MB3 restart not allowing enough time to restart

[dcauger 3]

Seems that either auto-restart does not allow enough time for the existing server instance to cleanly exit before giving up on the instance its trying to spawn. Both instances get killed. see attached logs (short logs as these are from an auto-restart earlier this AM).

 

It appears that the new server instance gives the old server instance 10 seconds to exit. My server can not seem to get exit in the 10 second window. I am unsure what it is doing for those ten 10 seconds, or how long it really takes to finally exit, but I can tell say that when I check my processes no media browser components are hung. I can cleanly start the server. I think I need more than 10 seconds.

 

my server is set to auto-restart at 5AM. The two logs attached should show the transition between the old and new instances.

 

In my case,

- I have a solid state disk mounted at ..AppData\Roaming\MediaBrowser-Server. This is about 40% free (1.3GB free)

- All of the media browser meta data is stored outside of the media folders on another solid state disk which is shared with image-by-name data and transcoding. In the logs this is identified as the "I:" drive. This is roughly 30% free (13GB free).

 

I don't know what MB3 is doing when it is shutting down, but whatever it is, it should be doing it all on SSD. I'm puzzled why it would take longer than the 10 seconds allocated for the process.

 

thoughts?

 

DCA

server-63554111999.txt

server-63554130002.txt

[dcauger 3]

Update:

 

I went to the beta version 3.0.5464.40000 - still have this problem. I can provide more logs if that helps. For now, I've eliminated scheduled restarts.

 

DCA