Sjors (and Relion users)
I have been stuck for over a week with the same error when
running a very long ptcl extraction job from my movies after getting
a good initial map:
This job involves extracting 42,000 movie ptcls and then writing
the final movie.star metadata file,
The cryptic error after ALL the movie files are made!!
91.52/91.52 hrs
............................................................~~(,_,">
Joining all metadata in one STAR file...
Killed
I tried 3 different things and it always MAKES all the movie.mrcs and
small movie.star files but gives this error when writing the final
big movie.star file that is needed for the polishing step.
I ran a test job with the first 4-5 ptcls in the input star file under the same conditions
which works fine and creates a small, final movie.star meta data file.
Is there any way I can debug this problem, it is not a naming issue.
These final movie star files are huge > well over 500,000 lines! (based on my
inspection of a previous project that worked in our group, that had somewhat fewer ptcls).
When you have 40,000 ptcls it may be a problem with memory on the head node of
the cluster where I am running this job. (I am using this node since it
has the best IO bandwidth to the raid with the movies on it and the job
already takes 3-4 days!!).
C Akey
Prof.
|