Locations that don't exist on first run are never tried again (Trac #48)

[boxbackup-bot]

Torsten reports:

If i have a location not existing locally and not existing on the store but activated as a location in bbackupd.conf. With this configuration box backup throws this errors on start.

ERROR: Failed to stat file: '/test-192.168.1.87-torsten': No such file or directory (2)
WARNING: Exception thrown: CommonException(OSFileError) at BackupClientFileAttributes.cpp(309)
ERROR: Failed to get attributes for path '/test-192.168.1.87-torsten', skipping location 'test-192.168.1.87-torsten'
WARNING: Failed to access directory: /test-192.168.1.87-torsten: No such file or directory
NOTICE: About to notify administrator about event backup-ok, running
script '/root/exobackup/boxbackup/NotifySysadmin.sh backup-ok'

• this errors are only printed once. The next backup runs do not consider to backup this location. That is not good for me. I have locations on remote storage mounted via samba. So only one mount error and i will never again be notified that this location is not backed up.

• is this the right behavior, that backup-ok is called even though there was this error?

I attached a little patch that makes box backup to retry this location. But i dont think that this is a clean way.

Migrated from https://www.boxbackup.org/ticket/48

{
    "status": "new", 
    "changetime": "2008-11-08T22:18:55", 
    "description": "Torsten reports:\n\nIf i have a location not existing locally and not existing on the store but activated as a location in bbackupd.conf. With this configuration box backup throws this errors on start.\n\n{{{\nERROR: Failed to stat file: '/test-192.168.1.87-torsten': No such file or directory (2)\nWARNING: Exception thrown: CommonException(OSFileError) at BackupClientFileAttributes.cpp(309)\nERROR: Failed to get attributes for path '/test-192.168.1.87-torsten', skipping location 'test-192.168.1.87-torsten'\nWARNING: Failed to access directory: /test-192.168.1.87-torsten: No such file or directory\nNOTICE: About to notify administrator about event backup-ok, running\nscript '/root/exobackup/boxbackup/NotifySysadmin.sh backup-ok'\n}}}\n\n * this errors are only printed once. The next backup runs do not consider to backup this location. That is not good for me. I have locations on remote storage mounted via samba. So only one mount error and i will never again be notified that this location is not backed up.\n\n * is this the right behavior, that backup-ok is called even though there was this error?\n\nI attached a little patch that makes box backup to retry this location. But i dont think that this is a clean way.", 
    "reporter": "chris", 
    "cc": "ddmails@web.de", 
    "resolution": "", 
    "_ts": "1226182735000000", 
    "component": "bbackupd", 
    "summary": "Locations that don't exist on first run are never tried again", 
    "priority": "normal", 
    "keywords": "backup location not available retry backup-ok notification", 
    "version": "0.11rc1", 
    "time": "2008-11-08T22:16:06", 
    "milestone": "0.12", 
    "owner": "chris", 
    "type": "defect"
}

[boxbackup-bot]

Trac update at 20081108T22:16:22: chris changed attachment from "" to "bb_retry_dir_not_existing.diff"

[boxbackup-bot]

Trac update at 20081108T22:18:55: chris changed description from:

Torsten reports:

''if i have a location not existing locally and not existing on the store but
activated as a location in bbackupd.conf. With this configuration box backup
throws this errors on start.''

{{{
ERROR: Failed to stat file: '/test-192.168.1.87-torsten': No such file or directory (2)
WARNING: Exception thrown: CommonException(OSFileError) at BackupClientFileAttributes.cpp(309)
ERROR: Failed to get attributes for path '/test-192.168.1.87-torsten', skipping location 'test-192.168.1.87-torsten'
WARNING: Failed to access directory: /test-192.168.1.87-torsten: No such file or directory
NOTICE: About to notify administrator about event backup-ok, running
script '/root/exobackup/boxbackup/NotifySysadmin.sh backup-ok'
}}}

• ''this errors are only printed once. The next backup runs do not consider to
  backup this location. That is not good for me. I have locations on remote
  storage mounted via samba. So only one mount error and i will never again be
  notified that this location is not backed up.''

• ''is this the right behavior, that backup-ok is called even though there was
  this error?''

''I attached a little patch that makes box backup to retry this location. But i
dont think that this is a clean way.''

to:

Torsten reports:

If i have a location not existing locally and not existing on the store but activated as a location in bbackupd.conf. With this configuration box backup throws this errors on start.

{{{
ERROR: Failed to stat file: '/test-192.168.1.87-torsten': No such file or directory (2)
WARNING: Exception thrown: CommonException(OSFileError) at BackupClientFileAttributes.cpp(309)
ERROR: Failed to get attributes for path '/test-192.168.1.87-torsten', skipping location 'test-192.168.1.87-torsten'
WARNING: Failed to access directory: /test-192.168.1.87-torsten: No such file or directory
NOTICE: About to notify administrator about event backup-ok, running
script '/root/exobackup/boxbackup/NotifySysadmin.sh backup-ok'
}}}

• this errors are only printed once. The next backup runs do not consider to backup this location. That is not good for me. I have locations on remote storage mounted via samba. So only one mount error and i will never again be notified that this location is not backed up.

• is this the right behavior, that backup-ok is called even though there was this error?

I attached a little patch that makes box backup to retry this location. But i dont think that this is a clean way.

[boxbackup-bot]

bb_retry_dir_not_existing.diff