Possible bug with CentOS 7, MariaDB and FOG services.
May have found a possible bug with CentOS 7, with MariaDB and FOG build 4375…
Rebooting CentOS 7, checking Systemctl, seeing the following:
FOGImageReplicator.service loaded inactive dead FOGImageReplicator
FOGMulticastManager.service loaded inactive dead FOGMulticastManager
FOGScheduler.service loaded inactive dead FOGScheduler
FOGSnapinReplicator.service loaded inactive dead FOGSnapinReplicator
Doing a systemctl status on one of the services…
FOGScheduler.service - FOGScheduler
Loaded: loaded (/usr/lib/systemd/system/FOGScheduler.service; enabled)
Active: inactive (dead) since Wed 2015-08-12 13:09:50 PDT; 3min 32s ago
Process: 878 ExecStart=/opt/fog/service/FOGTaskScheduler/FOGTaskScheduler & (code=exited, status=0/SUCCESS)
Main PID: 878 (code=exited, status=0/SUCCESS)
Aug 12 13:09:43 control systemd: Started FOGScheduler.
Aug 12 13:09:50 control FOGTaskScheduler: PHP Warning: mysqli::mysqli(): (HY000/2002): Connection refused in /var/www/html/fog/lib/db/MySQL.class.php on line 37
Aug 12 13:09:50 control FOGTaskScheduler: PHP Warning: mysqli::mysqli(): (HY000/2002): Connection refused in /var/www/html/fog/lib/db/MySQL.class.php on line 39
Aug 12 13:09:50 control FOGTaskScheduler: PHP Warning: mysqli::query(): Couldn’t fetch mysqli in /var/www/html/fog/lib/db/MySQL.class.php on line 63
However, when you try and restart it, works fine. So this could be a timing issue with MariaDB and/or CentOS 7…
I’m still working on it but current trunk should have a better mechanism to handle all of this. Hopefully.
@SKasai There’s a similar issue with Fedora. Here is my fix that Tom helped me with this past spring. I’ve simply beautified the steps here lol. https://wiki.fogproject.org/wiki/index.php/Fedora_21_Server#Set_the_FOG_services_to_start_30_seconds_after_boot
Ok… Then basically for now, just have a hack shellscript to do…
/usr/bin/systemctl restart FOGImageReplicator.service
/usr/bin/systemctl restart FOGMulticastManager.service
/usr/bin/systemctl restart FOGScheduler.service
/usr/bin/systemctl restart FOGSnapinReplicator.service
Got it. (Yes, I know it can be manually started, but easier just to do this for ‘deadhead’ purposes).
This is a known issue but not a known and nice method of actually fixing it. Particularly for centos 7, it’s fairly simple to just require mariadb service, but this is not necessarily the right approach as other systemd setups may not require mariadb.