I installed my server following this tutorial: The Perfect Server - Debian 9 (Stretch) with Apache, BIND, Dovecot, PureFTPD and ISPConfig 3.1 (howtoforge.com) and then upgraded to Debian 10. Every day at 6.25am, I get a spike in read activity on all disks (initial server and all mounts). Since upgrading to Debian 10, this results in an unmount of the mounted disks (which come back online after rebooting). I reduced the problem to this line in /etc/crontab 25 6 * * * root test -x /usr/sbin/anacron || ( cd / && run-parts --report /etc/cron.daily ) If I run /etc/cron.daily, I can reproduce the problem. These are the cron jobs in cron.daily: apache2 apt-compat bsdmainutils dpkg locate logrotate man-db ntp passwd quota rkhunter spamassassin sysstat webalizer I hoped, by running them individually, I would be able to figure out which of these was causing the problems. But all the programmes ran without any read spike on the disks. So I have run out of ideas to solve this.
If disks get unmounted unexpectedly, I would try looking at all log files at around that time for clues.
I did, but non of the logs gave me some idea. But, in the meantime I found the problem, it is handled in this post https://bugzilla.redhat.com/show_bug.cgi?id=829553 Due to the updates in Debian, the database of man-db becomse buggy, sudo mandb -c solves the problem.
