Parallel solvers not terminating

[Robert Lancaster]

Yeah, this could be something with his system.  It works fine on all my systems.  I was wondering if we could see the reason it doesn’t work for him.  I’m not sure what it could be yet, but maybe the cancel files are not getting made for some reason.  Maybe there is something up with his temp directory?

> On Nov 17, 2020, at 3:43 PM, Jasem Mutlaq <mutlaqja at ikarustech.com> wrote:
> 
> I tried this in load and slew. While it failed to solve the image, it terminated correctly.
> 
> --
> Best Regards,
> Jasem Mutlaq
> 
> 
> 
> On Tue, Nov 17, 2020 at 11:39 PM Robert Lancaster <rlancaste at gmail.com <mailto:rlancaste at gmail.com>> wrote:
> Hi Wolfgang,
> 
> Does this happen when you try to solve the image in the StellarSolver tester, or just in KStars?  Does it happen when using the internal solvers and the local astrometry.net <http://astrometry.net/>, or just the internal ones?  Does it happen on more than one computer or just this one?
> 
> Can you look in the temp directory and watch what happens?  The way astrometry.net <http://astrometry.net/> works, to abort the internal or external solvers, a “cancel” file must be made, and once it notices that file, it aborts the solve.  It is basically an empty file.  What you should see happen is that the one solver solves the image, and then Stellarsolver tells the other solvers to abort, it does this by creating a cancel file for each one.  Then when each one shuts down, it will delete its cancel file.
> 
> I am concerned from your log here that possibly on your system the cancel file does not have the correct name so it is not being made perhaps. But I need more information to be sure.
> 
> The current naming scheme is that each solver gets a number, and any temp files it creates would have this number appended.  The internal solvers will have very few of these temp files (just solved and cancel), but the external ones have multiple files.  The cancel files should look something like this: "internalSextractorSolver_20.cancel”
> 
> Can you see if this happens in this case.
> 
> Thanks,
> 
> Rob
>> On Nov 17, 2020, at 2:01 PM, Wolfgang Reissenberger <sterne-jaeger at openfuture.de <mailto:sterne-jaeger at openfuture.de>> wrote:
>> 
>> Hi Robert,
>> I am running StellarSover on commit 2cac32ffb3523dbcb4b3d43afd73e96959b52d8d, but the problem when solving NGC6888 remains. It seems like after a successful termination of one solver, the others keep running.
>> 
>> There appears a warning around the first successful solve:
>> 
>> [2020-11-17T19:47:27.267 CET INFO ][     org.kde.kstars.ekos.align] - "Successfully solved with child solver: 1"
>> [2020-11-17T19:47:27.269 CET INFO ][     org.kde.kstars.ekos.align] - "Shutting down other child solvers"
>> [2020-11-17T19:47:27.270 CET WARN ][                       default] - Empty filename passed to function
>> [2020-11-17T19:47:27.270 CET WARN ][                       default] - QFile::remove: Empty or null file name
>> [2020-11-17T19:47:27.271 CET WARN ][                       default] - QFile::remove: Empty or null file name
>> [2020-11-17T19:47:27.271 CET INFO ][     org.kde.kstars.ekos.align] - "Solver completed after 1.76 seconds.“
>> 
>> The entire log file is here:
>> https://drive.google.com/open?id=1UTy8HCBSg2GILYvuzHlrIWKqrTGw7rYL <https://drive.google.com/open?id=1UTy8HCBSg2GILYvuzHlrIWKqrTGw7rYL>
>> 
>> Wolfgang
>> 
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.kde.org/pipermail/kstars-devel/attachments/20201117/f4ad0bf9/attachment-0001.htm>