encore une fois il y avait tout dans la doc et plus précisement la doc du director
commençons par faire un pool de bande par niveau de backup (Full / Differentielle / Incrementale / Catalogue )

Pool {
  Name = Full
  Pool Type = Backup
  Recycle = yes                 # Recycle les volumes automatiquement
  AutoPrune = yes               # Prune (élague) les volumes expirés
  Volume Retention = 1 week     # 1 Semaine de rétention
  Accept Any Volume = yes       # accepte tout les volumes
  Maximum Volume Jobs = 5       # 5 Jobs Maxi / Volumes (passe en Used auto)
  Label Format = "File-Full-"   # Auto Label File-Full-0001, File-Full-0002, ...
}

Pool {
  Name = Diff
  Pool Type = Backup
  Recycle = yes
  AutoPrune = yes
  Volume Retention = 1 week
  Accept Any Volume = yes
  Maximum Volume Jobs = 5
  Label Format = "File-Diff-"
}

Pool {
  Name = Inc
  Pool Type = Backup
  Recycle = yes
  AutoPrune = yes
  Volume Retention = 1 week
  Accept Any Volume = yes
  Maximum Volume Jobs = 35
  Label Format = "File-Inc-"
}

Pool {
  Name = Catalog
  Pool Type = Backup
  Recycle = yes
  AutoPrune = yes
  Volume Retention = 1 week
  Accept Any Volume = yes
  Maximum Volume Jobs = 7
  Label Format = "File-Catalog-"
}

une fois celle effectuer il faut indiquer quel pool utiliser pour quel level, dans le JobDefs on ajoute

  Full Backup Pool = Full
  Differential Backup Pool = Diff
  Incremental Backup Pool = Inc

ce qui donne chez moi

JobDefs {
  Name = "DefaultJob"
  Type = Backup
  Level = Incremental
  Client = tarhu-fd
  FileSet = "Full Set"
  Schedule = "WeeklyCycle"
  Storage = File
  Messages = Standard
  Pool = Default
  Full Backup Pool = Full
  Differential Backup Pool = Diff
  Incremental Backup Pool = Inc
  Priority = 10
}

puis on change le pool pour le catalog ( Full Backup Pool = Catalog ) 

Job {
  Name = "backupCatalog"
  JobDefs = "DefaultJob"
  Level = Full
  FileSet="Catalog"
  Full Backup Pool = Catalog
  Schedule = "WeeklyCycleAfterBackup"
  # Dump du catalogue
  RunBeforeJob = "/etc/bacula/scripts/make_catalog_backup bacula -ubacula -pbacula"
  # depose sur dedibakup & supprime le bacula.sql.gz
  RunAfterJob  = "/usr/local/bin/storeOnDedibackup"
  Write Bootstrap = "/var/lib/bacula/BackupCatalog.bsr"
  Priority = 11                   # run after main backup
}

un gros reload/restart du director et ca roule tout seul

le contenu du script storeOnDedibackup

#!/bin/bash
#
# 2007-08-27 Creation Pierre (http://www.grumpf.net)
#

for xfile in $(find /backup/ -name File-* -type f 2>>/dev/null)
do
        if [ -s $xfile ] ; then
          ncftpput -V -u auto -p " " dedibackup.dedibox.fr / $xfile
        fi
done

[ -f /var/lib/bacula/bacula.sql.gz ] && rm /var/lib/bacula/bacula.sql.gz