Red Hat Directory server backup scheduled db2bak.pl keep starting over as long as fs is full

Issue

• For backup of the instance we use 'db2bak.pl'

  /usr/lib64/dirsrv/slapd-<instance>/db2bak.pl -D "cn=Directory Manager" -w <pwd>
  

• This works fine till the filesystem is 100% full.

• In the error log then we see the following entry:

  [06/Jul/2016:14:16:36 +0200] - Backing up file 107170 (/var/lib/dirsrv/slapd-<instance>/bak/<instance>-2016_7_6_14_4_23/userRoot/entryusn.db)
  [06/Jul/2016:14:16:36 +0200] - Backing up file 107171 (/var/lib/dirsrv/slapd-<instance>/bak/<instance>-2016_7_6_14_4_23/userRoot/sn.db)
  [06/Jul/2016:14:16:36 +0200] - Backing up file 107172 (/var/lib/dirsrv/slapd-<instance>/bak/<instance>-2016_7_6_14_4_23/log.0000000001)
  [06/Jul/2016:14:16:36 +0200] - dblayer_copyfile: failed to write: 0
  [06/Jul/2016:14:16:36 +0200] - Backup: error in copying file '/var/lib/dirsrv/slapd-<instance>/db/log.0000000001' (err=-1) -- Starting over...
  [06/Jul/2016:14:16:36 +0200] - Backing up file 107173 (/var/lib/dirsrv/slapd-<instance>/bak/<instance>-2016_7_6_14_4_23/userRoot/uid.db)
  [06/Jul/2016:14:16:36 +0200] - Backing up file 107174 (/var/lib/dirsrv/slapd-<instance>/bak/<instance>-2016_7_6_14_4_23/userRoot/aci.db)
  

• This is repeated until some free space in the filesystem is created.

• Without creating some free space, the error log is filled continuously with above messages.

• Important to know is that we have the backup filesystem separated from the db filesystem:

  /dev/mapper/vg.appl-lv.varlibdirsrvbak    25G   23G  648M  98% /var/lib/dirsrv/slapd-<instance>/bak