PDA

View Full Version : Amanda is ALWAYS doing full backups



el3ktro
December 15th, 2006, 05:31 AM
Hi,

I want to backup around 160GiB of data to 11 tapes total, on weekdays, with a full backup every week. The config file look like this:



dumpcycle 7
runspercycle 5
runtapes 1
tapecycle 11
[code]

This works fine, Amanda backups all the data and I can also restore it. Though, Amanda seems to do a full backup every single day, although it should do this only approximately every 7 days (once a week). The Amanda backup report looks like this:

[code]
These dumps were to tape DATA_DAILY_07.
The next tape Amanda expects to use is: DATA_DAILY_08.


STATISTICS:
Total Full Incr.
-------- -------- --------
Estimate Time (hrs:min) 0:12
Run Time (hrs:min) 2:38
Dump Time (hrs:min) 2:26 2:26 0:00
Output Size (meg) 175202.6 175202.6 0.0
Original Size (meg) 175202.5 175202.5 0.0
Avg Compressed Size (%) -- -- --
Filesystems Dumped 7 7 0
Avg Dump Rate (k/s) 20547.6 20547.6 --

Tape Time (hrs:min) 2:26 2:26 0:00
Tape Size (meg) 175202.8 175202.8 0.0
Tape Used (%) 92.2 92.2 0.0
Filesystems Taped 7 7 0

Chunks Taped 0 0 0
Avg Tp Write Rate (k/s) 20517.9 20517.9 --

USAGE BY TAPE:
Label Time Size % Nb Nc
DATA_DAILY_07 2:26 175203M 92.2 7 0


NOTES:
planner: disk amanda:/usr/local/bin, estimate of level 1 failed.
planner: disk amanda:/etc, estimate of level 1 failed.
planner: disk amanda:/srv/data, estimate of level 1 failed.
planner: Full dump of technikserver:/var/www/wiki promoted from 9 days ahead.
planner: Full dump of technikserver:/var/lib/dokuwiki promoted from 9 days ahead.
planner: Full dump of technikserver:/var/lib/mysql promoted from 9 days ahead.
planner: Full dump of amanda:/etc promoted from 9 days ahead.
planner: Full dump of technikserver:/etc/dokuwiki promoted from 9 days ahead.
planner: Full dump of amanda:/usr/local/bin promoted from 9 days ahead.
taper: tape DATA_DAILY_07 kb 179407712 fm 7 [OK]


DUMP SUMMARY:
DUMPER STATS TAPER STATS
HOSTNAME DISK L ORIG-MB OUT-MB COMP% MMM:SS KB/s MMM:SS KB/s
-------------------------- ------------------------------------- -------------
amanda /etc 0 4 4 -- 0:01 5551.8 0:02 2239.8
amanda /srv/data 0 175153 175153 -- 145:26 20554.2 145:26 20554.1
amanda -/local/bin 0 0 0 -- 0:00 36.7 0:05 13.1
technikserve -c/dokuwiki 0 0 0 -- 0:00 1597.6 0:02 190.9
technikserve -b/dokuwiki 0 1 1 -- 0:00 9355.9 0:02 558.8
technikserve -/lib/mysql 0 29 29 -- 0:03 11419.0 0:02 13721.0
technikserve -r/www/wiki 0 16 16 -- 0:01 11265.4 0:05 3198.8

(brought to you by Amanda version 2.5.0p2)


I'm especially concerned about these "estimate of level 1 failed" messages. What exactly do they mean? How can I get around this?

Tom

paddy
December 17th, 2006, 07:43 PM
Amanda attempts to as many full backups as possible in a dump cycle.
In your case estimation for level 1 is failing and level 0 dumps fit into the
backup media space available.

You need to investigate by level 1 estimation is failing. Take a look at
sendsize logs on server "amanda".

If you need help, please post information from /tmp/amanda/sendsize* logs for
the backup run.

Thanks,
Paddy

el3ktro
December 19th, 2006, 06:41 PM
Hello,

I appreciate your help. The backup ran again today, and it did a full backup again. Here's the latest sendsize logfile you where asking for:



# cat /var/log/amanda/sendsize.20061219183350.debug
sendsize: debug 1 pid 19608 ruid 34 euid 34: start at Tue Dec 19 18:33:50 2006
sendsize: version 2.5.0p2
sendsize[19610]: time 0.114: calculating for amname '/srv/data', dirname '/srv/data', spindle -1
sendsize[19610]: time 0.114: getting size via gnutar for /srv/data level 0
sendsize[19608]: time 0.115: waiting for any estimate child: 1 running
sendsize[19610]: time 0.115: spawning /usr/lib/amanda/runtar in pipeline
sendsize[19610]: argument list: /bin/tar --create --file /dev/null --directory /srv/data --one-file-system --numeric-owner --listed-incremental /var/lib/amanda/gnutar-lists/amanda_srv_data_0.new --sparse --ignore-failed-read --totals .
sendsize[19610]: time 707.163: Total bytes written: 181917767680 (170GiB, 246MiB/s)
sendsize[19610]: time 707.164: .....
sendsize[19610]: estimate time for /srv/data level 0: 707.048
sendsize[19610]: estimate size for /srv/data level 0: 177654070 KB
sendsize[19610]: time 707.164: waiting for /bin/tar "/srv/data" child
sendsize[19610]: time 707.166: after /bin/tar "/srv/data" wait
sendsize[19610]: time 707.190: getting size via gnutar for /srv/data level 1
sendsize[19610]: time 717.525: spawning /usr/lib/amanda/runtar in pipeline
sendsize[19610]: argument list: /bin/tar --create --file /dev/null --directory /srv/data --one-file-system --numeric-owner --listed-incremental /var/lib/amanda/gnutar-lists/amanda_srv_data_1.new --sparse --ignore-failed-read --totals .
sendsize[19610]: time 717.665: /bin/tar: Unexpected field value in snapshot file
sendsize[19610]: time 717.665: /bin/tar: Error is not recoverable: exiting now
sendsize[19610]: time 717.665: .....
sendsize[19610]: estimate time for /srv/data level 1: 0.140
sendsize[19610]: no size line match in /bin/tar output for "/srv/data"
sendsize[19610]: .....
sendsize[19610]: estimate size for /srv/data level 1: -1 KB
sendsize[19610]: time 717.665: waiting for /bin/tar "/srv/data" child
sendsize[19610]: time 717.665: after /bin/tar "/srv/data" wait
sendsize[19610]: time 717.666: done with amname '/srv/data', dirname '/srv/data', spindle -1
sendsize[19608]: time 717.666: child 19610 terminated normally
sendsize[19613]: time 717.666: calculating for amname '/etc', dirname '/etc', spindle -1
sendsize[19613]: time 717.666: getting size via gnutar for /etc level 0
sendsize[19608]: time 717.667: waiting for any estimate child: 1 running
sendsize[19613]: time 717.748: spawning /usr/lib/amanda/runtar in pipeline
sendsize[19613]: argument list: /bin/tar --create --file /dev/null --directory /etc --one-file-system --numeric-owner --listed-incremental /var/lib/amanda/gnutar-lists/amanda_etc_0.new --sparse --ignore-failed-read --totals .
sendsize[19613]: time 718.018: Total bytes written: 3901440 (3.8MiB, 15MiB/s)
sendsize[19613]: time 718.018: .....
sendsize[19613]: estimate time for /etc level 0: 0.270
sendsize[19613]: estimate size for /etc level 0: 3810 KB
sendsize[19613]: time 718.018: waiting for /bin/tar "/etc" child
sendsize[19613]: time 718.018: after /bin/tar "/etc" wait
sendsize[19613]: time 718.018: getting size via gnutar for /etc level 1
sendsize[19613]: time 718.053: spawning /usr/lib/amanda/runtar in pipeline
sendsize[19613]: argument list: /bin/tar --create --file /dev/null --directory /etc --one-file-system --numeric-owner --listed-incremental /var/lib/amanda/gnutar-lists/amanda_etc_1.new --sparse --ignore-failed-read --totals .
sendsize[19613]: time 718.057: /bin/tar: Unexpected field value in snapshot file
sendsize[19613]: time 718.057: /bin/tar: Error is not recoverable: exiting now
sendsize[19613]: time 718.057: .....
sendsize[19613]: estimate time for /etc level 1: 0.004
sendsize[19613]: no size line match in /bin/tar output for "/etc"
sendsize[19613]: .....
sendsize[19613]: estimate size for /etc level 1: -1 KB
sendsize[19613]: time 718.057: waiting for /bin/tar "/etc" child
sendsize[19613]: time 718.057: after /bin/tar "/etc" wait
sendsize[19613]: time 718.057: done with amname '/etc', dirname '/etc', spindle -1
sendsize[19608]: time 718.058: child 19613 terminated normally
sendsize[19616]: time 718.058: calculating for amname '/usr/local/bin', dirname '/usr/local/bin', spindle -1
sendsize[19616]: time 718.058: getting size via gnutar for /usr/local/bin level 0
sendsize[19608]: time 718.058: waiting for any estimate child: 1 running
sendsize[19616]: time 718.075: spawning /usr/lib/amanda/runtar in pipeline
sendsize[19616]: argument list: /bin/tar --create --file /dev/null --directory /usr/local/bin --one-file-system --numeric-owner --listed-incremental /var/lib/amanda/gnutar-lists/amanda_usr_local_bin_0.new --sparse --ignore-failed-read --totals .
sendsize[19616]: time 718.095: Total bytes written: 10240 (10KiB, 637KiB/s)
sendsize[19616]: time 718.095: .....
sendsize[19616]: estimate time for /usr/local/bin level 0: 0.019
sendsize[19616]: estimate size for /usr/local/bin level 0: 10 KB
sendsize[19616]: time 718.095: waiting for /bin/tar "/usr/local/bin" child
sendsize[19616]: time 718.095: after /bin/tar "/usr/local/bin" wait
sendsize[19616]: time 718.096: getting size via gnutar for /usr/local/bin level 1
sendsize[19616]: time 718.112: spawning /usr/lib/amanda/runtar in pipeline
sendsize[19616]: argument list: /bin/tar --create --file /dev/null --directory /usr/local/bin --one-file-system --numeric-owner --listed-incremental /var/lib/amanda/gnutar-lists/amanda_usr_local_bin_1.new --sparse --ignore-failed-read --totals .
sendsize[19616]: time 718.116: /bin/tar: Unexpected field value in snapshot file
sendsize[19616]: time 718.117: /bin/tar: Error is not recoverable: exiting now
sendsize[19616]: time 718.117: .....
sendsize[19616]: estimate time for /usr/local/bin level 1: 0.004
sendsize[19616]: no size line match in /bin/tar output for "/usr/local/bin"
sendsize[19616]: .....
sendsize[19616]: estimate size for /usr/local/bin level 1: -1 KB
sendsize[19616]: time 718.117: waiting for /bin/tar "/usr/local/bin" child
sendsize[19616]: time 718.117: after /bin/tar "/usr/local/bin" wait
sendsize[19616]: time 718.117: done with amname '/usr/local/bin', dirname '/usr/local/bin', spindle -1
sendsize[19608]: time 718.118: child 19616 terminated normally
sendsize: time 718.118: pid 19608 finish time Tue Dec 19 18:45:48 2006


So it seems Amanda cleary has problems determining the size of /srv/data (~175GiB right now) - the question is why!?

Tom

paddy
December 19th, 2006, 06:55 PM
sendsize[19613]: argument list: /bin/tar --create --file /dev/null --directory /etc --one-file-system --numeric-owner --listed-incremental /var/lib/amanda/gnutar-lists/amanda_etc_1.new --sparse --ignore-failed-read --totals .
sendsize[19613]: time 718.057: /bin/tar: Unexpected field value in snapshot file
sendsize[19613]: time 718.057: /bin/tar: Error is not recoverable: exiting now

This is the problem.

What version of tar are you using? Does /var/lib/amanda/gnutar-lists/ directory contain
any information?

If you are using tar 1.16, you might be running into this problem: http://savannah.gnu.org/bugs/?18487

Thanks,
Paddy

el3ktro
December 20th, 2006, 07:40 AM
Hi,

well I see but I'm not backing up /proc or /sys at all. I'm using tar 1.15.91 on Ubuntu Edgy. The directory /var/lib/amanda/gnutar-lists/ contains several files (like "localhost_srv_data_0") containing lots and lots of filenames.

Any ideas what could be wrong?

martineau
December 20th, 2006, 08:02 AM
amanda-2.5.0p2 is incompatible with tar-1.15.91

You should upgrade to amanda-2.5.1p2 or downgrade your tar.

el3ktro
December 20th, 2006, 08:11 AM
Oh, good to know. Thanks a lot! Well I'll downgrade tar then I guess.

Tom

martineau
December 20th, 2006, 08:25 AM
Oh, good to know. Thanks a lot! Well I'll downgrade tar then I guess.


Don't forget to remove all files in /var/lib/amanda/gnutar-lists/