

Thanks everyone (again!!!)

The Ever-Stumbling Madison
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Teddy Mills
2003-12-29 07:00:38 UTC
Permalink
Either the drive had a hardware OR the superblock is toast. It doesnt seem
likely to be both.
From your description, it appears that the drive is functioning, but the
superblock is gone. Is that correct?

----- Original Message -----
From: "Madison Kelly" <linux-***@public.gmane.org>
To: <tlug-***@public.gmane.org>
Sent: Monday, December 29, 2003 12:23 AM
Subject: [TLUG]: Data recovery emergency on a downed server... Help please!!


> Hi everyone,
>
> Well, the fact that I am posting this after 12:00am Sunday night (and on
> my birthday, to-boot!) should indicate just how dire the situation is...
> I have had a drive failure on a server and I can no longer mount the
> data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
> the replacement server (Fedora Core 1) claims that the drive's
> superblock is toast.
>
> Now, before anyone rips me a new one, I -DO- have a tape drive but that,
> too, has rather depressingly failed. When I try to recover the files
> from the IDE Travan 8GB drive it successdully recovers about 50 files
> and then fails with this...
>
> -= Begin Error =-
> [root-9CO/***@public.gmane.org tape]# tar zxvf /dev/ht0
> .
> [bunch-of-files]
> .
> home/OldC/Program Files/Common Files/Symantec
> Shared/VirusDefs/19970902.001/navex15.exp
> home/OldC/Program Files/Common Files/Symantec
> Shared/VirusDefs/19970902.001/navex15.sys
> home/OldC/Program Files/Common Files/Symantec
> Shared/VirusDefs/19970902.001/NAVEX15.VXD
>
> gzip: stdin: unexpected end of file
> tar: Unexpected EOF in archive
> tar: Unexpected EOF in archive
> tar: Error is not recoverable: exiting now
> [root-9CO/***@public.gmane.org tape]#
> -= End Error =-
>
> This is true on three of the tapes I have tried to recover from. The data
> was backed up via this simple script:
>
> -= Begin Script =-
> mt -f /dev/ht0 erase
> mt -f /dev/ht0 rewind
> tar zcvf /dev/ht0 /home/*
> -= End Script =-
>
> I have been Googling for a while now with no success. If anyone has any
> clues, by all means -please- share them... Also, if anyone here can
> recommend (as a fall back) a data recovery house that won't balk at an
> ext3 partition, please forward contact info.
>
> Thanks everyone (again!!!)
>
> The Ever-Stumbling Madison
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>
>

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-29 07:08:17 UTC
Permalink
That would seem to be the case. However, I have used -many- of these
Seagates and I have never seen one suddenly pooch before. I am debating
running the Seagate test on it or decide to just recommend having it sent
to a data recovery house...

I know that backups of the superblock are made, do you know by chance how
to locate them and/or how to tell the OS to use one for mounting? Also, at
least two partitions on the drive have gone bad...

Madison

> Either the drive had a hardware OR the superblock is toast. It doesnt
> seem
> likely to be both.
>>From your description, it appears that the drive is functioning, but the
> superblock is gone. Is that correct?
>
> ----- Original Message -----
> From: "Madison Kelly" <linux-***@public.gmane.org>
> To: <tlug-***@public.gmane.org>
> Sent: Monday, December 29, 2003 12:23 AM
> Subject: [TLUG]: Data recovery emergency on a downed server... Help
> please!!
>
>
>> Hi everyone,
>>
>> Well, the fact that I am posting this after 12:00am Sunday night (and
>> on
>> my birthday, to-boot!) should indicate just how dire the situation is...
>> I have had a drive failure on a server and I can no longer mount the
>> data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
>> the replacement server (Fedora Core 1) claims that the drive's
>> superblock is toast.
>>
>> Now, before anyone rips me a new one, I -DO- have a tape drive but
>> that,
>> too, has rather depressingly failed. When I try to recover the files
>> from the IDE Travan 8GB drive it successdully recovers about 50 files
>> and then fails with this...
>>
>> -= Begin Error =-
>> [root-9CO/***@public.gmane.org tape]# tar zxvf /dev/ht0
>> .
>> [bunch-of-files]
>> .
>> home/OldC/Program Files/Common Files/Symantec
>> Shared/VirusDefs/19970902.001/navex15.exp
>> home/OldC/Program Files/Common Files/Symantec
>> Shared/VirusDefs/19970902.001/navex15.sys
>> home/OldC/Program Files/Common Files/Symantec
>> Shared/VirusDefs/19970902.001/NAVEX15.VXD
>>
>> gzip: stdin: unexpected end of file
>> tar: Unexpected EOF in archive
>> tar: Unexpected EOF in archive
>> tar: Error is not recoverable: exiting now
>> [root-9CO/***@public.gmane.org tape]#
>> -= End Error =-
>>
>> This is true on three of the tapes I have tried to recover from. The
>> data
>> was backed up via this simple script:
>>
>> -= Begin Script =-
>> mt -f /dev/ht0 erase
>> mt -f /dev/ht0 rewind
>> tar zcvf /dev/ht0 /home/*
>> -= End Script =-
>>
>> I have been Googling for a while now with no success. If anyone has any
>> clues, by all means -please- share them... Also, if anyone here can
>> recommend (as a fall back) a data recovery house that won't balk at an
>> ext3 partition, please forward contact info.
>>
>> Thanks everyone (again!!!)
>>
>> The Ever-Stumbling Madison
>> --
>> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
>> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
>> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>>
>>
>
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Anton Markov
2003-12-29 15:13:01 UTC
Permalink
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Madison,

First, if you have a spare hard drive, make a copy of the data on the
bad disk (or better yet two) in case the hardware will die, and work on
the copy.

Then do:

# /sbin/mke2fs -n /dev/hda5 -b [blocksize]
Be sure to use the right block size here!

You will see output including:

Superblock backups stored on blocks:
8193, 24577, 40961, 57345, 73729
These are the locations of the superblocks.

Pass these one by one to:

e2fsck -b [backup superblock location] -y /dev/hda5

[read up on "man e2fsck"]

If one of these superblocks are OK, e2fsck will start recovering all
possible data (the -y switch means "yes to all"; otherwise you get
thousands of prompts).

All the possible files will be dumped into the /lost+found directory.
- From there, you should be able to do something like:

find /lost+found/* -name [some directory you are sure of the location of]

i.e.
find /long+found/* -name anton
allowed me to find my home directory.

the /lost+found directory basicly contains many hard links to the same
files.

You should be able to find your /home, /var, /etc, or whatever other
directories you have on the drive and move them back to their respective
spot.

Be careful - many files may be corrupt without warning.


Using this technique, I was able to recover an ext3 partition after
deleting it, resizing the partition, reformatting as reiserfs, and using
it for two days. About 70% of the files got recovered, but it obviously
depends on the damage done.


Madison Kelly wrote:
> That would seem to be the case. However, I have used -many- of these
> Seagates and I have never seen one suddenly pooch before. I am debating
> running the Seagate test on it or decide to just recommend having it sent
> to a data recovery house...
>
> I know that backups of the superblock are made, do you know by chance how
> to locate them and/or how to tell the OS to use one for mounting? Also, at
> least two partitions on the drive have gone bad...
>
> Madison
>
>

- --
Anton Markov <("anton" + "@" + "truxtar" + "." + "com")>

GnuPG Key fingerprint =
5546 A6E2 1FFB 9BB8 15C3 CE34 46B7 8D93 3AD1 44B4

"The difference between insanity and genius is measured only by success."
- Some bad guy from 007
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQE/8ESERreNkzrRRLQRAhrqAJ99AdL8vA5Wcqgz/xUbiJIv9xKBtQCfWEc7
vur6ID9kQLMKqte0uPHFRZo=
=/aCT
-----END PGP SIGNATURE-----
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-29 15:51:45 UTC
Permalink
I do have an identical spare drive but how can I make an exact copy of a
drive with a messed up file system and also, how can I be sure that the
copy is non-destructive? If I can do this, it'd be wonderful!

Madison

PS - I got a dd of th last tape that ran (finally) and it is about 3.6GB
(about right). I extracted it as a tar.gz (which should work) but it fails
claiming to not be a gzip archive... Any ideas on that front by chance?

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi Madison,
>
> First, if you have a spare hard drive, make a copy of the data on the
> bad disk (or better yet two) in case the hardware will die, and work on
> the copy.
>
> Then do:
>
> # /sbin/mke2fs -n /dev/hda5 -b [blocksize]
> Be sure to use the right block size here!
>
> You will see output including:
>
> Superblock backups stored on blocks:
> 8193, 24577, 40961, 57345, 73729
> These are the locations of the superblocks.
>
> Pass these one by one to:
>
> e2fsck -b [backup superblock location] -y /dev/hda5
>
> [read up on "man e2fsck"]
>
> If one of these superblocks are OK, e2fsck will start recovering all
> possible data (the -y switch means "yes to all"; otherwise you get
> thousands of prompts).
>
> All the possible files will be dumped into the /lost+found directory.
> - From there, you should be able to do something like:
>
> find /lost+found/* -name [some directory you are sure of the location of]
>
> i.e.
> find /long+found/* -name anton
> allowed me to find my home directory.
>
> the /lost+found directory basicly contains many hard links to the same
> files.
>
> You should be able to find your /home, /var, /etc, or whatever other
> directories you have on the drive and move them back to their respective
> spot.
>
> Be careful - many files may be corrupt without warning.
>
>
> Using this technique, I was able to recover an ext3 partition after
> deleting it, resizing the partition, reformatting as reiserfs, and using
> it for two days. About 70% of the files got recovered, but it obviously
> depends on the damage done.
>
>
> Madison Kelly wrote:
>> That would seem to be the case. However, I have used -many- of these
>> Seagates and I have never seen one suddenly pooch before. I am debating
>> running the Seagate test on it or decide to just recommend having it
>> sent
>> to a data recovery house...
>>
>> I know that backups of the superblock are made, do you know by chance
>> how
>> to locate them and/or how to tell the OS to use one for mounting? Also,
>> at
>> least two partitions on the drive have gone bad...
>>
>> Madison
>>
>>
>
> - --
> Anton Markov <("anton" + "@" + "truxtar" + "." + "com")>
>
> GnuPG Key fingerprint =
> 5546 A6E2 1FFB 9BB8 15C3 CE34 46B7 8D93 3AD1 44B4
>
> "The difference between insanity and genius is measured only by success."
> - Some bad guy from 007
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.1 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
>
> iD8DBQE/8ESERreNkzrRRLQRAhrqAJ99AdL8vA5Wcqgz/xUbiJIv9xKBtQCfWEc7
> vur6ID9kQLMKqte0uPHFRZo=
> =/aCT
> -----END PGP SIGNATURE-----
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Anton Markov
2003-12-29 17:07:15 UTC
Permalink
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


You should use:

# dd if=/dev/hda5 of=/dev/hdxx bs=512 conv=noerror,sync

where xx is the drive and partition (of same size)

That should copy the data exactly, and the "noerror" option will ignore
physically corrupt areas (if any). You should first create the new
partition with /sbin/fdisk or /sbin/parted.

Here is one of the origional helpful documents I used when I had to
recover data:
<http://www.unixwiz.net/techtips/recovering-ext2.html>
Most of the article is about recovering the partition table, which you
probably don't need to do (what does "/sbin/fdisk -l" say?), but the
beginning is interesting.

If I ever find the other articles I've used (kick myself for forgetting
to bookmark, I'll post the links.


ps. I don't know how to fix your tape drive problems.


Madison Kelly wrote:
> I do have an identical spare drive but how can I make an exact copy of a
> drive with a messed up file system and also, how can I be sure that the
> copy is non-destructive? If I can do this, it'd be wonderful!
>
> Madison
>
> PS - I got a dd of th last tape that ran (finally) and it is about 3.6GB
> (about right). I extracted it as a tar.gz (which should work) but it fails
> claiming to not be a gzip archive... Any ideas on that front by chance?
>
>

- --
Anton Markov <("anton" + "@" + "truxtar" + "." + "com")>

GnuPG Key fingerprint =
5546 A6E2 1FFB 9BB8 15C3 CE34 46B7 8D93 3AD1 44B4

"The difference between insanity and genius is measured only by success."
- Some bad guy from 007
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQE/8F9MRreNkzrRRLQRAtyXAKCfRZpfTpcnmNG4TscVM4gSJ5yQ0gCfV0nF
t57QJMZzwChXsd9SXNY/Sv8=
=TCCS
-----END PGP SIGNATURE-----
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-29 17:49:43 UTC
Permalink
Hi Anton,

Thanks for the details. Does the destination partition need to be
exactly the same size? I am worried about trying to read the old drive...
Oh, and yes, I am sending from OE but that is because of a utility I needed
to use that spawned the default mail client which was, drum roll, a
server-hosted Eudora install. I had to set this up temporarily (and oh do I
mean temporarily!)

Anywho, I will be back shortly. I have to go downtown to get the spare
drive...

Madison

----- Original Message -----
From: "Anton Markov" <anton-F0u+***@public.gmane.org>
To: <tlug-***@public.gmane.org>
Sent: Monday, December 29, 2003 12:07 PM
Subject: Re: [TLUG]: Data recovery emergency on a downed server... Help
please!!


> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
> You should use:
>
> # dd if=/dev/hda5 of=/dev/hdxx bs=512 conv=noerror,sync
>
> where xx is the drive and partition (of same size)
>
> That should copy the data exactly, and the "noerror" option will ignore
> physically corrupt areas (if any). You should first create the new
> partition with /sbin/fdisk or /sbin/parted.
>
> Here is one of the origional helpful documents I used when I had to
> recover data:
> <http://www.unixwiz.net/techtips/recovering-ext2.html>
> Most of the article is about recovering the partition table, which you
> probably don't need to do (what does "/sbin/fdisk -l" say?), but the
> beginning is interesting.
>
> If I ever find the other articles I've used (kick myself for forgetting
> to bookmark, I'll post the links.
>
>
> ps. I don't know how to fix your tape drive problems.
>
>
> Madison Kelly wrote:
> > I do have an identical spare drive but how can I make an exact copy of a
> > drive with a messed up file system and also, how can I be sure that the
> > copy is non-destructive? If I can do this, it'd be wonderful!
> >
> > Madison
> >
> > PS - I got a dd of th last tape that ran (finally) and it is about 3.6GB
> > (about right). I extracted it as a tar.gz (which should work) but it
fails
> > claiming to not be a gzip archive... Any ideas on that front by chance?
> >
> >
>
> - --
> Anton Markov <("anton" + "@" + "truxtar" + "." + "com")>
>
> GnuPG Key fingerprint =
> 5546 A6E2 1FFB 9BB8 15C3 CE34 46B7 8D93 3AD1 44B4
>
> "The difference between insanity and genius is measured only by success."
> - Some bad guy from 007
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.1 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
>
> iD8DBQE/8F9MRreNkzrRRLQRAtyXAKCfRZpfTpcnmNG4TscVM4gSJ5yQ0gCfV0nF
> t57QJMZzwChXsd9SXNY/Sv8=
> =TCCS
> -----END PGP SIGNATURE-----
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Tim Writer
2003-12-29 18:51:59 UTC
Permalink
"Madison Kelly" <linux-***@public.gmane.org> writes:

> Hi Anton,
>
> Thanks for the details. Does the destination partition need to be
> exactly the same size?

No, but it can't be smaller. If it's substantially bigger and you're able to
recover successfully, you can resize the file system (e.g. using resize2fs)
to take advantage of the additional space.

--
tim writer <tim-s/***@public.gmane.org> starnix inc.
905.771.0017 ext. 225 thornhill, ontario, canada
http://www.starnix.com professional linux services & products
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Walter Dnes
2003-12-29 23:12:30 UTC
Permalink
On Mon, Dec 29, 2003 at 10:51:45AM -0500, Madison Kelly wrote

> PS - I got a dd of th last tape that ran (finally) and it is about 3.6GB
> (about right). I extracted it as a tar.gz (which should work) but it fails
> claiming to not be a gzip archive... Any ideas on that front by chance?

*MAKE A BACKUP COPY OF THE TAR.GZ FILE FIRST* !!!

Straight from the man page...
> tar -xvvzf foo.tar.gz
> extract gzipped foo.tar.gz
..did you remember the "z" parameter ?

I've had similar problems with tar.bz2 archives. I find that doing it
one step at a time works. Note, you will need a *LOT* more diskspace.

Step 1) zcat filename.tar.gz > filename.tar
The "zcat" executable is named "gzcat" in some versions. Use whichever
one works on your system.

Step 2) tar -xvf filename.tar

--
Walter Dnes <waltdnes-SLHPyeZ9y/***@public.gmane.org>
Email users are divided into two classes;
1) Those who have effective spam-blocking
2) Those who wish they did
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Tom Legrady
2003-12-29 23:30:07 UTC
Permalink
Don't use zcat to decompress a file ... use the appropriate tool, gunzip:

gunzip file.tar.gz
ls file.tar

NB. gunzip -c file.tar.gz is identical to zcat/


He already has a backup of the tar.gz ... it's called a tape

Tom

Walter Dnes wrote:

>On Mon, Dec 29, 2003 at 10:51:45AM -0500, Madison Kelly wrote
>
>
>
>>PS - I got a dd of th last tape that ran (finally) and it is about 3.6GB
>>(about right). I extracted it as a tar.gz (which should work) but it fails
>>claiming to not be a gzip archive... Any ideas on that front by chance?
>>
>>
>
> *MAKE A BACKUP COPY OF THE TAR.GZ FILE FIRST* !!!
>
> Straight from the man page...
>
>
>> tar -xvvzf foo.tar.gz
>> extract gzipped foo.tar.gz
>>
>>
>..did you remember the "z" parameter ?
>
> I've had similar problems with tar.bz2 archives. I find that doing it
>one step at a time works. Note, you will need a *LOT* more diskspace.
>
>Step 1) zcat filename.tar.gz > filename.tar
> The "zcat" executable is named "gzcat" in some versions. Use whichever
>one works on your system.
>
>Step 2) tar -xvf filename.tar
>
>
>

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-30 01:48:52 UTC
Permalink
Yep, still under OE (*shudder*)

Thanks for the suggestion, Walter. I did make a copy of the tar.gz file but
I got nothing useful out of it. I am still waiting now for the darn
activation code for a MS-based ext3-aware data recovery tool... Maybe that
will finally crack my rough luck thus far - if it ever arrives!

Tom, thanks for the note about gunzip -c being the same as zcat. I will make
another backup and give that a go, though my hopes are narrow... Oh, and
given the current state of my hardware; I certainly would not trust my tape
to make another dump so Walter's suggestion is/was valid.

Madison (who has passed the 24hours mark and wishes she could just get a
break and then go home!!)

----- Original Message -----
From: "Tom Legrady" <legrady-bJEeYj9oJeDQT0dZR+***@public.gmane.org>
To: <tlug-***@public.gmane.org>
Sent: Monday, December 29, 2003 6:30 PM
Subject: Re: [TLUG]: Data recovery emergency on a downed server... Help
please!!


> Don't use zcat to decompress a file ... use the appropriate tool, gunzip:
>
> gunzip file.tar.gz
> ls file.tar
>
> NB. gunzip -c file.tar.gz is identical to zcat/
>
>
> He already has a backup of the tar.gz ... it's called a tape
>
> Tom
>
> Walter Dnes wrote:
>
> >On Mon, Dec 29, 2003 at 10:51:45AM -0500, Madison Kelly wrote
> >
> >
> >
> >>PS - I got a dd of th last tape that ran (finally) and it is about 3.6GB
> >>(about right). I extracted it as a tar.gz (which should work) but it
fails
> >>claiming to not be a gzip archive... Any ideas on that front by chance?
> >>
> >>
> >
> > *MAKE A BACKUP COPY OF THE TAR.GZ FILE FIRST* !!!
> >
> > Straight from the man page...
> >
> >
> >> tar -xvvzf foo.tar.gz
> >> extract gzipped foo.tar.gz
> >>
> >>
> >..did you remember the "z" parameter ?
> >
> > I've had similar problems with tar.bz2 archives. I find that doing it
> >one step at a time works. Note, you will need a *LOT* more diskspace.
> >
> >Step 1) zcat filename.tar.gz > filename.tar
> > The "zcat" executable is named "gzcat" in some versions. Use whichever
> >one works on your system.
> >
> >Step 2) tar -xvf filename.tar
> >
> >
> >
>
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Byron Sonne
2003-12-30 02:48:28 UTC
Permalink
> Madison (who has passed the 24hours mark and wishes she could just get a
> break and then go home!!)

Please let us know how this all turns out. I'm quite interested in the
ultimate resolution! Sounds like a text book case of the typical failed
drive/bad tape nightmare scenario.

I've been through this kind of hell but it was when I was part of the
dark side and had to restore MS boxen, or attempt to, from bad
Veritas/BackupExec and/or Legato backups (though most of the time it
worked). Part of the overall procedure (which mind you wasn't followed
regularly for whatever reasons) was periodic restores to make sure
everything would work when called upon. Nothing like having a shelf full
of backups that turn out to be shitworthy in the end.

Oh; I would avoid CDs for backup... I've used numerous brands across
multiple writers and I'm finding that they are just not trustworthy
enough over time. Some people seem to have something against tape, but
I'd be far more inclined to trust DLT over CD. And keep them drives
clean! It does help.

My experience with using RAID (level 5 or similar levels of redundancy)
is that it obviates the need for 90% of the restores I would have had to
do in the past. Except, of course, people that want a restore done to
get an old version of a file back. RAID is not a panacea, but who gives
a shit about a busted drive or a 12 hour restore when you can just pop
in a new drive with nary a hiccup? If you got the money, go RAID and
complement it with a good backup system (I like DLT). Given a solid and
well designed IT infrastructure, using good stuff, backups should be
solid, reliable, and never, ever needed. Or for those with massive
amounts of data, go SAN.

While I'm rambling on, make sure you have clean electrical power,
preferably backed up with a UPS. Numerous restores I've done in the past
would have been avoided if the machine didn't hiccup or crash due to a
brownout or power failure. They don't have to keep your boxen running
for days on end, just enough time to close open files and back down
gracefully. They also help for restores and backups; rather annoying
when a long, long restore craps out midway through due to a power
glitch. Tapes will also be more reliable if your power is clean too, but
I suppose that is conjecture.
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-30 21:09:37 UTC
Permalink
Hi again Byron,

Oh, I would say I know a little about RAID... ;)

I don't ship a server without -at least- RAID 1 and the new
replacement server does indeed have two Seagate Barracuda 7200.7 SATA
120GB drives in software RAID 1 but alas the server that dies was
inherited and built on "budget" parts...

I am also implementing a tape rotation on a new DDS4 drive where
there is a tape for Mon-Thu then four Fri tapes to be rotated. Others
have recommended the same scheme like you, and I have personally
implemented this scheme for years and it works like a charm...

I'll post an update in a sec...

Madison

Byron Sonne wrote:
>> Madison (who has passed the 24hours mark and wishes she could just get a
>> break and then go home!!)
>
>
> Please let us know how this all turns out. I'm quite interested in the
> ultimate resolution! Sounds like a text book case of the typical failed
> drive/bad tape nightmare scenario.
>
> I've been through this kind of hell but it was when I was part of the
> dark side and had to restore MS boxen, or attempt to, from bad
> Veritas/BackupExec and/or Legato backups (though most of the time it
> worked). Part of the overall procedure (which mind you wasn't followed
> regularly for whatever reasons) was periodic restores to make sure
> everything would work when called upon. Nothing like having a shelf full
> of backups that turn out to be shitworthy in the end.
>
> Oh; I would avoid CDs for backup... I've used numerous brands across
> multiple writers and I'm finding that they are just not trustworthy
> enough over time. Some people seem to have something against tape, but
> I'd be far more inclined to trust DLT over CD. And keep them drives
> clean! It does help.
>
> My experience with using RAID (level 5 or similar levels of redundancy)
> is that it obviates the need for 90% of the restores I would have had to
> do in the past. Except, of course, people that want a restore done to
> get an old version of a file back. RAID is not a panacea, but who gives
> a shit about a busted drive or a 12 hour restore when you can just pop
> in a new drive with nary a hiccup? If you got the money, go RAID and
> complement it with a good backup system (I like DLT). Given a solid and
> well designed IT infrastructure, using good stuff, backups should be
> solid, reliable, and never, ever needed. Or for those with massive
> amounts of data, go SAN.
>
> While I'm rambling on, make sure you have clean electrical power,
> preferably backed up with a UPS. Numerous restores I've done in the past
> would have been avoided if the machine didn't hiccup or crash due to a
> brownout or power failure. They don't have to keep your boxen running
> for days on end, just enough time to close open files and back down
> gracefully. They also help for restores and backups; rather annoying
> when a long, long restore craps out midway through due to a power
> glitch. Tapes will also be more reliable if your power is clean too, but
> I suppose that is conjecture.


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
lsorense-1wCw9BSqJbv44Nm34jS7GywD8/ (Lennart Sorensen)
2004-01-05 20:45:50 UTC
Permalink
On Tue, Dec 30, 2003 at 04:09:37PM -0500, Madison Kelly wrote:
> Hi again Byron,
>
> Oh, I would say I know a little about RAID... ;)
>
> I don't ship a server without -at least- RAID 1 and the new
> replacement server does indeed have two Seagate Barracuda 7200.7 SATA
> 120GB drives in software RAID 1 but alas the server that dies was
> inherited and built on "budget" parts...
>
> I am also implementing a tape rotation on a new DDS4 drive where
> there is a tape for Mon-Thu then four Fri tapes to be rotated. Others
> have recommended the same scheme like you, and I have personally
> implemented this scheme for years and it works like a charm...

Make sure the backup script does a read verify on the tape too. In my
experience DDS is among the least reliable tapes around and don't last
very many uses at all before they start failing (unlike DLT and other
non helical scan tapes). I doubt I would ever use a helical scan tape
drive for a backup (not that I currently have any tape drives). The
head/tape wear is just too high.

> I'll post an update in a sec...

Lennart Sorensen
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Ilya Palagin
2003-12-30 04:25:11 UTC
Permalink
Madison Kelly wrote:

> Yep, still under OE (*shudder*)
>
> Thanks for the suggestion, Walter. I did make a copy of the tar.gz file but
> I got nothing useful out of it. I am still waiting now for the darn
> activation code for a MS-based ext3-aware data recovery tool... Maybe that
> will finally crack my rough luck thus far - if it ever arrives!
>
> Tom, thanks for the note about gunzip -c being the same as zcat. I will make
> another backup and give that a go, though my hopes are narrow... Oh, and
> given the current state of my hardware; I certainly would not trust my tape
> to make another dump so Walter's suggestion is/was valid.
>
> Madison (who has passed the 24hours mark and wishes she could just get a
> break and then go home!!)

For your daily backup operations, a recommend you to use star instead of
tar. star has a nice "diff" option, which allows to verify backups. Like
tar, it's multiplatform, so you can take your tapes to a friend with the
same tape drive and try to restore data on his system to make sure that
your backups make sence.

>
> ----- Original Message -----
> From: "Tom Legrady" <legrady-bJEeYj9oJeDQT0dZR+***@public.gmane.org>
> To: <tlug-***@public.gmane.org>
> Sent: Monday, December 29, 2003 6:30 PM
> Subject: Re: [TLUG]: Data recovery emergency on a downed server... Help
> please!!
>
>
>
>>Don't use zcat to decompress a file ... use the appropriate tool, gunzip:
>>
>>gunzip file.tar.gz
>>ls file.tar
>>
>>NB. gunzip -c file.tar.gz is identical to zcat/
>>
>>
>>He already has a backup of the tar.gz ... it's called a tape
>>
>>Tom
>>
>>Walter Dnes wrote:
>>
>>
>>>On Mon, Dec 29, 2003 at 10:51:45AM -0500, Madison Kelly wrote
>>>
>>>
>>>
>>>
>>>>PS - I got a dd of th last tape that ran (finally) and it is about 3.6GB
>>>>(about right). I extracted it as a tar.gz (which should work) but it
>
> fails
>
>>>>claiming to not be a gzip archive... Any ideas on that front by chance?
>>>>
>>>>
>>>
>>> *MAKE A BACKUP COPY OF THE TAR.GZ FILE FIRST* !!!
>>>
>>> Straight from the man page...
>>>
>>>
>>>
>>>> tar -xvvzf foo.tar.gz
>>>> extract gzipped foo.tar.gz
>>>>
>>>>
>>>
>>>..did you remember the "z" parameter ?
>>>
>>> I've had similar problems with tar.bz2 archives. I find that doing it
>>>one step at a time works. Note, you will need a *LOT* more diskspace.
>>>
>>>Step 1) zcat filename.tar.gz > filename.tar
>>> The "zcat" executable is named "gzcat" in some versions. Use whichever
>>>one works on your system.
>>>
>>>Step 2) tar -xvf filename.tar
>>>
>>>
>>>
>>
>>--
>>The Toronto Linux Users Group. Meetings: http://tlug.ss.org
>>TLUG requests: Linux topics, No HTML, wrap text below 80 columns
>>How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>
>
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-30 21:12:09 UTC
Permalink
Thank you for the info on 'star', I didn't know about that and it
definately is a bonus. Thanks!

Madison

Ilya Palagin wrote:
> Madison Kelly wrote:
>
>> Yep, still under OE (*shudder*)
>>
>> Thanks for the suggestion, Walter. I did make a copy of the tar.gz
>> file but
>> I got nothing useful out of it. I am still waiting now for the darn
>> activation code for a MS-based ext3-aware data recovery tool... Maybe
>> that
>> will finally crack my rough luck thus far - if it ever arrives!
>>
>> Tom, thanks for the note about gunzip -c being the same as zcat. I
>> will make
>> another backup and give that a go, though my hopes are narrow... Oh, and
>> given the current state of my hardware; I certainly would not trust my
>> tape
>> to make another dump so Walter's suggestion is/was valid.
>>
>> Madison (who has passed the 24hours mark and wishes she could just get a
>> break and then go home!!)
>
>
> For your daily backup operations, a recommend you to use star instead of
> tar. star has a nice "diff" option, which allows to verify backups. Like
> tar, it's multiplatform, so you can take your tapes to a friend with the
> same tape drive and try to restore data on his system to make sure that
> your backups make sence.
>
>>
>> ----- Original Message ----- From: "Tom Legrady" <legrady-bJEeYj9oJeDQT0dZR+***@public.gmane.org>
>> To: <tlug-***@public.gmane.org>
>> Sent: Monday, December 29, 2003 6:30 PM
>> Subject: Re: [TLUG]: Data recovery emergency on a downed server... Help
>> please!!
>>
>>
>>
>>> Don't use zcat to decompress a file ... use the appropriate tool,
>>> gunzip:
>>>
>>> gunzip file.tar.gz
>>> ls file.tar
>>>
>>> NB. gunzip -c file.tar.gz is identical to zcat/
>>>
>>>
>>> He already has a backup of the tar.gz ... it's called a tape
>>>
>>> Tom
>>>
>>> Walter Dnes wrote:
>>>
>>>
>>>> On Mon, Dec 29, 2003 at 10:51:45AM -0500, Madison Kelly wrote
>>>>
>>>>
>>>>
>>>>
>>>>> PS - I got a dd of th last tape that ran (finally) and it is about
>>>>> 3.6GB
>>>>> (about right). I extracted it as a tar.gz (which should work) but it
>>
>>
>> fails
>>
>>>>> claiming to not be a gzip archive... Any ideas on that front by
>>>>> chance?
>>>>>
>>>>>
>>>>
>>>> *MAKE A BACKUP COPY OF THE TAR.GZ FILE FIRST* !!!
>>>>
>>>> Straight from the man page...
>>>>
>>>>
>>>>
>>>>> tar -xvvzf foo.tar.gz
>>>>> extract gzipped foo.tar.gz
>>>>>
>>>>>
>>>>
>>>> ..did you remember the "z" parameter ?
>>>>
>>>> I've had similar problems with tar.bz2 archives. I find that doing it
>>>> one step at a time works. Note, you will need a *LOT* more diskspace.
>>>>
>>>> Step 1) zcat filename.tar.gz > filename.tar
>>>> The "zcat" executable is named "gzcat" in some versions. Use whichever
>>>> one works on your system.
>>>>
>>>> Step 2) tar -xvf filename.tar


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
John Vetterli
2003-12-31 14:47:09 UTC
Permalink
> Ilya Palagin wrote:
> > For your daily backup operations, a recommend you to use star instead of
> > tar. star has a nice "diff" option, which allows to verify backups. Like

BTW, GNU tar has a diff option as well (tar df <filename>).

JV
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Robert Brockway
2003-12-30 14:37:45 UTC
Permalink
On Mon, 29 Dec 2003, Madison Kelly wrote:

Hi Maddy. I hope you got some sleep last night...

> I know that backups of the superblock are made, do you know by chance how
> to locate them and/or how to tell the OS to use one for mounting? Also, at

No one mentioned this (and I didn't read TLUG yesterday) so here it is:

/sbin/dumpe2fs /dev/hdxy (where xy is a1, b2, etc) will give you this
info. There is quite a lot of output so redirect it to a file and read it
with less, eg:

/sbin/dumpe2fs /dev/hde3 > /tmp/hde3.info

Search for the word "Superblock"

Now, take that number and use it with e2fsck -b,

Eg I have "Superblock at 163840" listed for /dev/hde3, so I could try:

e2fsck -b 163840 /dev/hde3.

Remember to only work on backup copies! See below.

> least two partitions on the drive have gone bad...

Different filesystems (partitions) being bad makes it far less likely it
is a filesystem problem.

Also, the use of dd was mentioned. This is good but there is no need to
write it to a physical drive. You can write the dd to a file and mount it
using loopback and attempt recovery on that. As long as you keep a good
copy of the filesystem file around you can keep trying things over and
over on copies.

Good luck.

Rob

--
Robert Brockway
Senior Technical Consultant, OpenTrend Solutions Ltd.
Phone: 416-669-3073, Email: rbrockway-wgAaPJgzrDxH4x6Dk/***@public.gmane.org, http://www.opentrend.net
OpenTrend Solutions: Reliable, secure solutions to real world problems.
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-30 21:13:19 UTC
Permalink
I have passed the drive off to a data recovery house but if the drive is
recovered and the drive is still intact when it's all over I will try
this out for personal practice. Thank you!!

Madison

Robert Brockway wrote:
> On Mon, 29 Dec 2003, Madison Kelly wrote:
>
> Hi Maddy. I hope you got some sleep last night...
>
>
>>I know that backups of the superblock are made, do you know by chance how
>>to locate them and/or how to tell the OS to use one for mounting? Also, at
>
>
> No one mentioned this (and I didn't read TLUG yesterday) so here it is:
>
> /sbin/dumpe2fs /dev/hdxy (where xy is a1, b2, etc) will give you this
> info. There is quite a lot of output so redirect it to a file and read it
> with less, eg:
>
> /sbin/dumpe2fs /dev/hde3 > /tmp/hde3.info
>
> Search for the word "Superblock"
>
> Now, take that number and use it with e2fsck -b,
>
> Eg I have "Superblock at 163840" listed for /dev/hde3, so I could try:
>
> e2fsck -b 163840 /dev/hde3.
>
> Remember to only work on backup copies! See below.
>
>
>>least two partitions on the drive have gone bad...
>
>
> Different filesystems (partitions) being bad makes it far less likely it
> is a filesystem problem.
>
> Also, the use of dd was mentioned. This is good but there is no need to
> write it to a physical drive. You can write the dd to a file and mount it
> using loopback and attempt recovery on that. As long as you keep a good
> copy of the filesystem file around you can keep trying things over and
> over on copies.
>
> Good luck.
>
> Rob


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
John Vetterli
2003-12-29 15:37:16 UTC
Permalink
Have you tried using debugfs (part of e2fsprogs)? It lets you tinker
with the superblock manually.

Also: happy birthday.

JV

On Mon, 29 Dec 2003, Madison Kelly wrote:
> Well, the fact that I am posting this after 12:00am Sunday night (and on
> my birthday, to-boot!) should indicate just how dire the situation is...
> I have had a drive failure on a server and I can no longer mount the
> data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
> the replacement server (Fedora Core 1) claims that the drive's
> superblock is toast.
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-29 15:53:33 UTC
Permalink
I haven't yet because I am pretty worried about doing anything potentially
destructive to the data...

Madison (who's been at it now for over 15h... )

> Have you tried using debugfs (part of e2fsprogs)? It lets you tinker
> with the superblock manually.
>
> Also: happy birthday.
>
> JV
>
> On Mon, 29 Dec 2003, Madison Kelly wrote:
>> Well, the fact that I am posting this after 12:00am Sunday night (and
>> on
>> my birthday, to-boot!) should indicate just how dire the situation is...
>> I have had a drive failure on a server and I can no longer mount the
>> data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
>> the replacement server (Fedora Core 1) claims that the drive's
>> superblock is toast.
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
David J Patrick
2003-12-30 03:48:26 UTC
Permalink
Madison Kelly wrote:

>Hi everyone,
>
> Well, the fact that I am posting this after 12:00am Sunday night (and on
>my birthday, to-boot!) should indicate just how dire the situation is...
>I have had a drive failure on a server and I can no longer mount the
>data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
>the replacement server (Fedora Core 1) claims that the drive's
>superblock is toast.
>
have you considered a boot linux (like knoppix) in order to move trapped
data and/ or assess damage ?
works for me, YYMV
djp

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-30 21:41:17 UTC
Permalink
First and foremost; THANK YOU to everyone who has helped me out on
this!! I decided yesterday to try "Stellar Pheonix Linux" which is a
MS-based utility someone here (thank you!) that can attempt data
recovery of ext2 and ext3 partitions. It discovered a little over 5,000
files and then Win2k lost the drive entirely. At that, I decided it was
time to send it to a data recovery house which I did today.

Given the value of the data I decided that I had reached the limit of my
(with your help!) expertise. If interested, I will let you know how much
(if any) data is saved.

Thank you all again!

Madison

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Jason Shein
2004-01-07 17:13:54 UTC
Permalink
came across this article today

System recovery with Knoppix
What to do when good disks go bad
Level: Intermediate

http://www-106.ibm.com/developerworks/linux/library/l-knopx.html?ca=dgr-lnxw10Knoppix

some useful information.

Madison Kelly wrote:
> First and foremost; THANK YOU to everyone who has helped me out on
> this!! I decided yesterday to try "Stellar Pheonix Linux" which is a
> MS-based utility someone here (thank you!) that can attempt data
> recovery of ext2 and ext3 partitions. It discovered a little over 5,000
> files and then Win2k lost the drive entirely. At that, I decided it was
> time to send it to a data recovery house which I did today.
>
> Given the value of the data I decided that I had reached the limit of my
> (with your help!) expertise. If interested, I will let you know how much
> (if any) data is saved.
>
> Thank you all again!
>
> Madison
>
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>

--
" Eventually people tire of repairing broken Windows,
And decide to replace them with something stronger"
(o_
//\ Linux - The Choice Of A GNU Generation
V_/_ Jason Shein
Linux Registered User #281100
jason-gaRZxGPHtpBxZtjKW1aY+1aTQe2KTcn/@public.gmane.org
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Fraser Campbell
2004-01-07 18:01:02 UTC
Permalink
On Wednesday 07 January 2004 12:13, Jason Shein wrote:
> came across this article today
>
> System recovery with Knoppix
> What to do when good disks go bad

Knoppix is an awesome way to recover from failed disks. Especially handy if
your backups are on an amanda server. Amanda utilities so recovery is
boot->partition->make filesystems/raid/lvm/etc->extract backup->install
bootloader->reboot->go home.

--
Fraser Campbell <fraser-Txk5XLRqZ6CsTnJN9+***@public.gmane.org> http://www.wehave.net/
Georgetown, Ontario, Canada Debian GNU/Linux

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Tim Writer
2004-01-07 20:10:30 UTC
Permalink
Fraser Campbell <fraser-Txk5XLRqZ6CsTnJN9+***@public.gmane.org> writes:

> On Wednesday 07 January 2004 12:13, Jason Shein wrote:
> > came across this article today
> >
> > System recovery with Knoppix
> > What to do when good disks go bad
>
> Knoppix is an awesome way to recover from failed disks. Especially handy if
> your backups are on an amanda server. Amanda utilities so recovery is
> boot->partition->make filesystems/raid/lvm/etc->extract backup->install
> bootloader->reboot->go home.

And with mdadm and vgscan (both of which are included), it can recover LVM
and RAID volumes directly from the disk content, i.e. without needing any
auxiliary info such as /etc/raidtab.

--
tim writer <tim-s/***@public.gmane.org> starnix inc.
905.771.0017 ext. 225 thornhill, ontario, canada
http://www.starnix.com professional linux services & products
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Kevin Cozens
2003-12-31 17:31:15 UTC
Permalink
At 12:23 AM 12/29/2003 -0500, Madison wrote:
>I have had a drive failure on a server and I can no longer mount the
>data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
>the replacement server (Fedora Core 1) claims that the drive's
>superblock is toast.

Sorry to hear about your hard drive and tape drive problems. Hopefully the
recovery service will be able to help you. As was suggested earlier I was
wondering if using one of the other superblocks would have helped you
recover the files if the primary superblock was bad.

> Now, before anyone rips me a new one, I -DO- have a tape drive but that,
>too, has rather depressingly failed. When I try to recover the files
>from the IDE Travan 8GB drive it successdully recovers about 50 files
>and then fails with this...

I don't have anything against using tape for backups but I do have
something against tape drives. I have had four tape drives die on me.
Actually, it was three drives. The fourth drive was made up of the two main
pieces of two other identical drives that had previously failed in
different ways. None of the drives had seen heavy usage. In fact, they had
each been used for barely a dozen or so backups (IIRC).

This is why I am now looking at the use of DVDs for backup. I already have
software that allows me to backup everything in my Windows side of the
machine to DVD. It backed up a bit over 6G on to two DVDs and the bonus is
that they are bootable so I can recover from bare metal directly from the
DVDs. I'm hoping to eventually be able to do something similar under Linux.

One other difference between the use of tape and CD/DVD for backup is that
the CD/DVD drives (AFAIK) are a lot cheaper to buy/replace should something
happen to the drive. There is a place I know of which can repair the tape
drives I have at a flat fee of $350 per drive. Only problem is, I don't
have confidence that a repaired drive would last any longer then the new
ones did.


Cheers!

Kevin. (http://www.interlog.com/~kcozens/)

Owner of Elecraft K2 #2172 |"What are we going to do today, Borg?"
E-mail:kcozens at interlog dot com|"Same thing we always do, Pinkutus:
Packet:ve3syb-XXPEJ3/***@public.gmane.org#con.on.ca.na| Try to assimilate the world!"
#include <disclaimer/favourite> | -Pinkutus & the Borg

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2003-12-31 17:48:42 UTC
Permalink
I have been hearing over and over again how tape drives are far less
reliable than I once believed... I am now giving serious thought to
alternative backup schemes but I -have- to find something that will be
easy to maintain lest user apathy kick in and the backups stop occuring.

So far, I have helped negate the nee for a future recovery a certain
amount by implementing a RAID1 mirror however that will do nothing to
prevent more than a single disk failure. I have thought about DVD-/+RW
drives but that would require more effort on the part of the users; see
above fear. Finally, I have been thinking about getting a collection of
laptop-size HDD and placing them into external USB2.0-powered chassis
and writting a script to do the copy. This would allow for a simple swap
of a chassis each day and would be a lot faster, to boot. The only fear
there is the initial setup cost.

As for the superblock; if I was more confident in my own abilities I
would have tried exactly that. The problem is that the last PC actually
lost the drive indicating a hardware failure. Given the value of the
data, I was not willing to risk any further proding by me. If the data
is recovered and the original drive is still intact than for practice I
will try just that.

Again, I can't express enough how much I appreciate all the help
everyone has given me in this!!

Madison (who has finally gotten herself a good night's sleep!!)

Kevin Cozens wrote:
> At 12:23 AM 12/29/2003 -0500, Madison wrote:
>
>> I have had a drive failure on a server and I can no longer mount the
>> data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
>> the replacement server (Fedora Core 1) claims that the drive's
>> superblock is toast.
>
>
> Sorry to hear about your hard drive and tape drive problems. Hopefully
> the recovery service will be able to help you. As was suggested earlier
> I was wondering if using one of the other superblocks would have helped
> you recover the files if the primary superblock was bad.
>
>> Now, before anyone rips me a new one, I -DO- have a tape drive but
>> that,
>> too, has rather depressingly failed. When I try to recover the files
>> from the IDE Travan 8GB drive it successdully recovers about 50 files
>> and then fails with this...
>
>
> I don't have anything against using tape for backups but I do have
> something against tape drives. I have had four tape drives die on me.
> Actually, it was three drives. The fourth drive was made up of the two
> main pieces of two other identical drives that had previously failed in
> different ways. None of the drives had seen heavy usage. In fact, they
> had each been used for barely a dozen or so backups (IIRC).
>
> This is why I am now looking at the use of DVDs for backup. I already
> have software that allows me to backup everything in my Windows side of
> the machine to DVD. It backed up a bit over 6G on to two DVDs and the
> bonus is that they are bootable so I can recover from bare metal
> directly from the DVDs. I'm hoping to eventually be able to do something
> similar under Linux.
>
> One other difference between the use of tape and CD/DVD for backup is
> that the CD/DVD drives (AFAIK) are a lot cheaper to buy/replace should
> something happen to the drive. There is a place I know of which can
> repair the tape drives I have at a flat fee of $350 per drive. Only
> problem is, I don't have confidence that a repaired drive would last any
> longer then the new ones did.
>
>
> Cheers!
>
> Kevin. (http://www.interlog.com/~kcozens/)


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Wil McGilvery
2003-12-31 17:55:58 UTC
Permalink
I find DAT tapes to be a waste of time and energy. I do find DLT and Ultruim tapes to be a reliable source for backing up data. These are more expensive that dat tapes but you get what you pay for.

Backups in my opinion will never be a set it and forget it procedure. Periodic restores are necessary for peace of and the assurance that you can restore when necessary. It is also important to regularly clean your tape drives.

I have not heard of anyone successfully using cd's for backup purposes, but I do know of people using removable hard drives as an alternate to tape backups.

Good Luck.

Regards,

Wil McGilvery
Manager
Lynch Digital Media Inc



416-744-7949
416-716-3964 (cell)
1-866-314-4678
416-744-0406  FAX
www.LynchDigital.com


-----Original Message-----
From: Madison Kelly [mailto:linux-***@public.gmane.org]
Sent: Wednesday, December 31, 2003 12:49 PM
To: tlug-***@public.gmane.org
Subject: Re: [TLUG]: Data recovery emergency on a downed server... Help please!!

I have been hearing over and over again how tape drives are far less
reliable than I once believed... I am now giving serious thought to
alternative backup schemes but I -have- to find something that will be
easy to maintain lest user apathy kick in and the backups stop occuring.

So far, I have helped negate the nee for a future recovery a certain
amount by implementing a RAID1 mirror however that will do nothing to
prevent more than a single disk failure. I have thought about DVD-/+RW
drives but that would require more effort on the part of the users; see
above fear. Finally, I have been thinking about getting a collection of
laptop-size HDD and placing them into external USB2.0-powered chassis
and writting a script to do the copy. This would allow for a simple swap
of a chassis each day and would be a lot faster, to boot. The only fear
there is the initial setup cost.

As for the superblock; if I was more confident in my own abilities I
would have tried exactly that. The problem is that the last PC actually
lost the drive indicating a hardware failure. Given the value of the
data, I was not willing to risk any further proding by me. If the data
is recovered and the original drive is still intact than for practice I
will try just that.

Again, I can't express enough how much I appreciate all the help
everyone has given me in this!!

Madison (who has finally gotten herself a good night's sleep!!)

Kevin Cozens wrote:
> At 12:23 AM 12/29/2003 -0500, Madison wrote:
>
>> I have had a drive failure on a server and I can no longer mount the
>> data partition on the hard drive (/dev/hda5 [ext3 under RH7.3]) because
>> the replacement server (Fedora Core 1) claims that the drive's
>> superblock is toast.
>
>
> Sorry to hear about your hard drive and tape drive problems. Hopefully
> the recovery service will be able to help you. As was suggested earlier
> I was wondering if using one of the other superblocks would have helped
> you recover the files if the primary superblock was bad.
>
>> Now, before anyone rips me a new one, I -DO- have a tape drive but
>> that,
>> too, has rather depressingly failed. When I try to recover the files
>> from the IDE Travan 8GB drive it successdully recovers about 50 files
>> and then fails with this...
>
>
> I don't have anything against using tape for backups but I do have
> something against tape drives. I have had four tape drives die on me.
> Actually, it was three drives. The fourth drive was made up of the two
> main pieces of two other identical drives that had previously failed in
> different ways. None of the drives had seen heavy usage. In fact, they
> had each been used for barely a dozen or so backups (IIRC).
>
> This is why I am now looking at the use of DVDs for backup. I already
> have software that allows me to backup everything in my Windows side of
> the machine to DVD. It backed up a bit over 6G on to two DVDs and the
> bonus is that they are bootable so I can recover from bare metal
> directly from the DVDs. I'm hoping to eventually be able to do something
> similar under Linux.
>
> One other difference between the use of tape and CD/DVD for backup is
> that the CD/DVD drives (AFAIK) are a lot cheaper to buy/replace should
> something happen to the drive. There is a place I know of which can
> repair the tape drives I have at a flat fee of $350 per drive. Only
> problem is, I don't have confidence that a repaired drive would last any
> longer then the new ones did.
>
>
> Cheers!
>
> Kevin. (http://www.interlog.com/~kcozens/)


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Terrence Enger
2003-12-31 18:31:45 UTC
Permalink
At 12:55 2003-12-31 -0500, Wil McGilvery wrote:
>
> I have not heard of anyone successfully using cd's for backup purposes,
but I do know of people using removable hard drives as an alternate to tape
backups.
>

I have been using CDs for a while now. Quite frankly, the
cost of writer and media was the motivating factor, and I
thought that tape would be "better". However, the earlier
comments here make me think maybe I made a lucky choice.

The backup CDs carried me through one recovery after I hosed
my system quite thoroughly, and every month or so they save
me after a slip of the brain. But, now that you mention it,
it *is* time to test complete recovery again.

BTW, I have Windows and Linux machines at my desk. I find
myself moveing more and more to the Linux box just because
my Linux backup procedure is easy enough that I can run it
frequently.

Terry.


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Wil McGilvery
2003-12-31 18:41:26 UTC
Permalink
One success story is better than none :)

Thanks for the info.

Regards,

Wil McGilvery
Manager
Lynch Digital Media Inc



416-744-7949
416-716-3964 (cell)
1-866-314-4678
416-744-0406  FAX
www.LynchDigital.com


-----Original Message-----
From: Terrence Enger [mailto:tenger-ew0EfhANLmVEfu+***@public.gmane.org]
Sent: Wednesday, December 31, 2003 1:32 PM
To: tlug-***@public.gmane.org
Subject: RE: [TLUG]: Data recovery emergency on a downed server... Help please!!

At 12:55 2003-12-31 -0500, Wil McGilvery wrote:
>
> I have not heard of anyone successfully using cd's for backup purposes,
but I do know of people using removable hard drives as an alternate to tape
backups.
>

I have been using CDs for a while now. Quite frankly, the
cost of writer and media was the motivating factor, and I
thought that tape would be "better". However, the earlier
comments here make me think maybe I made a lucky choice.

The backup CDs carried me through one recovery after I hosed
my system quite thoroughly, and every month or so they save
me after a slip of the brain. But, now that you mention it,
it *is* time to test complete recovery again.

BTW, I have Windows and Linux machines at my desk. I find
myself moveing more and more to the Linux box just because
my Linux backup procedure is easy enough that I can run it
frequently.

Terry.


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Kevin Cozens
2003-12-31 21:36:08 UTC
Permalink
At 12:55 PM 12/31/2003 -0500, Wil McGilvery wrote:
>Backups in my opinion will never be a set it and forget it procedure.
>Periodic restores are necessary for peace of and the assurance that you
>can restore when necessary.

Tape and CD/DVD changers help but if the data is critical, someone still
has to physically move the backups off-site and put fresh tapes or CD/DVDs
in to the changer.

>I have not heard of anyone successfully using cd's for backup purposes,
>but I do know of people using removable hard drives as an alternate to
>tape backups.

No one has successfully used CD's or DVD's? What sort of stories have you
heard? Perhaps they were using el cheapo CD's? If one is serious about
having good backups, you don't go for the cheapest backup media you can get.


Cheers!

Kevin. (http://www.interlog.com/~kcozens/)

Owner of Elecraft K2 #2172 |"What are we going to do today, Borg?"
E-mail:kcozens at interlog dot com|"Same thing we always do, Pinkutus:
Packet:ve3syb-XXPEJ3/***@public.gmane.org#con.on.ca.na| Try to assimilate the world!"
#include <disclaimer/favourite> | -Pinkutus & the Borg

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Wil McGilvery
2003-12-31 21:59:16 UTC
Permalink
I do not like using CD's because the storage is too small. Even DVD's are not quite big enough to completely back up a system. They are also cheap and you can still experience unreliability. Even the "good" cd's are still too unreliable for me.

When I first got started I used dat tapes and sold some to customers. That was a big mistake. I was replacing the tape drive every year and once I had a new tape drive it was a crap shoot as to whether I could perform a restore or not from an old tape.

If you are backing up your own data on your own system at your house then do what ever works because obviously price is a consideration, but if you are responsible for mission critical data, don't go cheap. It is just not worth it. Spend the money and do it properly.

Like I said before, you get what you pay for.

Regards,

Wil McGilvery
Manager
Lynch Digital Media Inc



416-744-7949
416-716-3964 (cell)
1-866-314-4678
416-744-0406  FAX
www.LynchDigital.com


-----Original Message-----
From: Kevin Cozens [mailto:kcozens-qazKcTl6WRFWk0Htik3J/***@public.gmane.org]
Sent: Wednesday, December 31, 2003 4:36 PM
To: tlug-***@public.gmane.org
Subject: RE: [TLUG]: Data recovery emergency on a downed server... Help please!!

At 12:55 PM 12/31/2003 -0500, Wil McGilvery wrote:
>Backups in my opinion will never be a set it and forget it procedure.
>Periodic restores are necessary for peace of and the assurance that you
>can restore when necessary.

Tape and CD/DVD changers help but if the data is critical, someone still
has to physically move the backups off-site and put fresh tapes or CD/DVDs
in to the changer.

>I have not heard of anyone successfully using cd's for backup purposes,
>but I do know of people using removable hard drives as an alternate to
>tape backups.

No one has successfully used CD's or DVD's? What sort of stories have you
heard? Perhaps they were using el cheapo CD's? If one is serious about
having good backups, you don't go for the cheapest backup media you can get.


Cheers!

Kevin. (http://www.interlog.com/~kcozens/)

Owner of Elecraft K2 #2172 |"What are we going to do today, Borg?"
E-mail:kcozens at interlog dot com|"Same thing we always do, Pinkutus:
Packet:ve3syb-XXPEJ3/***@public.gmane.org#con.on.ca.na| Try to assimilate the world!"
#include <disclaimer/favourite> | -Pinkutus & the Borg

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
James Knott
2003-12-31 22:12:08 UTC
Permalink
I have 2 identical 60G hard drives. I simply clone the entire drive
onto the other, which is mounted in a removable tray.

Also, the best backup system is redunancy. Make lots of copies, with
off site storage, to minimize the potential loss. Live mirrors also
help protect against physical damage.



Wil McGilvery wrote:
> I do not like using CD's because the storage is too small. Even DVD's are not quite big enough to completely back up a system. They are also cheap and you can still experience unreliability. Even the "good" cd's are still too unreliable for me.
>
> When I first got started I used dat tapes and sold some to customers. That was a big mistake. I was replacing the tape drive every year and once I had a new tape drive it was a crap shoot as to whether I could perform a restore or not from an old tape.
>
> If you are backing up your own data on your own system at your house then do what ever works because obviously price is a consideration, but if you are responsible for mission critical data, don't go cheap. It is just not worth it. Spend the money and do it properly.
>
> Like I said before, you get what you pay for.
>
> Regards,
>
> Wil McGilvery
> Manager
> Lynch Digital Media Inc
>
>
>
> 416-744-7949
> 416-716-3964 (cell)
> 1-866-314-4678
> 416-744-0406 FAX
> www.LynchDigital.com
>
>
> -----Original Message-----
> From: Kevin Cozens [mailto:kcozens-qazKcTl6WRFWk0Htik3J/***@public.gmane.org]
> Sent: Wednesday, December 31, 2003 4:36 PM
> To: tlug-***@public.gmane.org
> Subject: RE: [TLUG]: Data recovery emergency on a downed server... Help please!!
>
> At 12:55 PM 12/31/2003 -0500, Wil McGilvery wrote:
>
>>Backups in my opinion will never be a set it and forget it procedure.
>>Periodic restores are necessary for peace of and the assurance that you
>>can restore when necessary.
>
>
> Tape and CD/DVD changers help but if the data is critical, someone still
> has to physically move the backups off-site and put fresh tapes or CD/DVDs
> in to the changer.
>
>
>>I have not heard of anyone successfully using cd's for backup purposes,
>>but I do know of people using removable hard drives as an alternate to
>>tape backups.
>
>
> No one has successfully used CD's or DVD's? What sort of stories have you
> heard? Perhaps they were using el cheapo CD's? If one is serious about
> having good backups, you don't go for the cheapest backup media you can get.


--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
William Park
2003-12-31 22:58:52 UTC
Permalink
On Wed, Dec 31, 2003 at 05:12:08PM -0500, James Knott wrote:
> I have 2 identical 60G hard drives. I simply clone the entire drive
> onto the other, which is mounted in a removable tray.
>
> Also, the best backup system is redunancy. Make lots of copies, with
> off site storage, to minimize the potential loss. Live mirrors also
> help protect against physical damage.

1. My advice to those experiencing flaky harddisk... Spend the money on
- good power supply and case
- good harddisk
- good motherboard
Of the above, power supply is most often overlooked. It's the most
important thing, in my opinion; and, I don't mean wattage rating on
those.

2. Since I'm running Slackware, I don't do complete system backup. I
only backup users data (ie. things that were added or modified since
install). When doing total restore, I just run a script to do the
complete installation of Linux from distribution CD; then, restore the
users data. For other distro, it may be different, because this depends
on whether you can do unattended complete install.

--
William Park, Open Geometry Consulting, <opengeometry-FFYn/***@public.gmane.org>
Linux solution for data management and processing.
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Jason Shein
2003-12-31 23:13:46 UTC
Permalink
It will be interesting when the next-gen dvd format becomes more
affordable. The blu-ray format will store 27gb on a songle dvd.

http://www.cdrinfo.com/Sections/News/Details.asp?RelatedID=4429
http://www.blu-ray.com/
http://news.com.com/2100-1041_3-990826.html

Any prices I have found are in the 3500 USD range.

Hmm... come to think of it, were'nt DVD writers about 600 buck last
christmas? About 100 bucks now. Let's hope that the blu-ray devices drop
just as fast.


William Park wrote:
> On Wed, Dec 31, 2003 at 05:12:08PM -0500, James Knott wrote:
>
>>I have 2 identical 60G hard drives. I simply clone the entire drive
>>onto the other, which is mounted in a removable tray.
>>
>>Also, the best backup system is redunancy. Make lots of copies, with
>>off site storage, to minimize the potential loss. Live mirrors also
>>help protect against physical damage.
>
>
> 1. My advice to those experiencing flaky harddisk... Spend the money on
> - good power supply and case
> - good harddisk
> - good motherboard
> Of the above, power supply is most often overlooked. It's the most
> important thing, in my opinion; and, I don't mean wattage rating on
> those.
>
> 2. Since I'm running Slackware, I don't do complete system backup. I
> only backup users data (ie. things that were added or modified since
> install). When doing total restore, I just run a script to do the
> complete installation of Linux from distribution CD; then, restore the
> users data. For other distro, it may be different, because this depends
> on whether you can do unattended complete install.
>

--
" Eventually people tire of repairing broken Windows,
And decide to replace them with something stronger"
(o_
//\ Linux - The Choice Of A GNU Generation
V_/_ Jason Shein
Linux Registered User #281100
jason-gaRZxGPHtpBxZtjKW1aY+1aTQe2KTcn/@public.gmane.org
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Steve A.
2003-12-31 23:15:22 UTC
Permalink
On Wed, Dec 31, 2003 at 04:59:16PM -0500 or thereabouts, Wil McGilvery wrote:
> I do not like using CD's because the storage is too small. Even DVD's are not quite big enough to completely back up a system. They are also cheap and you can still experience unreliability. Even the "good" cd's are still too unreliable for me.
>
> When I first got started I used dat tapes and sold some to customers. That was a big mistake. I was replacing the tape drive every year and once I had a new tape drive it was a crap shoot as to whether I could perform a restore or not from an old tape.
>
> If you are backing up your own data on your own system at your house then do what ever works because obviously price is a consideration, but if you are responsible for mission critical data, don't go cheap. It is just not worth it. Spend the money and do it properly.
>
> Like I said before, you get what you pay for.

I'm surprised to hear you say that DVD's aren't quite big enough... I'm
in the digital asset business, and we gave tape up a long time ago.
We use DVD with a jukebox, and the whole thing is automated in terms of
doing the actual backup, we have no issue carrying over files that fill
up one disk to another, the software is intelligent enough to know where
all the parts are. Our backups each day aren't tiny...
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Robert Brockway
2004-01-01 02:20:59 UTC
Permalink
On Wed, 31 Dec 2003, Wil McGilvery wrote:

> I do not like using CD's because the storage is too small. Even DVD's
> are not quite big enough to completely back up a system. They are also
> cheap and you can still experience unreliability. Even the "good" cd's
> are still too unreliable for me.
>
> When I first got started I used dat tapes and sold some to customers.
> That was a big mistake. I was replacing the tape drive every year and
> once I had a new tape drive it was a crap shoot as to whether I could
> perform a restore or not from an old tape.

One of the biggest problems with DDS (and other) tapes is the inability to
restore except on the tape drive used to do that backup. This is caused
by head misalignment on the original drive, and makes the idea of disaster
recovery a joke when you consider that in a real disaster the original
drive probably went up in smoke.

Anyone using tapes (particularly DDS) to backup important data[1] better
be testing restores on a _different_ tape drive.

[1] If you are troubled enough to back it up chances are it is important.

Rob

--
Robert Brockway
Senior Technical Consultant, OpenTrend Solutions Ltd.
Phone: 416-669-3073, Email: rbrockway-wgAaPJgzrDxH4x6Dk/***@public.gmane.org, http://www.opentrend.net
OpenTrend Solutions: Reliable, secure solutions to real world problems.
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Teddy Mills
2004-01-01 03:47:07 UTC
Permalink
I dont like backups onto CDs or any tape formats. Be it helical, travan or
whatever.
Backing up just data is fine and reliable. But systems are a NOS and data.
NOS being the operating system and all the installed programs running.

1. Getting a true backup of a system onto tape is often difficult due to
open files of the OS and various(important) data and programs.
2. If the system crashes, restoring a system entirely from tape is equally
as difficult. On servers with real data. I will never make that mistake
agan. Anyone whose ever had to rebuild a server just from tapes might agree.

The system I chose for now, is running a mirrored drive system, and backing
up just the data onto tape.
A third drive I keep as a drive image backup.



----- Original Message -----
From: "Robert Brockway" <rbrockway-wgAaPJgzrDxH4x6Dk/***@public.gmane.org>
To: <tlug-***@public.gmane.org>
Sent: Wednesday, December 31, 2003 9:20 PM
Subject: RE: [TLUG]: Data recovery emergency on a downed server... Help
please!!


> On Wed, 31 Dec 2003, Wil McGilvery wrote:
>
> > I do not like using CD's because the storage is too small. Even DVD's
> > are not quite big enough to completely back up a system. They are also
> > cheap and you can still experience unreliability. Even the "good" cd's
> > are still too unreliable for me.
> >
> > When I first got started I used dat tapes and sold some to customers.
> > That was a big mistake. I was replacing the tape drive every year and
> > once I had a new tape drive it was a crap shoot as to whether I could
> > perform a restore or not from an old tape.
>
> One of the biggest problems with DDS (and other) tapes is the inability to
> restore except on the tape drive used to do that backup. This is caused
> by head misalignment on the original drive, and makes the idea of disaster
> recovery a joke when you consider that in a real disaster the original
> drive probably went up in smoke.
>
> Anyone using tapes (particularly DDS) to backup important data[1] better
> be testing restores on a _different_ tape drive.
>
> [1] If you are troubled enough to back it up chances are it is important.
>
> Rob
>
> --
> Robert Brockway
> Senior Technical Consultant, OpenTrend Solutions Ltd.
> Phone: 416-669-3073, Email: rbrockway-wgAaPJgzrDxH4x6Dk/***@public.gmane.org,
http://www.opentrend.net
> OpenTrend Solutions: Reliable, secure solutions to real world problems.
> --
> The Toronto Linux Users Group. Meetings: http://tlug.ss.org
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
>
>

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Henry Spencer
2004-01-01 04:43:26 UTC
Permalink
On Wed, 31 Dec 2003, Teddy Mills wrote:
> 2. If the system crashes, restoring a system entirely from tape is equally
> as difficult. On servers with real data. I will never make that mistake
> agan. Anyone whose ever had to rebuild a server just from tapes might agree.

Or might not. I had to do it a number of times, in the bad old days when
I was a sysadmin, and although there were sometimes problems, the rebuild-
from-tape part was never among them. But then, we were seriously careful
about how we did backups.

Henry Spencer
henry-***@public.gmane.org

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Robert Brockway
2004-01-01 04:45:24 UTC
Permalink
On Wed, 31 Dec 2003, Teddy Mills wrote:

> I dont like backups onto CDs or any tape formats. Be it helical, travan or

I'm ok with CDs as long as they are checked (like anything I guess).
Their big problem (IMHO) is capacity.

> whatever.

Travans - I've seen some interesting experiences there. I now avoid them.

> Backing up just data is fine and reliable. But systems are a NOS and data.
> NOS being the operating system and all the installed programs running.

We work on the assumption of full system recovery in as short a time as
possible (mainly looking at a business perspective here). For this it is
necessary to backup everything. It takes far too long to reinstall all
the boxes and configure their apps, etc. I want the speed of the storage
device to be the limiting factor.

> 1. Getting a true backup of a system onto tape is often difficult due to
> open files of the OS and various(important) data and programs.

I've had surprisingly few problems here. True, having a file open can
be a problem, but with nightly backups it tends to be that the same files
are not open from night to night.

If someone was really worried about this, I'd be worth taking a system to
single user mode periodically and doing a full (rather than incremental)
backup in this state. This is not something I've found a particular need
for.

Nost systems are backuped when they are relatively idle. If a system was
active 24/7 it would be worth looking at snapshotting data and taking the
backups from there.

> 2. If the system crashes, restoring a system entirely from tape is equally
> as difficult. On servers with real data. I will never make that mistake
> agan. Anyone whose ever had to rebuild a server just from tapes might agree.

This is where DR (Disaster Recovery) testing comes in. If there is a
problem with the DR plan (and there often is) run a live DR test and find
out before you have to do it for real.

Live DR testing is timeconsuming and thus company accountants and managers
may be tempted to cut corners (seeing the time translate into a monetary
cost). Insist on a proper DR test and detail why it is needed.

With the reliance on information resources in todays companies, few things
as a important as a properly tested DR plan but impressing this on the
management in many companies has been a slow process.

It has been said that management in many companies are getting more
"Information Wise" but I won't consider them information wise until they
are "DR Wise" :)

Have a good NYE and don't stop partying just because some of us are sick
:)

Cheers,
Rob

--
Robert Brockway
Senior Technical Consultant, OpenTrend Solutions Ltd.
Phone: 416-669-3073, Email: rbrockway-wgAaPJgzrDxH4x6Dk/***@public.gmane.org, http://www.opentrend.net
OpenTrend Solutions: Reliable, secure solutions to real world problems.
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Peter L. Peres
2004-01-02 17:27:27 UTC
Permalink
On Wed, 31 Dec 2003, Robert Brockway wrote:

> One of the biggest problems with DDS (and other) tapes is the inability to
> restore except on the tape drive used to do that backup. This is caused
> by head misalignment on the original drive, and makes the idea of disaster
> recovery a joke when you consider that in a real disaster the original
> drive probably went up in smoke.
>
> Anyone using tapes (particularly DDS) to backup important data[1] better
> be testing restores on a _different_ tape drive.

I do not agree fully. Incompatible tape mechanisms and backup tape horror
stories are due to the strange ideas customers have about tape mechanism
MTBFs. A tape mechanism belongs in a service center for checkup once at
most every 1000 hours of use at the latest (better 500 hours, or once a
year), assuming it runs in a low dust conditioned environment with no
smoking and no copiers/laser printers allowed near it (some of the toner
ends up as fine dust inside the surrounding machines), and with tapes
properly stored and changed (no tape will last anywhere near 500 uses).
The tests performed include compatibility testing, cleaning, and recording
quality evaluation (ber) (which may mean you may have to buy a new tape
drive if it works out low). By comparison a hard disk can rake up 20,000
hours (more than two years continuous 24/7/365) and sometimes much more
without problems.

Peter
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
James Knott
2004-01-02 21:41:30 UTC
Permalink
Peter L. Peres wrote:
> On Wed, 31 Dec 2003, Robert Brockway wrote:
> I do not agree fully. Incompatible tape mechanisms and backup tape horror
> stories are due to the strange ideas customers have about tape mechanism
> MTBFs. A tape mechanism belongs in a service center for checkup once at
> most every 1000 hours of use at the latest (better 500 hours, or once a
> year), assuming it runs in a low dust conditioned environment with no
> smoking and no copiers/laser printers allowed near it (some of the toner
> ends up as fine dust inside the surrounding machines)

And no programmers, who insist on smoking near the computers. Years
ago, I serviced mini computers and there were a couple of programmers
who'd be smoking while changing tapes or disk packs, despite being told
not to. (this was before the no smoking laws) I also had to clean the
cigarette ashes out of a lot of terminal keyboards.




--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Robert Brockway
2004-01-02 21:48:03 UTC
Permalink
On Fri, 2 Jan 2004, James Knott wrote:

> And no programmers, who insist on smoking near the computers. Years
> ago, I serviced mini computers and there were a couple of programmers
> who'd be smoking while changing tapes or disk packs, despite being told

Wow. Few things are as bad for magnetic media as cigarette smoke. The
particles in the smoke bond to the media _really_ well.

Rob

--
Robert Brockway
Senior Technical Consultant, OpenTrend Solutions Ltd.
Phone: 416-669-3073, Email: rbrockway-wgAaPJgzrDxH4x6Dk/***@public.gmane.org, http://www.opentrend.net
OpenTrend Solutions: Reliable, secure solutions to real world problems.
--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
James Knott
2004-01-02 22:12:45 UTC
Permalink
Robert Brockway wrote:
> On Fri, 2 Jan 2004, James Knott wrote:
>
>
>>And no programmers, who insist on smoking near the computers. Years
>>ago, I serviced mini computers and there were a couple of programmers
>>who'd be smoking while changing tapes or disk packs, despite being told
>
>
> Wow. Few things are as bad for magnetic media as cigarette smoke. The
> particles in the smoke bond to the media _really_ well.

Yes, I know. I'm the guy who had to clean the heads on those drives.
The cleaning pads would have a yellow tint to them. We also had
occasional head crashes in the disk pack drives. I wonder how many of
those were caused by the smoke?



--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Madison Kelly
2004-01-03 21:11:29 UTC
Permalink
Hi all,

Another update of the data recovery job... Action Front Data
Recovery, the DR house we sent the drive to, seems to think they can
recover the drive. Apparently the directory structure is somewhat
corrupted but the data itself seems to be instact. Anywho, I will post a
final update when they are finished (~Wednesday).

THANK YOU ALL again for offering so much help, advice and concern!

Madison

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Byron Sonne
2004-01-05 06:31:17 UTC
Permalink
> THANK YOU ALL again for offering so much help, advice and concern!

Having gone to data recovery houses in the past 3 or 4 times, and having
a %0 percent success rate, I've always got my ears out for this kinda
thing.

Wondering why it never worked for me/my employer but there are numerous
success stories. Maybe we ran over some dude from Memofix's cat ;)


--

For Good, return Good. For Evil, return Justice.

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Kevin Cozens
2004-01-06 16:15:09 UTC
Permalink
At 01:31 AM 01/05/2004 -0500, you wrote:
>Having gone to data recovery houses in the past 3 or 4 times, and having a
>%0 percent success rate, I've always got my ears out for this kinda thing.
>
>Wondering why it never worked for me/my employer but there are numerous
>success stories. Maybe we ran over some dude from Memofix's cat ;)

Isn't there a data recovery company in Mississauga(?) that is supposed to
be good. I'm thinking of one that has been written about in one of the
local papers on a couple of occasions.


Cheers!

Kevin. (http://www.interlog.com/~kcozens/)

Owner of Elecraft K2 #2172 |"What are we going to do today, Borg?"
E-mail:kcozens at interlog dot com|"Same thing we always do, Pinkutus:
Packet:ve3syb-XXPEJ3/***@public.gmane.org#con.on.ca.na| Try to assimilate the world!"
#include <disclaimer/favourite> | -Pinkutus & the Borg

--
The Toronto Linux Users Group. Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml
Continue reading on narkive:
Search results for ': Data recovery emergency on a downed server... Help please!!' (newsgroups and mailing lists)
52
replies
Beowulf Questions
started 2003-01-03 08:46:41 UTC
beowulf@beowulf.org
90
replies
Recovery
started 2007-07-23 20:40:19 UTC
alt.sysadmin.recovery
Loading...