dump-users

Re: [Dump-users] compression

[Bernhard E. <be...@be...>]

Chuck Pierce wrote:
> 
> is there a way to compress a partition before you put it on to tape??  I'd like
> to keep all of my data on one tape, and I don't really care how long it takes..
> thanks - Chuck

/sbin/dump 0af - /partition | gzip -c | dd obs=32k of=$TAPE

[Dump-users] compression

[Chuck P. <ch...@on...>]

is there a way to compress a partition before you put it on to tape??  I'd like
to keep all of my data on one tape, and I don't really care how long it takes.. 
thanks - Chuck

Re: [Dump-users] RE: Dump HOWTO

[Stelian P. <st...@ca...>]

On Mon, 26 Jun 2000, Linas Vepstas wrote:

> If I read between the lines, there seems to be no way to
> put two or more filessystems on one tape.   

You're wrong. This was already discussed on this mailing list
one or two months ago...

> Yet I want to
> to this ... how? what's teh best way? 

Use the no rewinding tape device (such as /dev/nst0) and:

	mt -f /dev/nst0 rewind
	dump 0f /dev/nst0 /dev/sda1
	dump 0f /dev/nst0 /dev/sda2
	dump 0f /dev/nst0 /dev/sda3
	mt -f /dev/nst0 rewind

In order to do a restore, first go to the wanted place of the
tape (mt -f /dev/nst0 fsf 1 or fsf 2 or fsf 3), then do your
preferred flavour or restore ( -i, -r, -x etc).

> Amanda seems like
> overkill ...  (a hint on the manpage about this topic would
> be nice)

Yes, it would :)

Stelian.
-- 
Stelian Pop <sp...@ca...>| Too many things happened today
Captimark                      | Too many words I don't wanna say
Paris, France                  | I wanna be cool but the heat's coming up
                               | I'm ready to kill 'cause enough is enough
PGP key available on request   |              (Accept - "Up To The Limit")

[Dump-users] RE: Dump HOWTO

[Linas V. <li...@li...>]

Hi,

If I read between the lines, there seems to be no way to
put two or more filessystems on one tape.   Yet I want to
to this ... how? what's teh best way? Amanda seems like
overkill ...  (a hint on the manpage about this topic would
be nice)

--linas

[Dump-users] Re: linux dump failures

[John R. J. <jr...@ga...>]

>|   DUMP: master/slave protocol botched.

That probably means one of the dump child processes died.  Start looking
for core files (try /tmp/amanda first).

John R. Jackson, Technical Software Specialist, jr...@pu...

[Dump-users] linux dump failures

[Troy E. A. <ch...@me...>]

I'm using 2.4.2-19991216-beta1 of amanda with linux dump 0.4b17 (had
0.4b15 and have used it without this problem in the past), but now it's
running on kernel 2.4.0-test1-ac18 and I'm getting this kind of failure on
several filesystems, all of them have > 2G of data on them, but some of
them fail immediately.

I'm cc'ing the linux-kernel list just in case someone knows of any
fundamental changes in dev kernels that would cause something like this.

Thanks.

/-- moonlight  /home/media/misc lev 0 FAILED [/sbin/dump returned 3]
sendbackup: start [moonlight:/home/media/misc level 0]
sendbackup: info BACKUP=/sbin/dump
sendbackup: info RECOVER_CMD=/sbin/restore -f... -
sendbackup: info end
|   DUMP: Date of this level 0 dump: Tue Jun 20 05:38:27 2000
|   DUMP: Date of last level 0 dump: the epoch
|   DUMP: Dumping /dev/mvg0/mmed (/home/media/misc) to standard output
|   DUMP: Label: none
|   DUMP: mapping (Pass I) [regular files]
|   DUMP: mapping (Pass II) [directories]
|   DUMP: estimated 6011317 tape blocks.
|   DUMP: Volume 1 started at: Tue Jun 20 05:39:07 2000
|   DUMP: dumping (Pass III) [directories]
|   DUMP: dumping (Pass IV) [regular files]
|   DUMP: 2.43% done, finished in 3:21
|   DUMP: 5.49% done, finished in 2:52
|   DUMP: 8.63% done, finished in 2:38
|   DUMP: 11.79% done, finished in 2:29
|   DUMP: 14.96% done, finished in 2:22
|   DUMP: 17.89% done, finished in 2:17
|   DUMP: 21.08% done, finished in 2:11
|   DUMP: 24.22% done, finished in 2:05
|   DUMP: 27.39% done, finished in 1:59
|   DUMP: 30.54% done, finished in 1:53
|   DUMP: 33.71% done, finished in 1:48
|   DUMP: 36.87% done, finished in 1:42
|   DUMP: 40.02% done, finished in 1:37
|   DUMP: 43.16% done, finished in 1:32
|   DUMP: 46.30% done, finished in 1:26
|   DUMP: 49.46% done, finished in 1:21
|   DUMP: 52.60% done, finished in 1:16
|   DUMP: 55.73% done, finished in 1:11
|   DUMP: 58.86% done, finished in 1:06
|   DUMP: 62.02% done, finished in 1:01
|   DUMP: 65.17% done, finished in 0:56
|   DUMP: 68.31% done, finished in 0:51
|   DUMP: 71.45% done, finished in 0:45
|   DUMP: 74.62% done, finished in 0:40
|   DUMP: 77.78% done, finished in 0:35
|   DUMP: master/slave protocol botched.
|   DUMP: The ENTIRE dump is aborted.
sendbackup: error [/sbin/dump returned 3]
\--------

Re: [Dump-users] Dump HOWTO

[Stelian P. <po...@cy...>]

On Sun, 28 May 2000, José Juan Pina Camacho wrote:

> I'm new at this list and I'm looking forward a dump HOWTO wich enabled me to make a document. Any of you knows where I can fin smth interesting (apart from the man pages) ?

There is currently a lack of dump documentation (except if you
can read Japanese :) ).

But the man pages are rather complete and explicit, and since dump
is a UNIX standard tool, you can rely on any other documentation
you can find (not linux specific), like a good UNIX book...

Stelian.
-- 
    /\ 
   /  \     Stelian Pop
  / DS \                        Email: po...@cy...
  \____/ 

[Dump-users] Dump HOWTO

[<jos...@wa...>]

Hi!=20

I'm new at this list and I'm looking forward a dump HOWTO wich enabled =
me to make a document. Any of you knows where I can fin smth interesting =
(apart from the man pages) ?

Thanks in advance,

Jose Juan

Re: [Dump-users] Req: restore C success report

[Stelian P. <po...@cy...>]

On Mon, 17 Apr 2000, Kenneth Porter wrote:

> I back up 3 partitions to a tape, rewind it, and then verify the 3
> partitions with "restore C". I'd like to see a success report at the
> end of each verify, if all files compared equal, so I know that restore
> definitely didn't exit silently with an error. (Peace of mind issue.)

This should be rather simple to add to restore, but you should be
aware that this will work only if you are dumping a whole filesystem,
not a directory of a filesystem (because you will have 'file not found
on tape errors', inherent to the dump design).

Maybe you should analyse the restore output with grep...

> Also, when I perform the verify pass, I have to issue 5 restore
> operations to verify 3 partitions. I have to issue a dummy "restore Cf
> /dev/nst0" between the real ones, which responds immediately with
> "restore: tape read error: Success". What's up with that? The system is
> Red Hat 6.1 with a 2.2.14 kernel from the 6.2 distribution. (It's done
> this since 5.2, though.) The drive is an HP SureStore DDS3 DAT
> (12/24GB). The SCSI controller is an Adaptec AIC7xxx controller.

This seems to me like a module problem... try to insmod the module
before launching dump...

Stelian.
-- 
    /\ 
   /  \     Stelian Pop
  / DS \                        Email: po...@cy...
  \____/ 

[Dump-users] Re: illegal tape size

[Rob S. <rsi...@me...>]

Thanks!

Robert Simmons
Systems Administrator
http://www.wlcg.com/

On Tue, 23 May 2000, Bernhard Erdmann wrote:

> Rob Simmons wrote:
> > 
> > I just tried to run dump into /dev/null with the same args that amanda
> > gives dump (except for the "u") and I get the following error:
> > $ dump -0sf 1048576 /dev/null /dev/hdb1
> > dump: illegal tape size -- f
> 
> Don't use "-" before your parameter list:
> 
> # /sbin/dump -0sf 1048576 /dev/null /dev/hdc5
> /sbin/dump: illegal tape size -- f
> 
> # /sbin/dump 0sf 1048576 /dev/null /dev/hdc5
>   DUMP: Date of this level 0 dump: Tue May 23 07:45:07 2000
>   DUMP: Date of last level 0 dump: the epoch
>   DUMP: Dumping /dev/hdc5 (/) to /dev/null
>   DUMP: Label: none
>   DUMP: mapping (Pass I) [regular files]
>   DUMP: mapping (Pass II) [directories]
>   DUMP: estimated 63332 tape blocks on 0.00 tape(s).
>   DUMP: Volume 1 started at: Tue May 23 07:45:07 2000
>   DUMP: dumping (Pass III) [directories]
>   DUMP: dumping (Pass IV) [regular files]
>   DUMP: Closing /dev/null
>   DUMP: Volume 1 completed at: Tue May 23 07:45:16 2000
>   DUMP: Volume 1 took 0:00:09
>   DUMP: Volume 1 transfer rate: 7088 KB/s
>   DUMP: 63792 tape blocks (62.30MB) on 1 volume(s)
>   DUMP: finished in 9 seconds, throughput 7088 KBytes/sec
>   DUMP: Date of this level 0 dump: Tue May 23 07:45:07 2000
>   DUMP: Date this dump completed:  Tue May 23 07:45:16 2000
>   DUMP: Average transfer rate: 7088 KB/s
>   DUMP: DUMP IS DONE
> 
> 
> Look carefully to your sendbackup.debug file:
> 
> sendbackup: spawning "/sbin/dump" in pipeline
> sendbackup: argument list: "dump" "0usf" "1048576" "-" "/dev/hdb1"
> 

[Dump-users] Re: illegal tape size

[Bernhard E. <be...@be...>]

Rob Simmons wrote:
> 
> I just tried to run dump into /dev/null with the same args that amanda
> gives dump (except for the "u") and I get the following error:
> $ dump -0sf 1048576 /dev/null /dev/hdb1
> dump: illegal tape size -- f

Don't use "-" before your parameter list:

# /sbin/dump -0sf 1048576 /dev/null /dev/hdc5
/sbin/dump: illegal tape size -- f

# /sbin/dump 0sf 1048576 /dev/null /dev/hdc5
  DUMP: Date of this level 0 dump: Tue May 23 07:45:07 2000
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/hdc5 (/) to /dev/null
  DUMP: Label: none
  DUMP: mapping (Pass I) [regular files]
  DUMP: mapping (Pass II) [directories]
  DUMP: estimated 63332 tape blocks on 0.00 tape(s).
  DUMP: Volume 1 started at: Tue May 23 07:45:07 2000
  DUMP: dumping (Pass III) [directories]
  DUMP: dumping (Pass IV) [regular files]
  DUMP: Closing /dev/null
  DUMP: Volume 1 completed at: Tue May 23 07:45:16 2000
  DUMP: Volume 1 took 0:00:09
  DUMP: Volume 1 transfer rate: 7088 KB/s
  DUMP: 63792 tape blocks (62.30MB) on 1 volume(s)
  DUMP: finished in 9 seconds, throughput 7088 KBytes/sec
  DUMP: Date of this level 0 dump: Tue May 23 07:45:07 2000
  DUMP: Date this dump completed:  Tue May 23 07:45:16 2000
  DUMP: Average transfer rate: 7088 KB/s
  DUMP: DUMP IS DONE


Look carefully to your sendbackup.debug file:

sendbackup: spawning "/sbin/dump" in pipeline
sendbackup: argument list: "dump" "0usf" "1048576" "-" "/dev/hdb1"

[Dump-users] problem with dump0.4p16

[Rob S. <rsi...@me...>]

whatever value I send to dump in the -s arg returns "illegal tape size"

$ dump -0sf <any number here> /dev/null /dev/hda1
dump: illegal tape size -- f

Is there an ./configure option that I missed, or is this a bug in dump?

Robert Simmons
Systems Administrator
http://www.wlcg.com/

Re: [Dump-users] master/slave protocol botched

[Stelian P. <po...@cy...>]

On Sat, 13 May 2000, Tim Thornley wrote:

> Hi,
> 
> I am using a rather old version of DUMP on RedHat 4.2 and for some reason I
> have started getting the following error on one file system.
> 
> Is this a bug that has been fixed in newer versions or am I missing the
> meaning of "master/slave protocol botched" If so can you please explain it
> to me.
> 

This reminds me of a very old bug, which was corrected a while ago.

Try the latest version, it should work (and MANY bugs were fixed,
many features added since dump 0.3 - which I believe your version is).


Stelian.
-- 
    /\ 
   /  \     Stelian Pop
  / DS \                        Email: po...@cy...
  \____/ 

[Dump-users] master/slave protocol botched

[Tim T. <tho...@we...>]

Hi,

I am using a rather old version of DUMP on RedHat 4.2 and for some reason I
have started getting the following error on one file system.

DUMP: Date of this level 1 dump: Sat May 13 05:53:21 2000
DUMP: Date of last level 0 dump: Wed Apr  5 06:06:13 2000
DUMP: Dumping /dev/hdb1 (/home) to /backup/sehm.1.1305.00
DUMP: mapping (Pass I) [regular files]
DUMP: mapping (Pass II) [directories]
DUMP: estimated 1199274 tape blocks on 1.67 tape(s).
DUMP: dumping (Pass III) [directories]
DUMP: master/slave protocol botched.
DUMP: The ENTIRE dump is aborted.

Is this a bug that has been fixed in newer versions or am I missing the
meaning of "master/slave protocol botched" If so can you please explain it
to me.

Thanks
Tim

----------------------------------------------------------------------------
--

[Dump-users] suggestion for a dump schedule

[Eros A. <er...@la...>]

I am seeking suggestions, clue or comments for a dump schedule that should be a
compromise between workload and safety.

At present I am using a 28-day alternative differential (taken from the manual
examples of Digital unix)
0  5  4  7  6  9  8
1  5  4  7  6  9  8
2  5  4  7  6  9  8
3  5  4  7  6  9  8
with the dump at level 0 done on tape (manually) and the incrementals on disk
through a cron scripts.
-- 
Eros Albertazzi
CNR-LAMEL, Via P.Gobetti 101, 40129 Bologna, Italy  
Tel: (+39)-051-639 9179
Fax: (+39)-051-639 9216
E-mail: er...@la...

Re: [Dump-users] How do I get multiple partitions on one DUMP

[Bernhard E. <be...@be...>]

TJ McNeely wrote:
> 
> I am sorry if this has been asked and answered already, I am new to the
> list.
> 
> [tjm@ted tjm]$ df
> Filesystem           1k-blocks      Used Available Use% Mounted on
> /dev/hda2               101517     41031     55245  43% /
> /dev/hdc1             16682168   3778868  12055876  24% /backup
> /dev/hda6               300603     42113    242969  15% /home
> /dev/hda5              1511968    844516    590644  59% /usr
> /dev/hda7              1051632     12156    986056   1% /var
> 
> Here is the DUMP command we attempted to use:
>         dump -0ua -b 32 -f /dev/ht0 /
> 
> We got a nice looking directory structure, but no files :)

Why do you use a blocksize of 32 KB? Is it necessary for your tape
drive? How did you verify your "directory structure, but no files"? Does
/etc/dumpdates exist? Did you actually restore / from tape? Did you
issue "mkdir -p /tmp/restored; cd /tmp/restored; restore -rf /dev/ht0;
du ."?

> I want to back the entire system "/" up to a tape... it worked fine on my
> other server cause it was all one partition, this one only backed up
> /dev/hda2 .. is there is slice 2 like in Solaris that gets the WHOLE DRIVE..
> actually I need to get the backup partiton too. Any help would be
> appreciated.

You didn't fully unterstand the concept of dump yet. Dump works at
inode-level on a partition. Each partition has to be dumped separately.
You'd like to backup four partitions? So go and dump four partitions.
export $TAPE=/dev/(n?)ht0 # the non-rewinding tape device, check for it!
dump 0ua /
dump 0ua /home
dump 0ua /var
dump 0ua /usr

> I would be willing to get all of the hda drive backed up to the hdc drive
> (if possible)... then just do all of HDC --> tape..

dump 0uaf /backup/root-20000418.dump /
dump 0uaf /backup/home-20000418.dump /home
...

If you have a 2 GB file limit on your system, you have to use the -M
option (see "man dump").

Another possibility is to use gzip on the fly:
dump 0uaf - /home | gzip -c > /backup/home-20000418.dump.gz

You can write the dump images with dd to your tape drive:
cd /backup
for i in *.dump
do
dd if=$i of=/dev/(n?)ht0	# use the non-rewinding drive!
done

Compressed dump images can also be restored from tape:
dd if=$TAPE | gzip -dc | restore -rf -
(Maybe you have to specify of=- to dd, it depends on your OS and dd's
version.)

How do you forward and rewind your tape drive? Can you use mt? From the
device /dev/ht0 I assume you have an IDE tape.

> PS: also, and I feel stupid asking this, but how do I make a cron run only
> on SUNDAY night (well Monday morning).. right now our server (bill) does a
> remote backup to a file on ted every morning at 2:30.. I really only need it
> once a week. Thanks!

0 1 * * 1 /usr/local/bin/backup-me-now
will run your script every monday morning 1:00 am

"man 5 crontab" says:

       The time and date fields are:

              field          allowed values
              -----          --------------
              minute         0-59
              hour           0-23
              day of month   1-31
              month          1-12 (or names, see below)
              day of week    0-7 (0 or 7 is Sun, or use names)

[Dump-users] Req: restore C success report

[Kenneth P. <sh...@we...>]

I back up 3 partitions to a tape, rewind it, and then verify the 3
partitions with "restore C". I'd like to see a success report at the
end of each verify, if all files compared equal, so I know that restore
definitely didn't exit silently with an error. (Peace of mind issue.)
For two of the partitions (/ and /var), I expect to get a few
miscompares, but the /home partition is normally quiet at night when
the backup runs, and I'd like to know by a positive message that the
compare completed successfully.

Also, when I perform the verify pass, I have to issue 5 restore
operations to verify 3 partitions. I have to issue a dummy "restore Cf
/dev/nst0" between the real ones, which responds immediately with
"restore: tape read error: Success". What's up with that? The system is
Red Hat 6.1 with a 2.2.14 kernel from the 6.2 distribution. (It's done
this since 5.2, though.) The drive is an HP SureStore DDS3 DAT
(12/24GB). The SCSI controller is an Adaptec AIC7xxx controller.

Ken
mailto:sh...@we...
http://www.sewingwitch.com/ken/
http://www.harrybrowne2000.org/

[Dump-users] How do I get multiple partitions on one DUMP

[TJ M. <tj...@di...>]

I am sorry if this has been asked and answered already, I am new to the
list.

[tjm@ted tjm]$ df
Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/hda2               101517     41031     55245  43% /
/dev/hdc1             16682168   3778868  12055876  24% /backup
/dev/hda6               300603     42113    242969  15% /home
/dev/hda5              1511968    844516    590644  59% /usr
/dev/hda7              1051632     12156    986056   1% /var


Here is the DUMP command we attempted to use:
	dump -0ua -b 32 -f /dev/ht0 /


We got a nice looking directory structure, but no files :)

I want to back the entire system "/" up to a tape... it worked fine on my
other server cause it was all one partition, this one only backed up
/dev/hda2 .. is there is slice 2 like in Solaris that gets the WHOLE DRIVE..
actually I need to get the backup partiton too. Any help would be
appreciated.

I would be willing to get all of the hda drive backed up to the hdc drive
(if possible)... then just do all of HDC --> tape..

Thanks,

Knop Head

PS: also, and I feel stupid asking this, but how do I make a cron run only
on SUNDAY night (well Monday morning).. right now our server (bill) does a
remote backup to a file on ted every morning at 2:30.. I really only need it
once a week. Thanks!

[Dump-users] Solution to the 2GB issue with 32bit E2fs

[Knop <kn...@di...>]

This is only for dumping to a file. tapes are fine with large files... at
least I have not come across any issues (yet).


here is my "backup" script.. it backs up the entire / partition to a
server named ted (which in turn backs itself up to tape)


root@bill:~:# cat bin/backup 
#!/bin/bash
printf "\n\n --- START FULL BACKUP --- \n\n\n"
/sbin/dump -0uM -B 2000000 -f ted:/backup/bill.dump /dev/hda1
printf "\n\n --- END --- \n\n"
root@bill:~:#  


that basically limits dump to making 2 GB files (well a lttle less to
prevent issues)... I have, however, had issues with dump actually creating
the remote file... it is fine if I touch the file on ted ahead of time.

Re: [Dump-users] dump/restore on multiple partitions on one tape

[<cco...@sw...>]

On Thu, Apr 13, 2000 at 01:37:27PM -0500, chuck wrote:
> Bernhard Erdmann wrote:
> > 
> > On Thu, Apr 13, 2000 at 01:21:12PM -0500, chuck wrote:
> > > anyone know how to dump/restore multiple partitions onto one tape?? - Chuck
> > 
> > No experience yet, but I'd simply use several dumps in serial order
> > onto a nonrewinding tape drive.
> 
> 
> yea, but if you do that, how do you do an interactive restore (restore -i)??

You'd just use the mt commands (or their equivalents) between restore sessions:

mt rewind
restore -i
..
..
quit
mt fsf 1
restore -i
..
..
quit

etc...

best,
c

Re: [Dump-users] 2 GB file limit

[Bernhard E. <be...@be...>]

On Thu, Apr 13, 2000 at 06:26:45PM -0600, TJ McNeely wrote:
> I for some reason cannot dump a file bigger than 2 GB.. is this a DUMP
> limitation or an E2FS limitation? or am I crazy... wait don't answer that :)

Your e2fs on a 32 bit machine is limited to 2 GB filesize. Look at option -M in "man dump" to bypass it. (dump 0.4b16)

[Dump-users] 2 GB file limit

[TJ M. <tj...@di...>]

I for some reason cannot dump a file bigger than 2 GB.. is this a DUMP
limitation or an E2FS limitation? or am I crazy... wait don't answer that :)


Knop Head

Re: [Dump-users] dump/restore on multiple partitions on one tape

[chuck <ch...@on...>]

Bernhard Erdmann wrote:
> 
> On Thu, Apr 13, 2000 at 01:21:12PM -0500, chuck wrote:
> > anyone know how to dump/restore multiple partitions onto one tape?? - Chuck
> 
> No experience yet, but I'd simply use several dumps in serial order
> onto a nonrewinding tape drive.


yea, but if you do that, how do you do an interactive restore (restore -i)??

[Dump-users] dump/restore on multiple partitions on one tape

[chuck <ch...@on...>]

anyone know how to dump/restore multiple partitions onto one tape?? - Chuck

Re: [Dump-users] exabyte 8500 parameters

[Stelian P. <po...@cy...>]

On Wed, 5 Apr 2000, Greg Thompson wrote:

> what are reasonable values for blocksize, density, and feet for one of
> these?  i have a legacy dumping script that uses 126, 54000, and 13000,
> respectively, but none of those numbers make sense to me.  are density and
> feet only used by dump to calculate how much data it can shove on a tape?
> should i just use "-a" rather than try to figure them out?

They are used only to calculate the amount of data a tape will
manage. So you can safely use -a for that (or -B).

> blocksize: what value should i be using?  does it matter?  will i get better
> throughput to the tapes if i pick some magic blocksize that's just right for
> this drive?

Exactly. The best to do is to experiment with different values,
setting your tape drive to variable blocksize and fixed blocksize
(if it accepts those settings), and see by yourself...

Generally, a bigger blocksize gives better performance. But if
a tape block is corrupt, you will lose more data...

> density: i'm using mt to set the density code for the drive to 21 (0x15),
> but i can't find anyplace that'll tell me what that corresponds to in bpi.
> 
> length: even though exabyte recommends against it, i'm using 160m XL tapes.
> if m stands for meters, then that works out to just shy of 525 feet, which
> is nowhere near the 13000 that someone put in this old script i'm using.
> what's up with that?

Leave the length/density stuff alone. The only interesting settings
are -a (or -B if -a if your tape drive is buggy) and -b.

The length/density were used for old tape drives (and cartridges),
and it's a complete mess. If you are really interested, you can
look at the source code to see how this stuff is calculated. I also
found an interesting page on different tape drive settings somewhere
aroung www.amanda.org...

Hope this made some thing clearer for you (although I'm not
convinced :) ).

Stelian.
-- 
    /\ 
   /  \     Stelian Pop
  / DS \                        Email: po...@cy...
  \____/