Thanks for thinking out loud Ken! I have the server where I truncated the mdl_backup_courses table up and running again and there the automated backup was running ‘normal’ again.
I did this in three steps:
- truncate mdl_backup_courses and wait until the table was recreated again with all the courses (laststatus for all courses = 5).
- I updated the table with laststatus = 1, start/endtime with a valid timestamp (endtime after starttime) and nextstarttime with the timestamp for the next nightrun
- the automated backup uses the parameters that are valid for the (nightly) batchrun
This helped me to get it working again on one server (not the production server and with Moodle 3.5.4+!)
It’s not the ideal situation but a workaround to get the automated backup run back on track again.
Problem is the mismatch with the automated backup and courses that fail or take to long time to run. We can only try to exclude them manually from the ‘good’ courses. A bash script run by the Linux cron with the backup cli script was the only way to find out that a course was taking 7 hours for making a backup! These are the courses that cause a headache to administer this backup process.