Digging further in amdump log, I think amanda fails to get
estimate size for the shares (shows 0) which are seen failing
below.
Here is part of what amdump log looks like:
setup_estimate: localhost:/h: command 0, options:
last_level -1 next_level0 -12341 level_days 0
getting estimates 0 (0) -1 (-1) -1 (-1)
machine2:/h overdue 12341 days for level 0
setup_estimate: machine2:/h: command 0, options:
last_level -1 next_level0 -12341 level_days 0
getting estimates 0 (0) -1 (-1) -1 (-1)
machine2://windows/users overdue 12341 days for level 0
setup_estimate: machine2://windows/users: command 0, options:
last_level -1 next_level0 -12341 level_days 0
getting estimates 0 (0) -1 (-1) -1 (-1)
planner: time 10.321: setting up estimates for machine2:/vw
setup_estimate: machine2:/vw: command 0, options:
last_level 1 next_level0 8 level_days 2
getting estimates 0 (1366230) 1 (104580) 2 (0)
etc
etc
..
Does that have something to do with the frequent failing?
Amanda staticstic:
STATISTICS:
Total Full Daily
-------- -------- --------
Estimate Time (hrs:min) 0:24
Run Time (hrs:min) 0:51
Dump Time (hrs:min) 0:24 0:18 0:06
Output Size (meg) 978.2 840.1 138.1
Original Size (meg) 3913.9 2426.7 1487.2
Avg Compressed Size (%) 25.0 34.6 9.3 (level:#disks ...)
Filesystems Dumped 9 5 4 (1:3 2:1)
Avg Dump Rate (k/s) 693.0 786.4 402.3
Tape Time (hrs:min) 0:21 0:17 0:03
Tape Size (meg) 978.6 840.3 138.3
Tape Used (%) 50.8 43.6 7.2 (level:#disks ...)
Filesystems Taped 9 5 4 (1:3 2:1)
Avg Tp Write Rate (k/s) 809.1 826.4 717.7
Which says that it just wrote about 2 GB of data. But rest of
tape still has approx 10 GB blank. What makes amanda fail to
do full dumps when there is enough space in the tape for atleast
one to two full filesystem backups?
Can somebody please throw some light here?
Thanks!
Rohit
----- Original Message -----
From: Rohit
To: amanda-users AT amanda DOT org
Sent: Wednesday, October 15, 2003 4:03 PM
Subject: some imp partitions are not being backed up.
Guys:
I put amanda into use almost for 10 days now. Thanks to everyone who helped
me out. One
problem though: I still see few shares not being backed up regularly. All
the partitions
average around 8 to 10 GB of data and the tape which I use is HP DAT 12/24
GB. Any reason
why amanda wouldn't backup these shares? I also tried setting up priority to
high. But that
didn't help either (other high priority filesystems are being backed up
fine). For full
(level 0) dumps amanda complains "FAILED [dump larger than tape, but cannot
incremental
dump new disk". For INC (level > 0) it complains: FAILED [dump larger than
tape,
skipping incremental]
Here is the amoverview part:
date 10 10 10 10 10 10 10 10 10
host disk 07 08 09 10 11 12 13 14 15
machine1 /h 0 E E E E 1 2
machine1 /vl E 0 E E E E 0
machine1 /vm 0 1 1 0 0 0 1
machine1 /vw E 0 E 1 2 2 0
machine2 //windows/UsersG2 0 E 1 E E
machine2 //windows/UsersG3 E E E E E
machine2 //windows/users E E E E E
machine2 /h E E E E E E
machine2 /h2 E E E E E E
machine2 /h3 E E E E E E
machine2 /vl E E 0 E 0 1
machine2 /vw E E 0 E 1 1
localhos /h E E E E E E E
localhos /rp E 0 E 1 2 0 1
localhos /vl 0 1 1 0 1 0 1
localhos /vw 0 1 E 0 1 1 1
Snippet of my amamda.conf:
inparallel 4
netusage 600 Kbps
dumpcycle 2 weeks
runspercycle 5 weeks
tapecycle 20 tapes
Also, I forgot to set record option to 'yes' after I moved amanda to
production. I realized this
two three days ago and set it back to 'yes'. Could this be one of the
reasons? Amanda never did
level 0 backups for most of the partitions and these partitions hold lot of
our critical data.
I'm getting little concerned here.
Could anyone help me here?
Thanks a lot!
Rohit
|
