When you say "clean startup", if you mean after a fresh PC reboot, the answer for me is YES. My exit procedure is by clicking on END FLIGHT and then EXIT P3D. Approx 30 - 40 % of these exits will still show the program running until I kill process.

I have observed a consistent behavior .. If you start P3d and you get the training scenario dialog and exit P3d from there without actually going into flight/missions, P3d exits properly from the task manager. But if you get into a flight and then exit P3d from there, it continues to run In the task manager..

Do you guys use ASN?  For me I've noticed when this has happened the asn file (im not at my pc now to tell you the name) but it's the add-on driver that runs from the as_srv folder in prepar3d.  I will notice that is still running in the task manager.  If I force close that, then the hung instance of Prepar3d will close itself out.

Do you guys use ASN?  For me I've noticed when this has happened the asn file (im not at my pc now to tell you the name) but it's the add-on driver that runs from the as_srv folder in prepar3d.  I will notice that is still running in the task manager.  If I force close that, then the hung instance of Prepar3d will close itself out.

Yes, I do. And I had this module deactivated in 2.4 - where Prepar3d.exe closed properly. I don't have time for any tests right now, but this might give an idea.

 

Kind regards, Michael

I only run OPUS FSI 

This can happen:

 

1.  Check your EXE.XML ... any entries in this file have the potential of running AFTER P3D terminates (Ultimate Traffic 2, GoFlight, etc.) ... on a restart of P3D these existing processes can become a problem as EXE is loaded twice or if coded to only load single instance might reference  a state that is no longer active/valid.

 

2.  Any 3rd party product that goes outside the SDK/PDK and directly access memory offsets in P3D can cause this issue (FSUPIC, ASN, and others)

 

3.  Being very close to OOM (3.8 to 3.9 GB VAS used)

 

For the EXE.XML entries they provide the filename (exe being used) those are easy to spot in taskmgr and need to be terminated.  However, those EXEs may not like being terminated ... for example, if I do this (kill task) with GFDev.exe (GoFlight) then I'll need to actually start GFConfig to clear any module states in my GF devices.

 

Ultimate Traffic 2 will sometimes also not terminate and I need to make sure that is killed prior to re-loading P3D.

 

Finally, sometimes P3D just takes a while to shutdown.

 

Cheers, Rob.

When you say "clean startup", if you mean after a fresh PC reboot, the answer for me is YES. My exit procedure is by clicking on END FLIGHT and then EXIT P3D. Approx 30 - 40 % of these exits will still show the program running until I kill process.

Rob has already given the answer what I mend in his point 1 ;-)

just to let you all know, when i reported this problem i was on p3d v2.5 (build 12943). After i upgraded to the hotfix 12944, I am not seeing this problem anymore in the same situations when it used to occur.

 

So something got fixed with the 12944 hotfix patch.

 

 

So something got fixed with the 12944 hotfix patch.

 

I still see the issue in 12944 ... Windows will actually log the slow shutdown of Prepar3d.exe in the OS Event log (use Event Viewer to see entry).  The delayed shutdown usually happens when I'm very close to OOM 3.8GB VAS usage and exit P3D.  In some extreme cases it can take as long as 1 minute to finally terminate.

 

I've only been able to duplicate this scenario with a heavy set of add-ons (fully loaded DLL.XML and EXE.XML).  

 

But to add more information I would have a slow shutdown in v2.4 also -- especially if I had UT2 loaded which has an entry in the EXE.XML.  In v2.4, UT2 would sometimes NOT unload when exiting which would cause all kinds of issues the next time I loaded P3D.  My guess is LM did some additional work in P3D v2.5 shutdown to wait for these EXE processes to unload before it unloads.

 

If you have a base v2.5 (no add-ons at all) try to see if you can duplicate the slow shutdown ... I wasn't able to create the situation, but maybe others can on a base v2.5.

 

Cheers, Rob.