Amanda-Users

Re: estimate timeout

2005-10-10 10:55:46
Subject: Re: estimate timeout
From: Gene Heskett <gene.heskett AT verizon DOT net>
To: amanda-users AT amanda DOT org
Date: Mon, 10 Oct 2005 10:45:20 -0400
On Monday 10 October 2005 03:20, Shai Ayal wrote:
>Hi all,
>
>I have searched the archives but none of the emails with similar
> subjects helped me.
>
>I have a FC2 amanda 2.4.4 server with 2 linux clients. The server is
> using vtapes for daily backups. It all ran very nicely for many months
> until we ran out of disk space in the server. After a few days of bad
> backups due to full disk, we installed an additional disk, moved some
> of the virtual tapes to it using symlinks, flushed the old backups
> etc... and sat back to enjoy amanda at work.
>
>However:
>
>While one client is being backed up perfectly well, the other keeps
> getting estimates timeout. On this client, everything seem ok except
> for showing 2 amandad processes during estimates, one of them defunct
> -- I attach the 2 amandad debug reports.

It is possible that the defunct amandad has open locks on files, thereby
blocking the estimate.  2 things might help, first I'd reboot the machine
the failure is on to remove them, and then I think I'd install a newer
amanda, 2.4.4 is getting a bit long in the tooth these days.  I can't
recall the exact version I was running when that happened on my firewall
box, mainly because I wasn't doing virtual tapes yet and was having so
many other tape related issues back then that a stuck amandad just wasn't
an event to record at length in my wetram.

If you still jave the same scripts you used to build the 2.4.4 on each
box, then 2.4.5-20051006 should install and run exactly the same.

However, I just checked the /home/amanda directory on my single linux
client, and its equally elderly, at 2.4.4-20030529, and its working fine
other than a 10 second delay in checking clients when amcheck is run,
about 80% of the time.

But this is as good a time to bring it uptodate as any, so its building on
that box now.  Using the same script I built the older version with.  
Oops,
forgot to run ldconfig after the install, done now.

Humm, I note that, and this has been random in the past, true about 80%
of the time, but there is no longer a 10 second delay in checking the
clients now, more like .35 seconds.  At least for the several iterations
of it I've done.  Maybe thats fixed now?

>On the server I have set an etimeout of 300 which should be enough, but
>even bumping this to 7200 did not help.
>
>I have no firewall on client and server
 
I do, but it not between the client and server, its betwen client and
the rest of the planet.  That box is the gateway.

>tar version is tar (GNU tar) 1.13.25 o the client

Thats a good one, although I'm running 1.15-1 on the server.  But the
client box is rh7.3, and the glib version won't let me build, or install
1.15.1.

>This is really frustrating since this setup used to work !
>
>Thanks in advance
>Shai

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
99.35% setiathome rank, not too shabby for a WV hillbilly
Yahoo.com and AOL/TW attorneys please note, additions to the above
message by Gene Heskett are:
Copyright 2005 by Maurice Eugene Heskett, all rights reserved.


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