Amanda-Users

RE: How does the amanda server identify itself to clients?

2004-02-06 13:27:00
Subject: RE: How does the amanda server identify itself to clients?
From: Michael Kahle <michael.kahle AT scc-wi DOT com>
To: "'amanda-users AT amanda DOT org'" <amanda-users AT amanda DOT org>
Date: Fri, 6 Feb 2004 12:13:13 -0600
Sorry, my mail reader sucks (outlook)...
-----Original Message-----
From: Stefan G. Weichinger [mailto:monitor AT oops.co DOT at] 
Sent: Friday, February 06, 2004 12:09 PM
To: amanda-users AT amanda DOT org
Subject: Re: How does the amanda server identify itself to clients?

<snip>
The server has been compiled with -default-tape-server=localhost ?

If you still have the sources lying around I would recommend a simple

make uninstall
./configure --with-tape-server=<fqdn> --other-options --... make install

It does not break/delete/change your config/index-dirs or such.

But you get predictable behavior. It is a well known fact on this list to
not use "localhost". As it seems it doesn't come through the docs well
enough.
</snip>

Ok, this is (I think) the info I was looking for.  I just didn't do a good
job wording it properly the first time (and second?).  :)

The Debian package maintainer compiled this with
"--default-tape-server=localhost"  I did not roll my own.  I try not to
whenever possible on Debian.  From the looks of it, I will have to.

Why doesn't amanda allow for runtime configuration of this stuff?  I.e., in
the /etc/amanda folder you could put a "master.conf" file that would contain
info like, default user, default tape server, etc.  This kind of stuff
should be separated from the binaries!  Weird.

According to Oreilly's "Unix Backup & Recovery" chapter on Amanda p.151 it
reads:
"Some operating system distributions provide precompiled versions of AMANDA,
but because AMANDA hard codes some values into the programs, they may not
match the configuration."

I just answered my own question above, as the paragraph continues...
"Work is being done to move these values to runtime configuration files, but
for now AMANDA should be built from source."

I guess I should take the authors advice.  When I read things like that I
think, "Well that may be true for LESSER distributions."  Looks like my
"blind faith" in the infallibility of Debian has just been shaken to the
core.  I will go compile now... and weep.

Thanks.

Michael