python - Celery beat task stops fails after a few beats (opencv) -

i have celery task runs on beat every 60 seconds grab webcam, take still it, , write file server can serve on webpage later. image taking uses opencv via cv2 library in python.

it looks this:

    @app.task                               # task , save webcam image     def getwebcamimage()                         logger.debug("capturing image attempt")     c = cv2.videocapture(0)                 #returns videocapture object     flag, frame = c.read()                  #grabs , decodes next frame     cv2.imwrite('file/loc/img.jpg, frame)   #writes frame file     logger.debug('saved image...hopefully')      c.release()                             #releases videocapture object     return 0 

everything goes great first few minutes, spits out normal errors not being able webcam properties , stops. , never starts again!

    ...debug/beat] celerytest.getwebcamimage[process]:capturing image attempt     vidoc_querymenu:invalid argument     vidoc_querymenu:invalid argument     vidoc_querymenu:invalid argument 

this how looks when work, keeps going. idea stopping it? running on rasperry pi 240mb dedicated arm core , 16 graphics.

things have done try fix it:

    $sudo sysctl vm.overvommit_memory=1 

so doesent throw mem. allocation errors $sudo rmmod ucvideo $sudo modprobe ucvideo nodrop=1 timeout=5000 quirks=0x80 driver more patient cam. have tried few cams, not issue.

please whatever suggestions or tests think of. tricky 1 because is: on pi, , pi/cellery cross. going try running script guvcview, feel py/opencv option cleaner because easier install, , either way know why doesn't work!

news:

dmesg reveals celery process being killed because out of memory. way fix this? how, if @ all, objects created in tasks destroyed?

this turned out unrelated celery, , related cv.

if every time cv2.videocapture object released, not deleted, memory still reserved. on pi, has little memory, took memory , killed. investigated memory usage top command in linux.

it happens on larger computers, symptoms more severe on pi.