[prev in list] [next in list] [prev in thread] [next in thread] 

List:       amanda-hackers
Subject:    Re: Backing up a system twice, simultaneously ?
From:       Mitch Collinsworth <mitch () ccmr ! cornell ! edu>
Date:       2001-08-22 14:31:29
[Download RAW message or body]


On Wed, 22 Aug 2001, Paul Lussier wrote:

> 
> In a message dated: Tue, 21 Aug 2001 21:01:23 EDT
> Mitch Collinsworth said:
> 
> >I don't remember the exact error, but yes it does cause a problem.
> >Two amandad's running at the same time on the same client just
> >doesn't work.
> >
> >What happens when we're backing up more than one filesystem on
> >the same machine is the server sends the whole list of filesystems
> >for that client in it's request to amandad.
> 
> This is exactly what I thought would happen.  However, Marc brought 
> up a good point, since it's running via inetd, shouldn't it just 
> launch a second instance of the amandad process?  And since the 
> second instance is being called via a different IP, it would be 
> called with a completely separate set of filesystems.  Isn't the 
> whole point of inetd to allow multiple, concurrent instances of the 
> same daemon to be connected to?
 
Well yes, Marc is correct that inetd will happily launch a 2nd amandad.
The problem occurs somewhere after this point.  Sorry it's been too
long since I ran into this to remember the details, but there is some
conflict that happens at this point and the 2nd amandad fails.  It's
not a bug per se, it's just that the original system design never
allowed for this possibility.  Some enterprising person could probably
change it to add this feature.

> I'm confused now.  I think it *should* work, but can also see why it 
> might not.  I guess the only thing I can do is try it out and see 
> what happens :)

Sure, feel free to try it if you like...

-Mitch

[prev in list] [next in list] [prev in thread] [next in thread]