Extraction of particles hangs forever (stain data)

open

#1

do not experience this issue with cryo data, only stain. I manually picked particles for subsequent template, and the extraction (<1000 particles) worked fine here. I checked the box “stain data”. However after template pick when I have >300k particles and I try to rextract it doesn’t work. I see in the output that it goes through all micrographs and extracts successfully, writes output to the folder, but at the end it never successfully ends the job. It just stays running forever. I let it go overnight.


#2

Dear all,
We are also facing more or less the same problem,
we are trying to extract the particles using “Extract from Micrographs” for the Negative stain data. the Job went to Queued but it never ran. I have used the output of CTFfind4 (Micrographs) and Inspect picks (particles) as a inputs for this job.

Any suggestions are welcome.

Best,
Eswar


#3

Hi @eswar909,

It seems like your issue is a little different than @orangeboomerang. Please refer to the resource manager tab (from the bottom bar) and you will be able to see the reason the job has stayed in the queue. There could be a few possible reasons, such as inputs not being ready or compute resources not being available.


#4

@orangeboomerang, we’re looking into the issue you reported. If you re-run the job with the same parameters, do you observe the same issue?

Best,
Ali


#5

I do see the same issue yes. Actually I am now seeing the issue while extracting even cryo data. Also I see it hanging even with refinements too, so I think it has to do perhaps with memory. I am running these jobs on a computing cluster, and I submit multiple jobs at once (trying to be productive!). I find if I submit more than 2 or 3 then they ALL get hung like this and never complete. Last night I started an extraction, it landed on the final image before going to bed, but I woke up this morning and still it was on that final image. Not sure if this is a cryosparc issue or something about the way our cluster is set up, but often cryosparc requests a GPU card and if it’s not available it will have one of the following 3 inconvenient outcomes (rather than just queuing up and starting when available):

  1. error “no cuda device available”
  2. memory allocation fail
  3. it will hang, as described above.