Amanda-Users

Re: Problem Calculating Estimates on RHEL 4

2005-04-14 14:32:06
Subject: Re: Problem Calculating Estimates on RHEL 4
From: Matt Hyclak <hyclak AT math.ohiou DOT edu>
To: Amanda Users <amanda-users AT amanda DOT org>
Date: Thu, 14 Apr 2005 14:22:57 -0400
On Thu, Apr 14, 2005 at 01:14:01PM -0500, Frank Smith enlightened us:
> --On Thursday, April 14, 2005 12:44:24 -0500 Matt Hyclak <hyclak AT 
> math.ohiou DOT edu> wrote:
> 
> > I ran into an interesting problem today, and I thought I'd post here to
> > get
> > some feedback/make the problem known.
> > 
> > First the specs:
> > Server:
> > RHEL 3 x86
> > amanda 2.4.4p3 (rebuilt with fqdn's and not localhost)
> > tar 1.13.25-13
> > 
> > RHEL 4 x86 (CentOS to be exact, but same difference)
> > amanda 2.4.4p3 (rebuilt as above)
> > tar 1.14-4
> > 
> > The machine in question was formerly a FC2 box, which I wiped and
> > installed
> > CentOS 4 on yesterday. I got amanda set up today and did a backup. My
> > /home
> > directory went just fine, but the / partition errored out saying that
> > the
> > partition was too big for tape. Funny, since it's a 20GB partition and
> > an
> > LTO-1 (100GB native) tape drive.
> > 
> > So I get into the debug logs on the client and run the tar command used
> > to
> > create the estimate and it informed me that the size is 1.2TB. I am
> > wondering about some of the output below. The socket ignored is normal,
> > however the "Cannot seek to 0" errors make me scratch my head. 
> 
> Two possibilities I can think of:
> Have you fsck'd the filesystem?  I've seen odd sizes from this before due
> to corrupted directory entries and/or overlapping partitions (which are
> really nasty, since an fsck can appear to fix it until something on the other
> partion writes to the overlap, so double-check your partition table).
> Second, do you have large sparse files that don't take up much physical
> space on disk but some versions of tar expand out to the full size?
>

I haven't fsck'd it personally, but it was *just* installed with a clean
format, so I don't think the problem is there. I think that the problem is
the sparse files...I haven't created any myself, but the output below makes
me think that this is the problem reported in
http://lists.gnu.org/archive/html/bug-tar/2005-02/msg00003.html

I've filed a bug report with RedHat.
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=154882 for those that
are interested.

Matt


> 
> >Before
> > running the command, I made sure
> > /var/lib/amanda/gnutar-lists/euclid.math.ohiou.edu__0.new didn't exist
> > and
> > also of note is that the exclude file is empty.
> > 
> > [root@euclid /]# /bin/tar --create --file /dev/null --directory /
> > --one-file-system --listed-incremental
> > /var/lib/amanda/gnutar-lists/euclid.math.ohiou.edu__0.new --sparse
> > --ignore-failed-read --totals --exclude-from
> > /var/log/amanda/sendsize._.20050414123909.exclude .
> > /bin/tar: ./tmp/.gdm_socket: socket ignored
> > /bin/tar: ./tmp/gaim_hyclak.0: socket ignored
> > /bin/tar: ./tmp/xmms_hyclak.0: socket ignored
> > /bin/tar: ./tmp/.X11-unix/X0: socket ignored
> > /bin/tar: ./tmp/.font-unix/fs7100: socket ignored
> > /bin/tar: ./tmp/orbit-hyclak/linc-1b5d-0-c1e4a46c44ed: socket ignored
> > /bin/tar: ./tmp/orbit-hyclak/linc-5b21-0-1c928a1496c03: socket ignored
> > /bin/tar: ./tmp/orbit-hyclak/linc-5b9f-0-66f105dcd65a4: socket ignored
> > /bin/tar: ./tmp/orbit-hyclak/linc-5ba1-0-26e0b1c6e6539: socket ignored
> > /bin/tar: ./tmp/ssh-Lisbo22174/agent.22174: socket ignored
> > /bin/tar: ./tmp/ssh-flQTJg4250/agent.4250: socket ignored
> > /bin/tar: ./tmp/ssh-hGbnhR6745/agent.6745: socket ignored
> > /bin/tar: ./tmp/ssh-jhRtQt3934/agent.3934: socket ignored
> > /bin/tar: ./tmp/ssh-rdYCJE6386/agent.6386: socket ignored
> > /bin/tar: ./var/db/nscd/group: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/db/nscd/hosts: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/db/nscd/passwd: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/lib/rpm/Basenames: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/lib/rpm/Filemd5s: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/lib/rpm/Providename: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/lib/rpm/Requirename: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/lib/rpm/Sha1header: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/lib/rpm/__db.001: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/lib/rpm/__db.003: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/log/lastlog: Warning: Cannot seek to 0: Bad file
> > descriptor
> > /bin/tar: ./var/run/acpid.socket: socket ignored
> > /bin/tar: ./var/run/dbus/system_bus_socket: socket ignored
> > /bin/tar: ./var/run/nscd/socket: socket ignored
> > Total bytes written: 1257689999360 (1.2TiB, 9.7GiB/s)
> > 
> > It seems like I need to upgrade tar anyway if I'd like to be able to
> > actually restore files on this system, however from the recent post it
> > would
> > seem creating the archive shouldn't be a problem...
> > 
> > Thoughts?
> > 
> > Thanks,
> > Matt

> 
> 
> 
> -- 
> Frank Smith                                      fsmith AT hoovers DOT com
> Sr. Systems Administrator                       Voice: 512-374-4673
> Hoover's Online                                   Fax: 512-374-4501
> 



-- 
Matt Hyclak
Department of Mathematics 
Department of Social Work
Ohio University
(740) 593-1263

Attachment: pgpW46wwt7Xne.pgp
Description: PGP signature

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