We recognize lots of Apache Restarts when setting up users and the configuration by API (mailboxes, ssh users, etc.). 1 User (1 vhost, 10 aliasvhost, 5 DB...) lead to 76 Apache restarts. The ISP Backend is almost down during it work of the appr. 200 jobs. How can we reduce that??? The issue with apache configuration check (restart vs. reload) is known, but are 76 restarts necessary/normal?
A restart is required for the config check as that's the only way to determine reliably if the config is valid. Disable the apache config check under system > server config to reduce restarts. Apache is not restarted for mail things and not for ssh users etc. It can't be 76 restarts from 1 domain plus 10 alias vhosts as only the apache or nginx module restart the web server and the code of these modules is not triggered for mail, ssh, ftp and dns.
Hi Till, we checked the creation by api again. ISP restarts the webserver for each domain entry as you said. The problem here is, if we migrate a user with 40 Subdomain and other modul entries, all websides on the server are not accessible for about 1 minute. Is there a possibility to check all apache entries at once so that the server restarts only one time? Btw. how the active processes are ended. Is there a gracefully shutdown or are they killed on a restart?
That's not possible as the script does not know if there might come a future update for the same vhost and without checking each vhost change, ispconfig is not able to see which of the changes caused the failure to roll back the config. If you want to have just one restart for all changes, switch off config check. In that case, only one restart will happen for all changes that got applied, but the config cannot be checked anymore then, so you have to ensure that. Depends probably on how the init script / systemd is doing it. We can not use reload or force-reload as a reload does not apply SSL file changes.
