Restart script requires 'sudo'

[bradford 5]

I brought this up in my other thread, but I think this issue deserves its own.
 
Server OS: Debian 8.6 LXC container in Proxmox 4.4 (Debian 8.6)
Emby Server: 3.1.2

Anytime I restart emby from the management page, emby dies but doesn't come back. I tried to track down in the logs what was happening, and the only thing I saw was a reference to restart.sh. I tried manually running 

bash /usr/lib/emby-server/restart.sh

but there's no indication it actually ran, and my bashfu is weak. But I did echo the command it runs on my system, which is actually 

service emby-server restart

I twiddled around with it, and found that 'sudo' isn't installed by default on debian, at least not on my Debian 8.6-based Proxmox server nor my Debian 8.6 lxc container. 
 
Removing 'sudo' from the script enabled the script to restart emby (running the script as root).
 
I then installed sudo from the regular Debian (jessie) repositories, added 'sudo' back to the script, and it runs fine as well.
 
What's puzzling is that when I try to restart the server from the management board, emby dies, but it never comes back - it appears that it uses a different script when initiated from within emby (I suspect that restart.sh is used by the system if emby crashes).

 

Anyway, I tried restarting from within emby again to see if installing sudo fixed the issue and it does.

 

I know I'm an edge case that doesn't have sudo. Does emby need sudo in order to run a privileged command like restarting itself? If that's the case, at what point did I allow emby to run the privileged command? Certainly not just by installing sudo, right?

 

In any case, this can be a reference to anyone else with the same issue, and possibly something the devs can use to include in their deployment strategy.

Edited December 29, 2016 by bradford

[bradford 5]

Perhaps the restart script is suid root? In that case sudo isn't needed. Just a thought.

[hurricanehrndz 149]

I brought this up in my other thread, but I think this issue deserves its own.

 

Server OS: Debian 8.6 LXC container in Proxmox 4.4 (Debian 8.6)

Emby Server: 3.1.2

 

Anytime I restart emby from the management page, emby dies but doesn't come back. I tried to track down in the logs what was happening, and the only thing I saw was a reference to restart.sh. I tried manually running

bash /usr/lib/emby-server/restart.sh

but there's no indication it actually ran, and my bashfu is weak. But I did echo the command it runs on my system, which is actually

service emby-server restart

I twiddled around with it, and found that 'sudo' isn't installed by default on debian, at least not on my Debian 8.6-based Proxmox server nor my Debian 8.6 lxc container.

 

Removing 'sudo' from the script enabled the script to restart emby (running the script as root).

 

I then installed sudo from the regular Debian (jessie) repositories, added 'sudo' back to the script, and it runs fine as well.

 

What's puzzling is that when I try to restart the server from the management board, emby dies, but it never comes back - it appears that it uses a different script when initiated from within emby (I suspect that restart.sh is used by the system if emby crashes).

 

Anyway, I tried restarting from within emby again to see if installing sudo fixed the issue and it does.

 

I know I'm an edge case that doesn't have sudo. Does emby need sudo in order to run a privileged command like restarting itself? If that's the case, at what point did I allow emby to run the privileged command? Certainly not just by installing sudo, right?

 

In any case, this can be a reference to anyone else with the same issue, and possibly something the devs can use to include in their deployment strategy.

The emby package itself contains a sudoers file. Setuid is flagged as a bad packaging practice

 

Sent from my ONEPLUS A3000 using Tapatalk