Issue Backup Restore Job takes Sites offline

[dannray]

Hi,

Is it normal when you do a large backup restore job for it to restart PHP or take all other sites offline temporarily?

Thanks

 

[Bitpalast]

It does this on one of these two events:
- You have clicked the checkbox that during the restore the site shall be offline.
- The load is so high compared to the capabilities of the server that the site takes so long to respond that it reaches a timeout and appears to be offline.

 

[dannray]

Hi Peter,

Thanks for the reply, is there any way to avoid this?

When the restore was taking place it seemed like all the sites on the server were pointing to the login or default index.html screen like the apache config or Nginx config were broken.

I had to rebuild the web using Plesk repair web -y to fix this.

Thanks

 

[Bitpalast]

A restore also reconfigures all webserver configuration files. This process can take several minutes, depending on the number of configuration files that need to be created. The process first deletes all such files and then restores the files for each subscription separately. It is correct that during this phase the web server points all traffic to the default page, because the web server does not know the virtual hosts yet. This is an unavoidable behavior. I can think of ways how this could be done better, e.g. create all new config files in a shadow directory and then move them into their destination all at once, but currently it's a one by one algorithm.

It is normally not necessary to run a "plesk repair web" on it. This will only reconfigure the server again. Instead, when the issue occurs, look into the system with
# ps aux | grep mng
or
# ps aux | config
which will reveal the ongoing reconfiguration processes.