I'm assuming that setting "Heartbeat Interval = 300" is the same as 5 minutes. I want to be clear on it's usage though.. it would be for when a Firewall or Router is not honoring KeepAlive?
Hello, Mingus: 300 == 5 minutes. It should be added to all involved self-defining daemon resources (Director - bacula-dir.conf, Storage - bacula-sd.conf and FileDaemon - bacula-fd.conf). You might try lower values, but it is not a warranty that will solve this problem.
I'm having issues with backup failures on a mix of Windows clients. These all seem to have bsock.c timeout issues. Sometimes the jobs complete, mostly the FD completes writing files to SD, but loses connection to the Director and doesn't write the job to the database. It's become a hot button issue for me that I can't seem to resolve.
Do you have any antivirus and/or IPS solution? I had this customer once that I had to disable his antivirus solution with scripts during backup, because of this same behavior.
Besides that I think you should look for regular physical network troubleshooting (packet loss, router and firewall logs etc.).
I've added the Heartbeat Interval to the Director and FD configs. I've verified that the security team is not seeing any dropped connections in their FW logs, and the timeout is 4h on connections. I've curated the FileSet to minimize the data transferred. The systems are all VMs and some version of Windows (Windows 2k8, 2k12). Other than that I haven't been able to find anything common.
I'd appreciate any ideas. The server is Bacula 7.0.5 and clients are all 5.x
It would be nice to upgrade everything to 7.4, if possible.
