Ah!! I found the log.
This is what Amanda said from the last run:
taper: read label `Homenet18' date `X'
taper: wrote label `Homenet18' date `20050310'
planner: time 231.726: got result for host ziyal disk /data2/public: 0
-> 2071100K, 1 -> 2071100K, 2 -> 2071100K
planner: time 231.726: got result for host ziyal disk /data2/media: 0
-> 94800K, 1 -> 94800K, -1 -> -1K
planner: time 231.726: got result for host ziyal disk
/data2/images/group3: 0 -> 12787860K, 1 -> 12787860K, 2 -> 12787860K
planner: time 231.726: got result for host ziyal disk
/data2/images/group2: 0 -> 2160330K, 1 -> 2160330K, -1 -> -1K
planner: time 231.726: got result for host ziyal disk
/data2/images/group1: 0 -> 10131800K, 1 -> 10131800K, 2 -> 10131800K
planner: time 231.726: got result for host ziyal disk /data2/homes: 0
-> 13306280K, 1 -> 13306280K, 2 -> 13306280K
planner: time 231.726: got result for host ziyal disk /data2/downloads:
0 -> 8009890K, 1 -> 8009890K, 2 -> 8009890K
planner: time 231.726: got result for host ziyal disk
/data2/distributions: 0 -> 10K, 1 -> 10K, -1 -> -1K
planner: time 231.726: got result for host ziyal disk /data2/cvsroot: 0
-> 1500K, 1 -> 1500K, -1 -> -1K
planner: time 231.726: got result for host ziyal disk /: 0 -> 2555090K,
1 -> 2555090K, 2 -> 2555090K
planner: time 1455.891: got result for host kira disk /: 0 -> 3971310K,
1 -> 3971310K, 2 -> 3971320K
planner: time 1455.891: getting estimates took 1455.616 secs
Looks like the estimates are the same for every level!
Why would this be?
Thanks in advance!
\marc
On Mar 10, 2005, at 8:36 PM, Kevin Dalley wrote:
Do you have something like this in your amdump file:
planner: time 1817.635: got result for host condor disk //bunting/e$:
0 -> 5524252K, 1 -> 108113K, 2 -> 108113K
It should verify that it calculates the level 0 and level 1 backups.
Is the level 0 actually written entirely to disk (or tape). Is it
possible that the backup fails before the level 0 has been completed?
"Marc N. Cannava" <mnc AT botcave DOT com> writes:
Hi all,
I've been using Amanda for a while with no issue.
I recently moved from RH9 to Gentoo and re-installed amanda with the
latest version available in the ebuild tree (amanda 2.4.4_p3). Using
my same config files and backing up to disk (using chg-disk).
Very, very odd problem now: No matter how many times Amanda runs, she
always calculates the dump sizes as if they are level 0, even when she
wants to do a level 1.
Example:
ziyal -ata2/homes 1 13306200 13306200 -- 13:5615917.9
13:5715903.2
This is supposedly a level 1 of this filesystem, but the filesystem's
data size is:
13526260 /data2/homes
So obviously Amanda isn't doing a level 1 here even though she thinks
she is.
The problem here is that every night my backups fail because there's
always too much data to put on the backup medium, and Amanda can never
seem to get a level 1. Note that this particular filesystem has been
backed up seven times, the first time was a level 0 and each time
after was a "level 1" but Amanda dumped the full 13gb partition each
time.
Help? Any ideas why Amanda can't seem to get level 1 calculations
correct?
Thanks!
\marc
--
mnc AT pobox DOT com
Marc N. Cannava
www.botcave.com
--
Kevin Dalley
kevin AT kelphead DOT org
--
mnc AT pobox DOT com
Marc N. Cannava
www.botcave.com
|
