Amanda-Users

Re: Recovering tapeless amanda server

2004-07-14 05:03:33
Subject: Re: Recovering tapeless amanda server
From: Paul Bijnens <paul.bijnens AT xplanation DOT com>
To: amanda-users AT amanda DOT org
Date: Wed, 14 Jul 2004 10:57:02 +0200
Joshua Baker-LePain wrote:
On Tue, 13 Jul 2004 at 7:43pm, Ralph at yahoogroups wrote

We've just had a disk crash on our Amanda 2.4.3 server on RedHat 7.3
Linux.  I had it configured to do tapeless backups onto a different
fileserver (via chg-multi (I think) and the file tapetype.  However, I was
shortsighted enough to backup the Amanda config files with Amanda herself.
So, now I am left with a load of "tape" files but don't know how to get
at their contents.

Can someone, please, point me in the right direction?


docs/RESTORE will still cover this I believe. Just treat the "tape" files as if they were tapes. 'dd if=tapefile of=header.out bs=32k count=1' will read the header of the first image in 'tapefile'...

Come to think of it, though, I don't know a priori how to get to the second image in one of those files (you could do it with counting blocks). Your easiest path is probably just to get amanda installed on the rebuilt server and use amrestore on those tape files. amrestore doesn't need indices.

Second image in one file?  The format the the virtual tapes is actually
quite easy.  Each image on tape is one file on disk (+ some other files
to store typical tape parameters, like blocksize etc.). Even the name of
that file contains hostname and DLE.

Or did I miss something?


I would do something like:

1. boot from CD, partition the disks, and format them appropriate
(You could use the opportunity to install a newer version too.)
A knoppix CD is very handy in this case; you have lots of tools
available, including the network and utilities like netcat.
Then mount the new partition(s) somewhere, e.g.

        mount /dev/hda1 /mnt
        mkdir /mnt/home
        mount /dev/hda2 /mnt/home

2. Assuming you have netcat available this is very quick:
First on the crashed host:

        nc -l -p 1234 | gtar -zpxvf -

Second on the fileserver that has the backup images:

        dd if=00005.crashhost._.0 bs=32k skip=1 | nc -w 1 crashhost 1234

(fill in the real number.host._Disk_List_Entry.level,
maybe omit -z flag if backup was not compressed)
Repeat for each backup level.
Do this for the other partitions as well.

If you don't have nc, then you'll need to set up some network
pipes using rsh or ssh (fighting against all kinds of authentication
problems, which I usually hate in these circumstances.)

If you do have "amrestore" installed on that fileserver, you
can replace the dd with:

        ammt -f file:/your/virtual/tape rewind
        amrestore -p file:/your/virtual/tape  | nc -w 1 crashhost 1234


3. If you use lilo as boot loader, execute

      chroot /mnt lilo

For grub, I've not yet had the opportunity to test this out :-)
If anyone has this knowledge online, let me know.


--
Paul Bijnens, Xplanation                            Tel  +32 16 397.511
Technologielaan 21 bus 2, B-3001 Leuven, BELGIUM    Fax  +32 16 397.512
http://www.xplanation.com/          email:  Paul.Bijnens AT xplanation DOT com
***********************************************************************
* I think I've got the hang of it now:  exit, ^D, ^C, ^\, ^Z, ^Q, F6, *
* quit,  ZZ, :q, :q!,  M-Z, ^X^C,  logoff, logout, close, bye,  /bye, *
* stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt,  abort,  hangup, *
* PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e,  kill -1 $$,  shutdown, *
* kill -9 1,  Alt-F4,  Ctrl-Alt-Del,  AltGr-NumLock,  Stop-A,  ...    *
* ...  "Are you sure?"  ...   YES   ...   Phew ...   I'm out          *
***********************************************************************



<Prev in Thread] Current Thread [Next in Thread>