Troubleshooting Bacula

From Tech Wiki
Jump to navigation Jump to search

Initial Steps

1. Firstly check that Bacula is installed and running.

The version on Windows needs to be 5.0.3 (or potentially greater). If it's 3.0.3a then you'll need to re-install with the correct version.

2. Check server is on IPv6.

A server needs a spare NIC to move onto an IPv6 backup network. If all NICs are in use then you'll need to add a USB Nic.

FYI: USB NICs are incompatible with VMWare Hardware Nodes!!

You should then be able to follow this KB:

Bacula Linux IPv6 Setup

3. Check the client can communicate to the backup server on 9103

   telnet backupserverIPv6address 9103

Connection refused - Outbound 9103 on the client is blocked. Check ip6tables and Windows Firewall.

Check here for ip6tables: Bacula Linux IP6Tables Rules

A lot of CPanel servers are now running the ConfigServer Firewall which will automatically start ip6tables. You can check using 'ps aufx | grep lfd'. If they are, give them the IPv6 address of the backup server and ask them to open 9102 inbound and 9103 outbound to/from this address.

4. Check the backup server can connect to the client on 9102.

telnet clientserverIPv6address 9102

Connection refused - 9102 is blocked inbound. Check software firewalls as above.

5. Check a status client

If the server connects over IPv4 check all of the above, and rebuild_backups. You should see it then connects over IPv6.

Connecting to Client srv-94_229_167_17 at 2a02:22d0:bac:0:20e:cff:fe31:8801:9102

Common backup errors

1. Network connection error

This one is obvious, but is often caused by reboots. Check the uptime on the server.

Ensure you're able to run status client successfully. Run a test backup if necessary to ensure data is able to transfer.

2. Bad Response to Open Command

This can be caused by the hostname in the bacula-fd.conf file not being set to a FQDN. Ensure this is set to the hostname of the server. 3. Input/Output Error

I've only ever seen this as a result of a busy disk on the client. In this case it's worth seeing if we can check disk I/O at the time backups run, and to see if there's a more suitable time the backup can be run at.

This can also be caused by a corrupt volume on the backup server which can be resolved by deleting the error-ed volumes from within bconsole.

4. bconsole can't connect to the director

Firstly check if the disk is full (if it is, move MySQL to /home and run the reclaim backups script). It may also be caused by /boot being full as well.

To help clear out the /boot partition you can clear out the old kernels. To do this take a look at the following page:

Clear Out Old Kernels

To move MySQL to the /home partition you can take a look at the following page:

Move MySQL To Home Partition

BSMTP Tickets

You can test bsmtp by calling it directly with the debug flag

/usr/sbin/bsmtp -h -f [email protected] -s "Bacula: %t %e of %c %l" %r -d 50

This should output any errors or show if it's sending out correctly.