2

I have a server with Proxmox 3.4 with several OpenVZ-containers. Host and all guests are running Debian 7 x64. Now I want to migrate my old setup to a new server with Proxmox 4 (based on Debian 8).

I transferred multiple OpenVZ-containers to the new environment and imported them successfully.

Now we get to the problem: After I upgraded them to Debian 8 most mysqld do not start any more. Setup showed me changes in my.cnf so I took the package maintainers version. One machine still starts mysql, some others do not although the configs are identical.

The mysql-error-log does not show anything, syslog neither. The startup script in '/etc/init.d/mysql' and config in '/etc/mysql/my.cnf' are identical on the machines.

root@web2:~# service mysql restart
Job for mysql.service canceled.

Starting the mysqld directly works:

root@web2:~# mysqld_safe
or
root@web2:~# /usr/sbin/mysqld --basedir=/usr --datadir=/var/lib/mysql --plugin-dir=/usr/lib/mysql/plugin --user=mysql --log-error=/var/log/mysql/error.log --pid-file=/var/run/mysqld/mysqld.pid --socket=/var/run/mysqld/mysqld.sock --port=3306

Can somebody give me a hint about how to fix this startup- or service-issue?

Any help or hints where to search are appreciated.

Marius
  • 141
  • 10
  • I did some more checking: It must be something that happens in dist-upgrade (not in upgrade). Also interesting: after the dist-upgrade mysql runs and works until the box is rebooted. – Marius Feb 26 '16 at 07:53

2 Answers2

2

The problem was related to ISPConfig, which created entries in /etc/fstab. The new default init-system systemd does not accept the lines any more (compared to sysvinit).

I found the solution by using journalctl -xb. Perhaps it helps someone else as well.

I added nofail to the entries in /etc/fstab to fix the boot-problems, e.g.:

# cat /etc/fstab
/var/log/ispconfig/httpd/example.com /var/www/clients/clientXX/webYY/log    none    bind,nobootwait,nofail    0 0
Marius
  • 141
  • 10
0

I had some issues with a bunch of random packages losing all their locales information when I moved containers from OpenVZ to LXC. I don't know much about MySQL but maybe it's not sure what to do about that. Try 'dpkg-reconfigure locales' as root, that fixed Perl for me when it was having a fit about there being no default locales. Don't know why it happened, but every one of the containers I migrated did the same thing.

K4KFH
  • 101
  • There was something wrong with locales indeed. But fixing this didn't change the mysql-issue. – Marius Feb 22 '16 at 16:01