Dear cisTEM team,
I am new to a cisTEM processing workflow and have a problem with finding the output files after a movie alignment run using either Apoferritin or my own dataset. The only difference is that, with apoferritin dataset, there is no problem occuring, but in case of my own data, one of the movie files was causing problem and I couldn't find out which one is that out of ~1000 movies. As a consequence, cisTEM processes were stalled and the job failed to be finished. Later on I managed to figure out which movie was causing problem, but am wondering if there is some sort of *.err file that cisTEM generates and one could access and find the problem easily?
Please let me know if you need some more details regarding the run and thank you for your help!
Kind wishes,
Emilia.
Hi,
The output aligned movie sums will be in Assets/Images in your project directory. If one is not working, one way to tell which one it is is to see which file in that directory has a zero file size.
I'm afraid there are no text output files beyond what you see in the GUI. It should be that no movie will give you a problem that stalls the job however - Do you know what about the movie caused it to break? Or would you be able to share it with us, so we can make sure similar problems don't cause stalls in the future.
Thank you!
Tim
Hello Tim,
thank you very much for clarifications. I was searching for aligned movie sums in Assets/Movies folder instead of Assets/Images…
Regarding the problem with stalls of Movie Alignment job and a partial solution to it: I was able to trace one of the movies being not properly copied from the source directory to the cluster where cisTEM calculations are done by comparing MD5 hashes. After copying it properly, the job was finished successfully with proper outputs. Nonetheless, I would call it a partial solution, since when attempted to run the same job with fewer frames to be included in final sum, a similar problem appeared... here is the link to a screenshot with GUI showing an error message (https://imgur.com/a/hwsw5). The way it is set up on our local HPC cluster is to split the process by executing 128 cisTEM jobs (with 10 ms launch delay) and during the Movie alignment step (this problem did not occur for further processing steps), one job all the time is left in the batch. I can terminate the job from GUI manually but after that, there is no output sum. Is there an obvious solution to it that we have overlooked?
Cheers,
Emilia.
Hi Emilia,
Is it always the same movie that hangs? I.e. that has 0 size in the Assets/Images directory? If so, do you think you would be able to send it to me?
Cheers,
Tim