I'm having this issue with bacula 5.2.6 on Debian Wheezy, after a failed run of 'estimate' command on bconsole for whatever reason (e.g. bad password), it will fail for all subsequently runs on jobs that has previously worked, complaining about a 'connection reseted by peer' for the failed host.
Also, I'm getting an email message only for the first failed run, when I exit bconsole, I get another email message with all failed runs, but always refering to the original failing job. Then, when I run bconsole again, everything works again.
root@bacula2mt:~# bconsole
1000 OK: bacula2mt-dir Version: 5.2.6 (21 February 2012)
Enter a period to cancel a command.
*estimate
The defined Job resources are:
1: NightlySave mailserver
2: NightlySave webserver
Select Job resource (1-2): 1
Using Catalog "MyCatalog"
2000 OK estimate files=27,985 bytes=1,749,993,667
*estimate
The defined Job resources are:
1: NightlySave mailserver
2: NightlySave webserver
Select Job resource (1-2): 2
Using Catalog "MyCatalog"
Failed to connect to Client.
You have messages.
*mess
12-Sep 20:37 bacula2mt-dir JobId 0: Fatal error: Unable to authenticate with File daemon at "
webserver.example.com:9102". Possible causes:
Passwords or names not the same or
Maximum Concurrent Jobs exceeded on the FD or
FD networking messed up (restart daemon).
*estimate
The defined Job resources are:
1: NightlySave mailserver
2: NightlySave webserver
Select Job resource (1-2): 1
Using Catalog "MyCatalog"
Failed to connect to Client.
You have messages.
*mess
12-Sep 20:37 bacula2mt-dir JobId 0: Error: bsock.c:389 Write error sending 37 bytes to Client: webserver-fd:
webserver.example.com:9102: ERR=Connection reset by peer
12-Sep 20:37 bacula2mt-dir JobId 0: Fatal error: Error sending Hello to File daemon at "
webserver.example.com:9102". ERR=Connection reset by peer