Gene Heskett wrote:
On Sunday 21 March 2010, Steve Wray wrote:
Hi there
I'm preparing some documentation on our backup/restore process.
Some time ago I linked to amanda documentation on how to extract from tapes
without using any amanda tools or indexes.
This link was:
http://www.amanda.org/docs/using.html#restoring_without_amanda
and is now broken.
This isn't quite what I'm after as it uses amrestore:
http://wiki.zmanda.com/index.php/How_To:Recover_Lost_Amanda_Configuration
Can anyone point me at the current correct link please?
Thanks!
cd to /tmp, then dd the first block of the tape to std out. You will see a
command line that will recover that file to /tmp. Repeat till out of tape.
I keep the amanda printouts, so I know which tape has the last level0 backup
of the amanda stuff, and pull that out first. It will probably be 3 or 4
tapes to get it all, but once you've pulled the config, then get the indexes
& put both back where they go. Next get the directory you built amanda in,
probably /home and copy that back. cd to /home/amanda (if amanda is the user
that built and installed it, and do another install as root.
That ought to get you enough to run amrecover at which point load it up with
the full monty and stand by to give it the tapes it asks for.
Or, you can be a bit more preemptive.
I have a script that runs after the Amanda backup is done and copies all
the Amanda home directory to another server. This includes all the
configuration and indexes. You can do this with just about any mechanism
you're comfortable with -- cpio, tar, scp, rsync, ... whatever. Each of
my Amanda servers sends everything over to another Amanda server. Sort
of a round robin. I use a script with a combination of find and cpio to
achieve a sequence of full and incremental backups that are kept for a week.
Without something like the above, you end up being a backup behind on
things like the indexes. Backups of the Amanda home directory that are
on tape can't include the latest changes to the Amanda indexes made as
backups that go on that tape are being collected.
I've also tried just reading the Amanda tapes directly. It works even if
it is a little tedious. In my case the sequence of commands involves
ufsrestore, but that too is documented on the tape itself in the first
record.
--
---------------
Chris Hoogendyk
-
O__ ---- Systems Administrator
c/ /'_ --- Biology & Geology Departments
(*) \(*) -- 140 Morrill Science Center
~~~~~~~~~~ - University of Massachusetts, Amherst
<hoogen...@bio.umass.edu>
---------------
Erdös 4
