Amanda can't run dump to estimate dump size but works from cmd line..

[William Hargrove]

Hello All,

I'm having a bit of a problem in getting Amanda to work.. specifically it's
failing when estimating the dump sizes when amdump is run. The configuration
is, machine backup RH Linux 2.4.20-30.9 Amanda server with an HP-DAT
drive, this machine is not a client. The client is machine augustus, same
spec as backup but is configured as a client.

amcheck runs fine on backup:

Amanda Tape Server Host Check
-
Holding disk /mnt/hd1/backups: 23141220 KB disk space available, using
21044068 KB
NOTE: skipping tape-writable test
Tape DailySet1-000 label ok
NOTE: info dir /opt/amanda/var/amanda/DailySet1/curinfo: does not exist
NOTE: it will be created on the next run
Server check took 0.048 seconds

Amanda Backup Client Hosts Check

Client check: 1 host checked in 0.144 seconds, 0 problems found

(brought to you by Amanda 2.4.4p2)

My disklist entry for this test is:

augustus..com /etc nocomp-test

amdump fails with the following error taken from
/tmp/amanda/sendsizedebug on augustus.

sendsize: debug 1 pid 26767 ruid 537 euid 537: start at Fri Apr 16 13:48:09
2004
sendsize: version 2.4.4p2
sendsize[26770]: time 0.005: calculating for amname '/etc', dirname '/etc',
spin
dle -1
sendsize[26770]: time 0.006: getting size via dump for /etc level 0
sendsize[26770]: time 0.007: calculating for device '/etc' with ''
sendsize[26770]: time 0.007: running /sbin/dump 0Ssf 1048576 - /etc
sendsize[26770]: time 0.008: running /opt/amanda/libexec/killpgrp
sendsize[26767]: time 0.009: waiting for any estimate child: 1 running
sendsize[26770]: time 0.040:   DUMP: Cannot open /dev/md1
sendsize[26770]: time 0.041:   DUMP: The ENTIRE dump is aborted.
sendsize[26770]: time 0.042: .
sendsize[26770]: estimate time for /etc level 0: 0.035
sendsize[26770]: no size line match in /sbin/dump output for /etc
sendsize[26770]: .
sendsize[26770]: estimate size for /etc level 0: -1 KB
sendsize[26770]: time 0.042: asking killpgrp to terminate
sendsize[26770]: time 1.050: done with amname '/etc', dirname '/etc',
spindle -1
sendsize[26767]: time 1.050: child 26770 terminated normally
sendsize: time 1.050: pid 26767 finish time Fri Apr 16 13:48:10 2004

Now, if I run that command line as the amanda user, it succeds:

[EMAIL PROTECTED] amanda]$ /sbin/dump 0Ssf 1048576 - /etc
  DUMP: Added inode 8 to exclude list (journal inode)
  DUMP: Added inode 7 to exclude list (resize inode)
18119680

18mb or so.. however the total disk size of /dev/md1 is much larger than
this. I've had a look through the know issues and can't see a problem with
the directory sitting on a disk this size.

Many thanks for people looking at this.. I can't see what's wrong with it,
though there is clearly an issue with it.

William Hargrove.

Some further information is below:

The disk layout of the client machine is:

df -k
Filesystem   1K-blocks  Used Available Use% Mounted on
/dev/md1  74807784  49629028  21378728  70% /
/dev/md0 98971 14331 79530  16% /boot
none256296 0256296   0% /dev/shm

/etc/fstab as the follow entries:

more /etc/fstab
/dev/md1/   ext3defaults1 1
/dev/md0/boot   ext3defaults1 2
none/dev/ptsdevpts  gid=5,mode=620  0 0
none/proc   procdefaults0 0
none/dev/shmtmpfs   defaults0 0
/dev/md2swapswapdefaults0 0
/dev/cdrom  /mnt/cdrom  udf,iso9660
noauto,owner,kudzu,ro 0 0
/dev/fd0/mnt/floppy autonoauto,owner,kudzu 0
0

The amanda user is a member of the disk group:

[EMAIL PROTECTED] amanda]$ groups
backup disk

The ownership of /dev/md1 is:

[EMAIL PROTECTED] root]# ls -l /dev/md1
brw-rw1 root disk   9,   1 Jan 30  2003 /dev/md1

and it's constituent disks are:

[EMAIL PROTECTED] root]# ls -l /dev/hde3
brw-rw1 root disk  33,   3 Jan 30  2003 /dev/hde3
[EMAIL PROTECTED] root]# ls -l /dev/hdg3
brw-rw1 root disk  34,   3 Jan 30  2003 /dev/hdg3

The output of amdump is:

These dumps were to tape DailySet1-000.
The next tape Amanda expects to use is: a new tape.

FAILURE AND STRANGE DUMP SUMMARY:
  augustus.k /etc lev 0 FAILED [disk /etc, all estimate failed]


STATISTICS:
  Total   Full  Daily
      
Estimate Time (hrs:min)0:00
Run Time (hrs:min) 0:00
Dump Time (hrs:min)0:00   0:00   0:00
Output Size (meg)   0.00.00.0
Original Size (meg) 0.00.00.0
Avg Compressed Size (%) -- -- --
Filesystems Dumped0  0  0
Avg Dump Rate (k/s

Amanda backup strategy thoughts, full and incremental backups to tape once a week.

[William Hargrove]

Hello All,

I would like to ask a general Amanda backup 'strategy' question here, given
by backup 'goals' below. I have a configured and working Amanda set-up but I
want to try and fit a backup policy that I believe is best suited to our
site around Amanda.

I have an 8-tape (DLT 35/70gb each) auto changer library linked to Amanda
via mtx, this works well.

I want to do a full backup of various file systems on a remote host once a
week and then, ideally, incremental backups once a day to that same tape. At
the end of a week's backup, the tape should switch. One week's backup is far
less than a single tape's capacity.

I would like to be able to hold 8 weeks worth of backups on tape, that is
one tape a week and then reuse the oldest tape.

I'm not sure this is possible with Amanda.. as I believe I can't append the
daily incremental backups onto the same tape as the full backups. Is it the
case that if I want to run Amanda every day that I would also have to use a
tape every day, thus meaning the tapes in the library are all used after a
week. I'm fixed in the number of tapes I can use and can't add/replace them,
so they would get reused after a week and wouldn't provide much of a 'backup
horizon'.

If this isn't possible does anyone have some suggestions? I was thinking I
could do the incremental backups to a hard disk area on the tape server each
night and then run Amanda once a week to archive the full backups plus the
HDD incremental ones.

Many thanks in advance for everyone's thoughts.

William.