Since I did the upgrade from jessie to stretch I have a few issues. They
both relate to services that start before services on which they rely
upon are not available yet.
I have iscsi that uses a LUN from a SAN. This LUN is used by LVM. On 1
LV there are the files for PostgreSQL. When PostgreSQL starts the LV is
not yet available and PostgreSQL does not start. Same for 2 other LVs
(for TomEE and Sonar).
I use autofs to automount NFS shares. The configuration is stored in
LDAP that runs on the same server. Autofs starts when slapd is starting
up. As a result there is nothing to automount.
I use cron as a (ugly) solution: @reboot /bin/sleep 60; /bin/systemctl
restart autofs; /bin/systemctl restart postgresql
Since I created the systemd scripts for Sonar and Tomee 'myself' I added
'Requires=postgresql.service' to the [Unit] section of both scripts. For
postgres and autofs I used 'systemctl edit' with '[Unit]' and
'Required=open-iscsi.service' or 'Required.slapd.service'. This did not
solve my issues.
Before on jessie it ran without a problem. How can I fix/debug these issues?
