XFS, Linux, and Amanda

2002-01-14 Thread Dan Rich

I am going slowly nuts (ok, so I'm already there :) trying to get Amanda
to backup my XFS partitions.  I've rebuilt from source (CVS), configured
it with XFS, and even used strings on the executables to prove that the
xfsdump command is making it in there.  However, I still get the following
in my nightly backups:

FAILURE AND STRANGE DUMP SUMMARY:
morpheus   hde11 lev 0 FAILED [disk hde11 offline on morpheus?]
morpheus   hde10 lev 0 FAILED [disk hde10 offline on morpheus?]
 and so on 

Any ideas of what I can look for?  I can't find anything in the list
archives or the Yahoo Amanada group that seems to help.

Thanks!

-- 
Dan Rich <[EMAIL PROTECTED]> |   http://www.employees.org/~drich/
   |  "Step up to red alert!"  "Are you sure, sir?
   |   It means changing the bulb in the sign..."
   |  - Red Dwarf (BBC)





Re: XFS, Linux, and Amanda

2002-01-14 Thread Joshua Baker-LePain

On Mon, 14 Jan 2002 at 1:17pm, Dan Rich wrote

> FAILURE AND STRANGE DUMP SUMMARY:
> morpheus   hde11 lev 0 FAILED [disk hde11 offline on morpheus?]
> morpheus   hde10 lev 0 FAILED [disk hde10 offline on morpheus?]
>  and so on 
> 
> Any ideas of what I can look for?  I can't find anything in the list
> archives or the Yahoo Amanada group that seems to help.

What is in /tmp/amanda/sendbackup*debug and /tmp/amanda/amandad*debug on 
morpheus at the time(s) of the attempted dump(s)?

What distribution?

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




Re: XFS, Linux, and Amanda

2002-01-14 Thread Dan Rich

On Mon, 14 Jan 2002, Joshua Baker-LePain wrote:

> On Mon, 14 Jan 2002 at 1:17pm, Dan Rich wrote
>
> > FAILURE AND STRANGE DUMP SUMMARY:
> > morpheus   hde11 lev 0 FAILED [disk hde11 offline on morpheus?]
> > morpheus   hde10 lev 0 FAILED [disk hde10 offline on morpheus?]
> >  and so on 
> >
> > Any ideas of what I can look for?  I can't find anything in the list
> > archives or the Yahoo Amanada group that seems to help.
>
> What is in /tmp/amanda/sendbackup*debug and /tmp/amanda/amandad*debug on
> morpheus at the time(s) of the attempted dump(s)?

A *whole* lot of data. :)  The sendbackup log only shows my one ext2
partition.  The amandad log looks good for all the partitions.  The debug
log that worries me is the sendsize log:
calculating for amname 'hde11', dirname '/misc/mp3'
sendsize: getting size via dump for hde11 level 0
sendsize: running "/sbin/dump 0Ssf 1048576 - /dev/hde11"
running /usr/lib/amanda/killpgrp
/dev/hde11: Bad magic number in super-block while opening filesystem
.
(no size line match in above dump output)
.
asking killpgrp to terminate
Shouldn't it be using xfsdump instead of dump to determine the size?

Oh, I should have mentioned, it's working fine for my non-XFS partitions.

> What distribution?

Redhat 7.0 with lots of packages replaced (mostly upgrades).  The system
is running a 2.4.9 kernel.

-- 
Dan Rich <[EMAIL PROTECTED]> |   http://www.employees.org/~drich/
   |  "Step up to red alert!"  "Are you sure, sir?
   |   It means changing the bulb in the sign..."
   |  - Red Dwarf (BBC)




Re: XFS, Linux, and Amanda

2002-01-14 Thread Joshua Baker-LePain

On Mon, 14 Jan 2002 at 2:08pm, Dan Rich wrote

> On Mon, 14 Jan 2002, Joshua Baker-LePain wrote:
> 
> > What is in /tmp/amanda/sendbackup*debug and /tmp/amanda/amandad*debug on
> > morpheus at the time(s) of the attempted dump(s)?
> 
> A *whole* lot of data. :)  The sendbackup log only shows my one ext2

There would be a senbackup log for every partition backed up, but...

> partition.  The amandad log looks good for all the partitions.  The debug
> log that worries me is the sendsize log:
>   calculating for amname 'hde11', dirname '/misc/mp3'
>   sendsize: getting size via dump for hde11 level 0
>   sendsize: running "/sbin/dump 0Ssf 1048576 - /dev/hde11"
>   running /usr/lib/amanda/killpgrp
>   /dev/hde11: Bad magic number in super-block while opening filesystem
>   .
>   (no size line match in above dump output)
>   .
>   asking killpgrp to terminate
> Shouldn't it be using xfsdump instead of dump to determine the size?

Indeed it should, and that's your problem -- amanda isn't getting 
estimates for your XFS partitions.  I assume that xfsdump was present when 
you ./configured amanda, right?  What does your fstab look like?

Can you try again with either 2.4.2p2 (plus the advfs patch) or 2.4.3b2?

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




Re: XFS, Linux, and Amanda

2002-01-14 Thread Brandon D. Valentine

On Mon, 14 Jan 2002, Dan Rich wrote:

>I am going slowly nuts (ok, so I'm already there :) trying to get Amanda
>to backup my XFS partitions.  I've rebuilt from source (CVS), configured
>it with XFS, and even used strings on the executables to prove that the
>xfsdump command is making it in there.  However, I still get the following
>in my nightly backups:
>
>FAILURE AND STRANGE DUMP SUMMARY:
>morpheus   hde11 lev 0 FAILED [disk hde11 offline on morpheus?]
>morpheus   hde10 lev 0 FAILED [disk hde10 offline on morpheus?]
> and so on 
>
>Any ideas of what I can look for?  I can't find anything in the list
>archives or the Yahoo Amanada group that seems to help.

I use Amanda to regularly backup over a terabyte of data on XFS
partitions, both Linux and SGI, and it works just great out of the box.
I do not and have never used xfsdump however.  I much prefer GNUtar for
Amanda use.  You might investigate using that instead.

Brandon D. Valentine
-- 
"Iam mens praetrepidans avet vagari."
- G. Valerius Catullus, Carmina, XLVI




Re: XFS, Linux, and Amanda

2002-01-14 Thread Dan Rich

On Mon, 14 Jan 2002, Joshua Baker-LePain wrote:

> Indeed it should, and that's your problem -- amanda isn't getting
> estimates for your XFS partitions.  I assume that xfsdump was present when
> you ./configured amanda, right?  What does your fstab look like?

I found part of the problem, it appears that the original RPM I installed
installs things in a very different location from "make install".  Amdump
was running the wrong copy of sendsize.  Now that that is fixed, I can
start figuring out the next error :)

Thanks for the help!

-- 
Dan Rich <[EMAIL PROTECTED]> |   http://www.employees.org/~drich/
   |  "Step up to red alert!"  "Are you sure, sir?
   |   It means changing the bulb in the sign..."
   |  - Red Dwarf (BBC)




Re: XFS, Linux, and Amanda

2002-01-14 Thread Joshua Baker-LePain

On Mon, 14 Jan 2002 at 2:57pm, Dan Rich wrote

> I found part of the problem, it appears that the original RPM I installed
> installs things in a very different location from "make install".  Amdump
> was running the wrong copy of sendsize.  Now that that is fixed, I can
> start figuring out the next error :)
> 
Hmmm, you did 'rpm -e' the RPM version, right?  Pre-build amanda=bad.

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




Re: XFS, Linux, and Amanda

2002-01-14 Thread Patrick Michael Kane

* Joshua Baker-LePain ([EMAIL PROTECTED]) [020114 15:37]:
> On Mon, 14 Jan 2002 at 2:57pm, Dan Rich wrote
> 
> > I found part of the problem, it appears that the original RPM I installed
> > installs things in a very different location from "make install".  Amdump
> > was running the wrong copy of sendsize.  Now that that is fixed, I can
> > start figuring out the next error :)
> > 
> Hmmm, you did 'rpm -e' the RPM version, right?  Pre-build amanda=bad.

Although that has been my general experience as well, the RedHat
7.1/7.2 RPMs work well enough in my setup, which is fairly complex.
Earlier ones barfed all over the place.

Best,
-- 
Patrick Michael Kane
<[EMAIL PROTECTED]>



Re: XFS, Linux, and Amanda

2002-01-14 Thread Brandon D. Valentine

On Mon, 14 Jan 2002, Joshua Baker-LePain wrote:

>Hmmm, you did 'rpm -e' the RPM version, right?  Pre-build amanda=bad.

Word.  Especially the moronic way in which RedHat has decided to build
it.

Brandon D. Valentine
-- 
"Iam mens praetrepidans avet vagari."
- G. Valerius Catullus, Carmina, XLVI




Re: XFS, Linux, and Amanda

2002-01-14 Thread Gene Heskett

On Monday 14 January 2002 04:17 pm, Dan Rich wrote:
>I am going slowly nuts (ok, so I'm already there :) trying to
> get Amanda to backup my XFS partitions.  I've rebuilt from
> source (CVS), configured it with XFS, and even used strings on
> the executables to prove that the xfsdump command is making it
> in there.  However, I still get the following in my nightly
> backups:
>
>FAILURE AND STRANGE DUMP SUMMARY:
>morpheus   hde11 lev 0 FAILED [disk hde11 offline on morpheus?]
>morpheus   hde10 lev 0 FAILED [disk hde10 offline on morpheus?]
> and so on 
>
>Any ideas of what I can look for?  I can't find anything in the
> list archives or the Yahoo Amanada group that seems to help.
>
>Thanks!

Humm, what I see is that you are using linux dump, specifically 
built for the ext2 filesystem, on a XFS filesystem?  Donnbesilly.

Use tar instead, its *lots* more dependable anyway.

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
98.3+% setiathome rank, not too shabby for a hillbilly



Re: XFS, Linux, and Amanda

2002-01-15 Thread Joshua Baker-LePain

On Mon, 14 Jan 2002 at 8:51pm, Gene Heskett wrote

> Humm, what I see is that you are using linux dump, specifically 
> built for the ext2 filesystem, on a XFS filesystem?  Donnbesilly.

No.  Amanda checks the fstab for filesystem type, and uses the appropriate 
dump tool for each filesystem (given that it was present when amanda was 
compiled).  So, on Linux, amanda will use xfsdump when appropriate.  The 
problem (we think) is that Dan had tried to use the RPMs, then hadn't 
quite cleaned them out before trying his own version.

> Use tar instead, its *lots* more dependable anyway.
> 
Than Linux dump -- yep.  Than xfsdump?  Well, xfsdump may be weird, but 
it's got years of service on IRIX.

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




Re: XFS, Linux, and Amanda

2002-01-15 Thread Paul Lussier


In a message dated: Mon, 14 Jan 2002 18:41:04 EST
"Brandon D. Valentine" said:

>On Mon, 14 Jan 2002, Joshua Baker-LePain wrote:
>
>>Hmmm, you did 'rpm -e' the RPM version, right?  Pre-build amanda=bad.
>
>Word.  Especially the moronic way in which RedHat has decided to build it.

I'm glad I'm not the only one who thought this!  I've had arguments 
with others over this issue, where they feel that you should always 
use the pre-build package if there's one available!

-- 

Seeya,
Paul


  God Bless America!

 If you're not having fun, you're not doing it right!

...we don't need to be perfect to be the best around,
and we never stop trying to be better. 
   Tom Clancy, The Bear and The Dragon





Re: XFS, Linux, and Amanda

2002-06-23 Thread Gene Heskett

On Monday 14 January 2002 04:17 pm, Dan Rich wrote:
>I am going slowly nuts (ok, so I'm already there :) trying to
> get Amanda to backup my XFS partitions.  I've rebuilt from
> source (CVS), configured it with XFS, and even used strings on
> the executables to prove that the xfsdump command is making it
> in there.  However, I still get the following in my nightly
> backups:
>
>FAILURE AND STRANGE DUMP SUMMARY:
>morpheus   hde11 lev 0 FAILED [disk hde11 offline on morpheus?]
>morpheus   hde10 lev 0 FAILED [disk hde10 offline on morpheus?]
> and so on 
>
>Any ideas of what I can look for?  I can't find anything in the
> list archives or the Yahoo Amanada group that seems to help.
>
>Thanks!

Humm, what I see is that you are using linux dump, specifically 
built for the ext2 filesystem, on a XFS filesystem?  Donnbesilly.

Use tar instead, its *lots* more dependable anyway.

-- 
Cheers, Gene
AMD K6-III@500mhz 320M
Athlon1600XP@1400mhz  512M
98.3+% setiathome rank, not too shabby for a hillbilly