[Rwp] a fun Reaper crashing experience

[Patrick Perdue]

Interesting. I haven't changed/updated drivers for any of the connected 
devices on this system in quite a while, and I'm literally only seeing 
this behavior as of yesterday.
I will install ReaAccess, and see if it does similar bad stuff.


On 4/9/2015 1:13 AM, James Teh via RWP wrote:
> 1. Does the freeze occur if you use ReaAccess instead of OSARA? It's
> possible the freeze is an OSARA bug, though I've no idea what might be
> causing it.
>
> 2. As for why the process hangs around even once you terminate it, I've
> seen this before. My (fairly certain) guess is that if a DAW doesn't
> release the audio device correctly and it still has outstanding requests
> (i.e. because it terminated abnormally), the driver doesn't cancel those
> requests correctly when Windows asks it to. Basically, when you exit a
> process, Windows asks drivers to cancel all outstanding requests, but if
> a driver is buggy (which sadly applies to a lot of audio drivers), it
> can get stuck. I see this quite a bit with abnormal exits with my
> M-Audio Fast Track C600. I also see occasional blue screens with this
> driver if I'm doing stuff other than just DAW work. The problem with
> semi-pro/pro audio hardware vendors is that if you do anything they
> consider slightly outside the norm for a DAW (like, you know, put your
> computer to sleep instead of shutting it down or have a DAW which
> occasionally crashes), they just say that's unsupported and thus justify
> not fixing their stupid driver bugs. Rant aside, if you want some
> technical detail about this unkillable process weirdness, see this page:
> http://blogs.technet.com/b/markrussinovich/archive/2005/08/17/unkillable-processes.aspx
>
>
> So, in short, if we can fix the freeze, you won't end up with an
> unkillable process... but the fact that the process is unkillable after
> the freeze is almost certainly a driver bug.
>
> Jamie
>
> On 9/04/2015 2:49 PM, Patrick Perdue via RWP wrote:
>> So, I've been using Reaper for about five years, regularly for a bit
>> over 4, and I've never seen this one.
>>
>> I'm using Reaper v4.77x64 with OSARA, and am having a fun issue
>> wherein Reaper crashes almost every time I leave it's main window and
>> come back.
>> Not quite a normal crash though.
>> It stops responding, so I press alt+f4 to close the window, which
>> gives the normal "reaper is not responding" error.
>> So, after the UI goes away, reaper.exe, ReaperHost32 and
>> ReaperHost64.exe are all still in the task manager. If I try to end
>> reaper.exe, nothing happens. Same for ending the process tree. If I go
>> to the command line and type "taskkill /im reaper.exe /f"
>> I get an error similar to
>> "The process reaper.exe with PID 1984 could not be terminated. Reason:
>> an instance of this process is not running."
>> So, I still have this old instance of Reaper hanging around, taking up
>> a bunch of memory. I can launch Reaper again, but when I do, I get
>> errors about not being able to access audio and midi devices,
>> presumably because the old Reaper still has access to those devices,
>> and won't close properly. Even if I log out of Windows and don't
>> restart, while the process is no longer running, those devices still
>> can't be accessed, unless I restart Windows.
>>
>> While editing a project, I've probably restarted this computer an
>> average of 7 times in an hour. A bit hard to be productive like that.
>>
>> If it matters, I am using bridged mode, with both 32 and 64-bit VST
>> plugins. No DirectX, converted or otherwise.
>> Reaper even crashes sometimes when going to effects windows for
>> tracks, but not consistently with one plugin. It could be that I have
>> a bad something in the chain of one of the tracks in this project,
>> which is causing Reaper to flop over, but I haven't been able to
>> figure out if that's the case, because it crashes when I try.
>>
>> Has anyone ever seen this particular issue, or something like it?
>> I realize that, without the exact combination of hardware, software
>> and project I'm using, probably no one will be able to replicate this
>> exactly, but I'd sure love to have a clue about why this is happening.
>> _______________________________________________
>> RWP mailing list
>> RWP at bluegrasspals.com
>> http://bluegrasspals.com/mailman/listinfo/rwp
>