amanda inparallel not working on large filesystems
2003-06-25 06:07:10
greetings!
I have a backup strategy that happily performs a full backup for 160Gb
of data every night, however, I noticed that on three large filesystems
on a client machine, the backups work sequentially. The tape device is
an HP Surestore 2/20 autoloader, capable of 200Gb per tape.
I am sure that it used to dump these
backups in parallel, but now they are definitely waiting. amstatus throughout
the night shows all backups kicking off within 1 hour of each other, yet
each backup waiting until the preceeding one completes, before commencing.
I have maxdumps set to 5 and inparallel set to 10 (also tried it at 30
and 63), yet these three systems only backup sequentially.
I ran amstatus every five minutes last night to confirm this, and it shows
on filesystem dumping while the other is listed as dumping, but sticks
at 32K until the other one finishes.
anybody any clues on to why amanda should behave this way?
Each file system is about 30Gb in size and I have four holding discs:
50Gb, 36Gb, 30Gb, 5Gb each.
Also, it tries to fill up the 5Gb one before it even attempts the others.
Here is my amanda.conf file, hope this
helps.
dumpuser "amanda"
# the user to run dumps under
printer "pcm2300f2"
inparallel 10 # maximum dumpers
that will run in parallel (max 63)
netusage 100000 Kbps #
maximum net bandwidth for Amanda, in KB per sec
dumpcycle 0 # the number
of days in the normal dump cycle
runspercycle 1
# the number of amdump runs in dumpcycle days
tapecycle 22 tapes #
the number of tapes in rotation
bumpsize 20 Mb
# minimum savings (threshold) to bump level 1 -> 2
bumpdays 1
# minimum days at each level
bumpmult 4
# threshold = bumpsize * bumpmult^(level-1)
ctimeout 1800
# number of seconds per filesystem timeouts on estimates
etimeout 1800
# number of seconds per filesystem for estimates.
dtimeout 24000
# number of idle seconds before a dump is aborted.
runtapes 1
# number of tapes to be used in a single run of amdump
tpchanger "/opt/etc/amanda/cc/chg-zd-mtx"
# the tape-changer glue script
changerfile "/opt/etc/amanda/cc/chg-zd-mtx"
# the tape-changer glue script
changerdev "/dev/changer"
# HP LTO on this box
tapedev "/dev/rmt/0un" #
the no-rewind tape device to be used
tapetype LTO
# what kind of tape it is (see tapetypes below)
labelstr "^cc[0-9][0-9][0-9][0-9]*$"
# label constraint regex: all tapes must
match
holdingdisk hd1 {
comment "main holding
disk"
directory "/amanda/dump1"
use 0
chunksize 500Mb
}
holdingdisk hd2 {
comment "JOW added
(local) secondary holding disk"
directory "/amanda/dump2"
use 0
chunksize 500Mb
}
holdingdisk hd3 {
comment "JOW added
secondary holding disk"
directory "/amanda/dump3"
use 0
chunksize 500Mb
}
holdingdisk hd4 {
comment "JOW added
secondary holding disk"
directory "/amanda/dump4"
use -500Mb
chunksize 500Mb
}
holdingdisk hd5 {
comment "JOW added
secondary holding disk"
directory "/amanda/dump6"
use -105Gb
chunksize 500Mb
}
maxdumps 5
reserve 30 # percent
infofile "/opt/etc/amanda/cc/curinfo"
# database DIRECTORY
logdir "/opt/etc/amanda/cc/log"
# log directory
indexdir "/opt/etc/amanda/cc/index"
# index directory
define tapetype LTO {
comment
"LTO tape drives"
length 200000
mbytes
lbl-templ
"/opt/etc/amanda/DIN-A4.ps"
filemark
513 kbytes
speed 20000
kps
}
define dumptype global {
comment "Global definitions"
index no
record no
compress none
dumpcycle 0
maxdumps 5
}
define dumptype lo_no_tar_full {
global
comment "Full dump of this filesystem always"
program "GNUTAR"
priority low
}
define dumptype me_no_tar_full {
global
comment "Full dump of this filesystem always"
program "GNUTAR"
priority medium
}
define dumptype hi_no_tar_full {
global
comment "Full dump of this filesystem always"
program "GNUTAR"
priority high
}
|
<Prev in Thread] |
Current Thread |
[Next in Thread> |
- amanda inparallel not working on large filesystems,
jason . walton <=
|
|
|