Amanda-Users

Re: all estimate failed - problem

2005-10-14 09:24:33
Subject: Re: all estimate failed - problem
From: Joshua Baker-LePain <jlb17 AT duke DOT edu>
To: Sylvia Gelman <sgelman AT rtr.tu-darmstadt DOT de>
Date: Fri, 14 Oct 2005 09:19:00 -0400 (EDT)
On Fri, 14 Oct 2005 at 2:44pm, Sylvia Gelman wrote

mailserver       /var/mail lev 0 FAILED [disk /var/mail, all estimate failed]
*snip*
Is there a way how I could test the estimate time "dry" without making really a backup?

Yep. sendsize*debug has the exact command it runs, which you could then run by hand.

sendsize[5579]: time 0.332: calculating for amname '/var/mail', dirname '/var/mail', spindle -1
sendsize[5579]: time 0.335: getting size via gnutar for /var/mail level 0
sendsize[5577]: time 0.344: waiting for any estimate child: 1 running
sendsize[5579]: time 0.375: spawning /usr/local/amanda/libexec/runtar in pipeline sendsize[5579]: argument list: /opt/gnu/bin/tar --create --file /dev/null --directory /var/mail --one-file-system --listed-incremental /usr/local/amanda/va r/amanda/gnutar-lists/glia_var_mail_0.new --sparse --ignore-failed-read --totals --exclude-from /tmp/amanda/sendsize._var_mail.20051013210005.exclude .

There's the command (I'm assuming the odd line break is a mailer issue.

*********************************************************************************
sendsize[5579]: time 349.922: Total bytes written: -1386622976

And there's your problem. tar is returning a *negative* size for /var/mail. You have to figure out why that's happening. Personally, I have no idea. :)

--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University

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