"Couldn't connect to host" "Could not resolve host" {"success": false} "tar: This does not look like a tar archive" "Version mismatch! Worker and master versions are not the same. Please update." "An unexpected error has occurred."

• If you have Conda installed, deactivate any active environments.
• Check that your LICENSE_ID environment variable is set correctly with this command

• Check your machine's connection to cryoSPARC's license servers at get.cryosparc.com with this curl command:

• ​Reinstall cryoSPARC, following the steps from the "Cannot download cryoSPARC" section

• In a command line, run cryosparcm status
• Check that the output looks like this

• Check that all items under "cryosparcm process status" that do not end in _dev are RUNNING. If any are not, run cryosparcm restart
• If any non-_dev components have a status other than RUNNING (such as STOPPED or EXITED), check their log for errors. For example, this command checks for errors on the database process:

• Check that you are on the master node or workstation where cryosparc_master is installed
• Run echo $PATH and check that it contains <installation directory>/cryosparc_master/bin

• If the web interface is inaccessible, check firewall settings to ensure cryoSPARC's base port number (default 39000) is exposed for network access
• ​Reinstall cryoSPARC​

1. 1.

   Run the following command in your terminal: cryosparcm listusers
2. 2.

   If you notice that your email address is incorrect or there is an extra space at the beginning or end, you'll need to modify it in the database. To do this, run the following commands:

   1. 1.

      Log into the MongoDB shell: cryosparcm mongo
   2. 2.

      Once in the MongoDB shell, run (replace the incorrect/correct email) db.users.update({ 'emails.0.address': '[email protected]' }, { $set: { 'emails.0.address': '[email protected]' } })
   3. 3.

      Exit the MongoDB shell: exit
3. 3.

   If you don't remember your password you can replace it with the following command (replace with your email address and new password): cryosparcm resetpassword --email "<email address>" --password "<new password>"

"License is invalid." "License signature invalid." "Could not find local license file. Please re-establish your connection to the license servers." "Local license file is expired. Please re-establish your connection to the license servers." "Token is invalid. Another cryoSPARC instance is running with the same license ID."

• Ensure you entered your license correctly during the installation step.
• Check your Internet connection
• Check your machine's connection to cryoSPARC's license servers at get.cryosparc.com with this curl command (substitute <license> with your unique license ID):

• If you see a license ID conflict such as this

"Another cryoSPARC instance is running with the same license ID."

"list index out of range" "Could not resolve hostname ... Name or service not known"

• Ensure at least one worker is connected to the master. See the Installation page for details.

• Check that all non-development cryoSPARC processes are running with the cryosparcm status command
• (For master/worker setups) check that SSH is configured between the master and worker machines.
• Check the log for the command_core log to find any application error messages

• If applicable, check that the cluster submission script is correct
• Refresh job types and reload cryoSPARC:

• Force-reinstall master and worker dependencies. This can help when a worker was not correctly installed.

• Restart cryoSPARC:

• Invalid or unspecified input slots
• Invalid or unspecified required parameters, including file/folder paths
• Incorrectly set up GPU (e.g., running a job on a node without enough GPUs or CUDA drivers not installed)
• Another process taking up memory on a GPU
• Cache not set up correctly for a worker

"AssertionError: Child process with PID ... has terminated unexpectedly!"

"cuInit failed: unknown error" "no CUDA-capable device is detected" "cuMemHostAlloc failed: OS call failed or operation not supported on this OS" "cuCtxCreate failed: invalid device ordinal kernel.cu ... error: identifier "__shfl_down_sync" is undefined

"cuMemAlloc failed: out of memory" "cuArrayCreate failed: out of memory" "cufftAllocFailed"

• Ensure a CUDA version ≥10 is installed and running on the workstation or each worker node
• Check the GPU configuration on the workstation or node where the job runs on. Log into that machine and navigate to the cryoSPARC installation directory. Run the cryosparcw gpulist command:

• Run nvidia-smi to check the GPU and CUDA driver status.

  • Check that CUDA version is within the allowed range. cryoSPARC supports CUDA versions ≥10
  • Check that no other processes are using GPU memory. cryoSPARC-related process appear with process name "python"

• Check the Job log: Select the Job card in the cryoSPARC interface and press the Spacebar on your keyboard to see the log. Scroll down to the bottom and look for the failure reasons in red

  

  ​
• Clear the job and select the "Building" status badge on the job card to enter the Job Builder
• If the job failed with a GPU-related error and multiple GPUs are available, try running the job on a different GPU. Press Queue, switch to the "Run on specific GPU" Queue type and select one or more GPUs

  
• If the job failed with AssertionError: Non-optional inputs from the following input groups and their slots are not connected then expand any input groups connected to the job. Missing required slots appear in red

  
• Check the job parameters To learn about setting specific parameters, hover over or touch the parameter names in the Job Builder to see a description of what they do

• Reduce the box-size of extracted particles. Some jobs need to fit thousands of particles in GPU memory at a time, and larger box sizes exceed GPU memory limits. Either extract with a smaller box size or with the Downsample Particles job

• Refresh job types and reload cryoSPARC:

• Look for extended error information with the cryosparcm joblog command (press Control + C on the keyboard to exit when finished)
• On occasion, a job fails due to an error in the cryoSPARC code (bug). When a bug is discovered, the cryoSPARC team promptly patches the issue and makes it available in the next update.

• ​Connect workers with SSD cache enabled. This speeds up processing for extracted particles during 2D Classification, ab-initio reconstruction, refinement and more. Ensure the "Cache particle images on SSD" parameter is enabled under "Compute settings" for the target particle-processing job
• Some jobs (motion correction, ctf estimation, 2D classification) can run on multiple GPUs. If your hardware supports it, increase the number of GPUs to parallelize over under

• Extracted particles with large box sizes (relative to their pixel size) take a long time to process. Consider Fourier-cropping (or "binning") the extracted particle blobs with the Downsample Particles job

• Minimize the number of processes using system resources on the workstation or worker nodes
• Check for zombie processes on worker machines. The process is similar to the steps under "Another cryoSPARC instance is running with the same license ID" under the License error or license not found section
• Cancel the job, clear and re-queue