How To Install Bacula Server on Ubuntu 14.04

A Bacula job is used to perform backup and restore actions. Job resources define the details of what a particular job will do, including the name of the Client, the FileSet to back up or restore, among other things.

Here, we will configure the jobs that will be used to perform backups of the local filesystem.

In the Director configuration, find the Job resource with a name of “BackupClient1” (search for “BackupClient1”). Change the value of Name to “BackupLocalFiles”, so it looks like this:

Job {
  Name = "BackupLocalFiles"
  JobDefs = "DefaultJob"
}

Next, find the Job resource that is named “RestoreFiles” (search for “RestoreFiles”). In this job, you want to change two things: update the value of Name to “RestoreLocalFiles”, and the value of Where to “/bacula/restore”. It should look like this:

Job {
  Name = "RestoreLocalFiles"
  Type = Restore
  Client=BackupServer-fd
  FileSet="Full Set"
  Storage = File
  Pool = Default
  Messages = Standard
  Where = /bacula/restore
}

This configures the RestoreLocalFiles job to restore files to /bacula/restore, the directory we created earlier.

A Bacula FileSet defines a set of files or directories to include or exclude files from a backup selection, and are used by jobs.

Find the FileSet resource named “Full Set” (it’s under a comment that says, “# List of files to be backed up”). Here we will make three changes: (1) Add the option to use gzip to compress our backups, (2) change the include File from /usr/sbin to /, and (3) change the second exclude File to /bacula. With the comments removed, it should look like this:

FileSet {
  Name = "Full Set"
  Include {
    Options {
      signature = MD5
      compression = GZIP
    }    
File = /
}
  Exclude {
    File = /var/lib/bacula
    File = /bacula
    File = /proc
    File = /tmp
    File = /.journal
    File = /.fsck
  }
}

Let’s go over the changes that we made to the “Full Set” FileSet. First, we enabled gzip compression when creating a backup archive. Second, we are including /, i.e. the root partition, to be backed up. Third, we are excluding /bacula because we don’t want to redundantly back up our Bacula backups and restored files.

Note: If you have partitions that are mounted within /, and you want to include those in the FileSet, you will need to include additional File records for each of them.

Keep in mind that if you always use broad FileSets, like “Full Set”, in your backup jobs, your backups will require more disk space than if your backup selections are more specific. For example, a FileSet that only includes your customized configuration files and databases might be sufficient for your needs, if you have a clear recovery plan that details installing required software packages and placing the restored files in the proper locations, while only using a fraction of the disk space for backup archives.

In the Bacula Director configuration file, the Storage resource defines the Storage Daemon that the Director should connect to. We’ll configure the actual Storage Daemon in just a moment.

Find the Storage resource, and replace the value of Address, localhost, with the private FQDN (or private IP address) of your backup server. It should look like this (substitute the highlighted word):

Storage {
  Name = File
# Do not use "localhost" here
  Address = backup_server_private_FQDN                # N.B. Use a fully qualified name here
  SDPort = 9103
  Password = "ITXAsuVLi1LZaSfihQ6Q6yUCYMUssdmu_"
  Device = FileStorage
  Media Type = File
}

This is necessary because we are going to configure the Storage Daemon to listen on the private network interface, so remote clients can connect to it.

A Pool resource defines the set of storage used by Bacula to write backups. We will use files as our storage volumes, and we will simply update the label so our local backups get labeled properly.

Find the Pool resource named “File” (it’s under a comment that says “# File Pool definition”), and add a line that specifies a Label Format. It should look like this when you’re done:

# File Pool definition
Pool {
  Name = File
  Pool Type = Backup
  Label Format = Local-
  Recycle = yes                       # Bacula can automatically recycle Volumes
  AutoPrune = yes                     # Prune expired volumes
  Volume Retention = 365 days         # one year
  Maximum Volume Bytes = 50G          # Limit Volume size to something reasonable
  Maximum Volumes = 100               # Limit number of Volumes in Pool
}

Save and exit. You’re finally done configuring the Bacula Director.

Let’s verify that there are no syntax errors in your Director configuration file:

sudo bacula-dir -tc /etc/bacula/bacula-dir.conf

If there are no error messages, your bacula-dir.conf file has no syntax errors.

Next, we’ll configure the Storage Daemon.

Find the Storage resource. This defines where the SD process will listen for connections. Add the SDAddress parameter, and assign it to the private FQDN (or private IP address) of your backup server:

Storage {                             # definition of myself
  Name = BackupServer-sd
  SDPort = 9103                  # Director's port
  WorkingDirectory = "/var/lib/bacula"
  Pid Directory = "/var/run/bacula"
  Maximum Concurrent Jobs = 20
  SDAddress = backup_server_private_FQDN
}

Next, find the Device resource named “FileStorage” (search for “FileStorage”), and update the value of Archive Device to match your backups directory:

Device {
  Name = FileStorage
  Media Type = File
  Archive Device = /bacula/backup 
  LabelMedia = yes;                   # lets Bacula label unlabeled media
  Random Access = Yes;
  AutomaticMount = yes;               # when device opened, read it
  RemovableMedia = no;
  AlwaysOpen = no;
}

Save and exit.

Let’s verify that there are no syntax errors in your Storage Daemon configuration file:

sudo bacula-sd -tc /etc/bacula/bacula-sd.conf

If there are no error messages, your bacula-sd.conf file has no syntax errors.

We’ve completed the Bacula configuration. We’re ready to restart the Bacula server components.

Begin by issuing a label command:

label

You will be prompted to enter a volume name. Enter any name that you want:

Enter new Volume name:
MyVolume

Then select the pool that the backup should use. We’ll use the “File” pool that we configured earlier, by entering “2”:

Select the Pool (1-3):
2

Bacula now knows how we want to write the data for our backup. We can now run our backup to test that it works correctly:

run

You will be prompted to select which job to run. We want to run the “BackupLocalFiles” job, so enter “1” at the prompt:

Select Job resource (1-3):
1

At the “Run Backup job” confirmation prompt, review the details, then enter “yes” to run the job:

yes

After running a job, Bacula will tell you that you have messages. The messages are output generated by running jobs.

Check the messages by typing:

messages

The messages should say “No prior Full backup Job record found”, and that the backup job started. If there are any errors, something is wrong, and they should give you a hint as to why the job did not run.

Another way to see the status of the job is to check the status of the Director. To do this, enter this command at the bconsole prompt:

status director

If everything is working properly, you should see that your job is running. Something like this:

Output — status director (Running Jobs)
Running Jobs:
Console connected at 09-Apr-15 12:16
 JobId Level   Name                       Status
======================================================================
     3 Full    BackupLocalFiles.2015-04-09_12.31.41_06 is running
====

When your job completes, it will move to the “Terminated Jobs” section of the status report, like this:

Output — status director (Terminated Jobs)
Terminated Jobs:
 JobId  Level    Files      Bytes   Status   Finished        Name
====================================================================
     3  Full    161,124    877.5 M  OK       09-Apr-15 12:34 BackupLocalFiles

The “OK” status indicates that the backup job ran without any problems. Congratulations! You have a backup of the “Full Set” of your Bacula server.

The next step is to test the restore job.

To demonstrate, we’ll restore all of the files in our last backup:

restore all

A selection menu will appear with many different options, which are used to identify which backup set to restore from. Since we only have a single backup, let’s “Select the most recent backup”—select option 5:

Select item (1-13):
5

Because there is only one client, the Bacula server, it will automatically be selected.

The next prompt will ask which FileSet you want to use. Select “Full Set”, which should be 2:

Select FileSet resource (1-2):
2

This will drop you into a virtual file tree with the entire directory structure that you backed up. This shell-like interface allows for simple commands to mark and unmark files to be restored.

Because we specified that we wanted to “restore all”, every backed up file is already marked for restoration. Marked files are denoted by a leading * character.

If you would like to fine-tune your selection, you can navigate and list files with the “ls” and “cd” commands, mark files for restoration with “mark”, and unmark files with “unmark”. A full list of commands is available by typing “help” into the console.

When you are finished making your restore selection, proceed by typing:

done

Confirm that you would like to run the restore job:

OK to run? (yes/mod/no):
yes

As with backup jobs, you should check the messages and Director status after running a restore job.

Check the messages by typing:

messages

There should be a message that says the restore job has started or was terminated with an “Restore OK” status. If there are any errors, something is wrong, and they should give you a hint as to why the job did not run.

Again, checking the Director status is a great way to see the state of a restore job:

status director

When you are finished with the restore, type exit to leave the Bacula Console:

exit

To verify that the restore job actually restored the selected files, you can look in the /bacula/restore directory (which was defined in the “RestoreLocalFiles” job in the Director configuration):

sudo ls -la /bacula/restore

You should see restored copies of the files in your root file system, excluding the files and directories that were listed in the “Exclude” section of the “RestoreLocalFiles” job. If you were trying to recover from data loss, you could copy the restored files to their appropriate locations.

You may want to delete the restored files to free up disk space. To do so, use this command:

sudo -u root bash -c "rm -rf /bacula/restore/*"

Note that you have to run this rm command as root, as many of the restored files are owned by root.