Description of problem: The systemd service file for bacula-fd contains "ExecStartPre=/usr/sbin/bacula-checkconf $CONFIG" but bacula-fd does not contain a Requires on bacula-dir, hence the start of the bacula-fd service fails: [root@localhost bacula]# systemctl status bacula-fd.service â bacula-fd.service - Bacula-FileDaemon, a Backup-client Loaded: loaded (/usr/lib/systemd/system/bacula-fd.service; enabled) Active: failed (Result: start-limit) since Fri 2015-05-22 15:15:47 BST; 7s ago Docs: man:bacula-fd(8) Process: 16399 ExecStartPre=/usr/sbin/bacula-checkconf $CONFIG (code=exited, status=203/EXEC) May 22 15:15:47 localhost systemd[1]: bacula-fd.service: control process exited, code=exited status=203 May 22 15:15:47 localhost systemd[1]: Failed to start Bacula-FileDaemon, a Backup-client. May 22 15:15:47 localhost systemd[1]: Unit bacula-fd.service entered failed state. May 22 15:15:47 localhost systemd[1]: bacula-fd.service failed. May 22 15:15:47 localhost systemd[1]: start request repeated too quickly for bacula-fd.service May 22 15:15:47 localhost systemd[1]: Failed to start Bacula-FileDaemon, a Backup-client. May 22 15:15:47 localhost systemd[1]: Unit bacula-fd.service entered failed state. May 22 15:15:47 localhost systemd[1]: bacula-fd.service failed. [root@localhost bacula]# /usr/sbin/bacula-checkconf /etc/bacula/bacula-fd.conf -bash: /usr/sbin/bacula-checkconf: No such file or directory Maybe other bacula packages also need a dependency on bacula-dir for /usr/sbin/bacula-checkconf, please check and add as necessary. Reported via forums: https://forums.mageia.org/en/viewtopic.php?f=15&t=9821 Reproducible: Steps to Reproduce:
Whiteboard: (none) => MGA5TOO
CC: (none) => thomas
Status: NEW => ASSIGNEDAssignee: shlomif => thomas
This bug is now fixed in svn. As soon as I can submit it to updates_testing I will assign it to QA
For the meantime, added to errata: https://wiki.mageia.org/en/Mageia_5_Errata#bacula_misses_a_dependency_on_.2Fusr.2Fsbin.2Fbacula-checkconf
CC: (none) => doktor5000Whiteboard: MGA5TOO => MGA5TOO, IN_ERRATA
This bug has now been fixed. The following packages are now in mga5 updates_testing: bacula-7.0.5-14.1.mga4.src.rpm lib64bacula7-7.0.5-14.1.mga4.x86_64.rpm lib64bacula-sql7-7.0.5-14.1.mga4.x86_64.rpm bacula-common-7.0.5-14.1.mga4.x86_64.rpm bacula-dir-7.0.5-14.1.mga4.x86_64.rpm bacula-sd-7.0.5-14.1.mga4.x86_64.rpm bacula-fd-7.0.5-14.1.mga4.x86_64.rpm bacula-console-7.0.5-14.1.mga4.x86_64.rpm bacula-bat-7.0.5-14.1.mga4.x86_64.rpm bacula-tray-monitor-7.0.5-14.1.mga4.x86_64.rpm lib64bacula-devel-7.0.5-14.1.mga4.x86_64.rpm nagios-plugins-bacula-7.0.5-14.1.mga4.x86_64.rpm bacula-debuginfo-7.0.5-14.1.mga4.x86_64.rpm and i586 packages Assigning it to QA
Version: Cauldron => 5Assignee: thomas => qa-bugsWhiteboard: MGA5TOO, IN_ERRATA => IN_ERRATA, CAULDRON TOO
Please fix in cauldron first.
Version: 5 => CauldronAssignee: qa-bugs => thomasWhiteboard: IN_ERRATA, CAULDRON TOO => IN_ERRATA MGA5TOO
(In reply to claire robinson from comment #4) > Please fix in cauldron first. It was fixed in Cauldron about 30 min. later. I increased the rel there to 15.
assigned back to QA
Assignee: thomas => qa-bugs
Changing it to a Mageia 5 bug then.
Version: Cauldron => 5Whiteboard: IN_ERRATA MGA5TOO => IN_ERRATA
Except package listed are mageia 4 Thomas. Is this fixed in Mageia 5? Is Mageia 4 relevant?
Whiteboard: IN_ERRATA => IN_ERRATA feedback
(In reply to claire robinson from comment #8) > Except package listed are mageia 4 Thomas. Ooops, copied them from my local build. Yes, they are fixed in mga5. I am going to list them again. > > Is this fixed in Mageia 5? > Is Mageia 4 relevant? No. mga4 is OK. The following packages are now in mga5 updates_testing: bacula-7.0.5-14.1.mga5.src.rpm lib64bacula7-7.0.5-14.1.mga5.x86_64.rpm lib64bacula-sql7-7.0.5-14.1.mga5.x86_64.rpm bacula-common-7.0.5-14.1.mga5.x86_64.rpm bacula-dir-7.0.5-14.1.mga5.x86_64.rpm bacula-sd-7.0.5-14.1.mga5.x86_64.rpm bacula-fd-7.0.5-14.1.mga5.x86_64.rpm bacula-console-7.0.5-14.1.mga5.x86_64.rpm bacula-bat-7.0.5-14.1.mga5.x86_64.rpm bacula-tray-monitor-7.0.5-14.1.mga5.x86_64.rpm lib64bacula-devel-7.0.5-14.1.mga5.x86_64.rpm nagios-plugins-bacula-7.0.5-14.1.mga5.x86_64.rpm bacula-debuginfo-7.0.5-14.1.mga5.x86_64.rpm and i586 packages
There was an additional error detected that prevented the program to write to the log. The reason was a spelling error in the configure section, logdir vs. log_dir. These packages have it corrected and are now in updates_testing: bacula-7.0.5-14.2.mga5.src.rpm lib64bacula7-7.0.5-14.2.mga5.x86_64.rpm lib64bacula-sql7-7.0.5-14.2.mga5.x86_64.rpm bacula-common-7.0.5-14.2.mga5.x86_64.rpm bacula-dir-7.0.5-14.2.mga5.x86_64.rpm bacula-sd-7.0.5-14.2.mga5.x86_64.rpm bacula-fd-7.0.5-14.2.mga5.x86_64.rpm bacula-console-7.0.5-14.2.mga5.x86_64.rpm bacula-bat-7.0.5-14.2.mga5.x86_64.rpm bacula-tray-monitor-7.0.5-14.2.mga5.x86_64.rpm lib64bacula-devel-7.0.5-14.2.mga5.x86_64.rpm nagios-plugins-bacula-7.0.5-14.2.mga5.x86_64.rpm bacula-debuginfo-7.0.5-14.2.mga5.x86_64.rpm and i586 packages
There will be another change coming the next few days.The Catalog isn't run because of perm issue.
Assigning back to Thomas in light of bug 16804
Assignee: qa-bugs => thomas
Hmmmm, I forgot about this. Well 16804 should take care of this one too. Shall we mark this one as resolved after 16804 has been pushed to repos?
16804 has been pushed, so closing this as fixed
Status: ASSIGNED => RESOLVEDResolution: (none) => FIXED