Thanks for the tip Leonard! Here the output of ls -l /proc/29669/fd, where 29669 is the PID of kstars when running an acquisition sequence from the CCD simulator (the same happens with the ASI camera, connected remotely or not):
...
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1570 -> socket:[876722]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1571 -> socket:[876723]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1572 -> socket:[871147]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1573 -> socket:[871148]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1574 -> socket:[874775]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1575 -> socket:[874776]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1576 -> socket:[875769]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1577 -> socket:[875770]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1578 -> socket:[871149]
lrwx------ 1 bertin bertin 64 Feb 25 00:14 1579 -> socket:[871150]
...And it goes on and on until I stop the acquisition. Aborting does not close the sockets, I have to quit kstars to free the descriptors. It happens only when the exposure time is about 1s or less, irrespective the raster size (the simulator takes only a few ms to generate the 128x128 rasters in this test). This is with fast looping off (I have to turn it off otherwise it creates all kinds of issues). The other processes (indiserver and indi_simulator_ccd) do not accumulate descriptors.
- Emmanuel.
Read More...
© 2003-2022 by INDI Library. All rights reserved.