Hi! For the avoidance of doubt: The bug I reported here and all the testing I did in 2019 around this bug report all took place on my Bacula production system.
In the meantime, I did set up an independent test system in order to have more flexibility in testing and also to exclude hardware problems other than those I already tested. Here's a quick overview of the test system compared to the production system: - System main components (e.g. mainboard, CPU, ...): all different from production system. - Fibre Channel controler: different (production system: Emulex LPe11002; test system: QLogic QLE2460) - Tape Library: same type as production system - LTO5 drive: exact same drive as previously tested in production system - Basic operating system: same as production system (Debian Stretch) - Actual kernel running: different (production system: 4.9.0-12-amd64; test system: 4.9.0-14-amd64) - Bacula version: same as production system: - bacula: 7.4.4+dfsg-6+deb9u2 - bacula-bscan: 9.4.2-1~bpo9+1 - bacula-client: 7.4.4+dfsg-6+deb9u2 - bacula-common: 9.4.2-1~bpo9+1 - bacula-common-mysql: 9.4.2-1~bpo9+1 - bacula-console: 9.4.2-1~bpo9+1 - bacula-director: 9.4.2-1~bpo9+1 - bacula-director-mysql: 9.4.2-1~bpo9+1 - bacula-fd: 9.4.2-1~bpo9+1 - bacula-sd: 9.4.2-1~bpo9+1 - bacula-sd-dbgsym: 9.4.2-1~bpo9+1 - bacula-server: 7.4.4+dfsg-6+deb9u2 Now, when I run tests on the test system with btape, both the "test" test and the "fill" test work without any problems, i.e. no crash occurs. Regards Sebastian