2011/2/25 Igor Zinovik <zinovik.igor AT gmail DOT com>:
>  Hello.
>
> I'm using bacula 5.0.3 since november 2010.  Now i'm facing
> problem with my backup jobs.  We have a mail server that
> also runs bacula-fd and i have to backup it, but this
> computer has very large file system partitions.  Bacula have to backup
> about 310 gigabytes of data on full backup.  Right now i'm facing problem
> that full backup does not fit my Max Run Time which is 42
> hours (17280000 seconds).  On full backup mail server goes
> first and when doing full backup bacula started to terminate
> mail server full backup jobs and all other server do not
> have their full jobs started.  Of course i can lower its priority
> and crank Max Run Time to something like 5 days,
> but i think it is not right solution for my problem.
>
> And here is my question: how I can parallelize backup jobs execution?
>
> I have following setup: 3 pools for full, differential and
> incremental jobs.  Maximum concurrent jobs are set to 20 in
> bacula-{dir,sd}.conf, but when scheduled jobs are executed
> they run consequentially, one starts, then it finishes and
> another one is launched for exection.  All pool backup
> data is stored on NFS share which is the signle device that
> is connected to bacula-sd.  Why my jobs are not concurrent?
>
> According documentation (26.2 Concurrent Disk Jobs) jobs
> could run simultaneously, but in fact it is not.  I want bacula
> to write simultaneously into one volume, not in separate
> volumes.  Soon I will add from 5 ti 10 heavy loaded servers  with
> big filesystems to my current 9 client and I would like to
> backup them concurrently.
>
> bacula-dir.conf:
> ...
> director {
>  name = bacula-dir
>  dirport = 9101
>  queryfile = "/usr/lib/bacula/query.sql"
>  working directory = "/var/lib/bacula"
>  pid directory = "/var/run"
>  maximum concurrent jobs = 20
>  password = "XXXXXX"
>  messages = Daemon
> }
>
>
> jobdefs {
>  name = "job defaults"
>  type = backup
>  level = incremental
>  fileset = "Full Set"
>  schedule = "monthly cycle"
>  storage = disk-storage
>  pool = "full pool"
>  full backup pool = "full pool"
>  differential backup pool = "diff pool"
>  incremental backup pool = "incr pool"
>  priority = 10
>  write bootstrap = "/var/lib/bacula/%c.bsr"
>  messages = Standard
>  # 48 hours
>  max run time = 1728000
> }
>
> schedule {
>  name = "monthly cycle"
>  # Full backup on each even month
>  run = full February,April,June,August,October,December 1st sun at 22:05
>  run = differential 2nd-5th sun at 22:05
>  run = incremental mon-sat at 22:05
> }
>
> storage {
>  name = disk-storage
>  address = 172.20.20.29
>  sd port = 9103
>  password = "ХХХ"
>  device = backup-disk-device
>  media type = File
> }
>
You need to set maximum concurrent jobs in the storage resource above
as well as where you have set it.

Also if you wanted a client to run concurrent jobs you need to enable
that in the client resource for each client in bacula-dir.conf by
setting the maximum concurrent jobs in the client resource.

> bacula-sd.conf:
> ...
> storage {
>  name = bacula-sd
>  sdport = 9103
>  working directory = "/var/lib/bacula"
>  pid directory = "/var/run"
>  maximum concurrent jobs = 20
>  heartbeat interval = 30 sec
> }
>
> device {
>  name = backup-disk-device
>  media type = File
>  archive device = /var/backup
>  label media = yes                   # lets Bacula label unlabeled media
>  always open = yes
>  automatic mount = yes               # when device opened, read it
>  random access = yes
>  removable media = no
> }
>

One single storage device that will restrict bacula to run only jobs
from 1 pool at a time since you can not load two different volumes in
the same device (even disk devices0 at the same time.

John

------------------------------------------------------------------------------
Free Software Download: Index, Search & Analyze Logs and other IT data in 
Real-Time with Splunk. Collect, index and harness all the fast moving IT data 
generated by your applications, servers and devices whether physical, virtual
or in the cloud. Deliver compliance at lower cost and gain new business 
insights. http://p.sf.net/sfu/splunk-dev2dev 
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users