I updated the master to 4.2. Then the auto update of the worker failed. I noticed default cuda is 12.0. I changed to 11.5. Now if I try to update the worker it says it is up to date, but it failed.
^
error: command '/usr/bin/gcc' failed with exit code 1
[end of output]
note: This error originates from a subprocess, and is likely not a problem with pip.
error: legacy-install-failure on pycuda. Can I delete the worked and install it fresh?
à Encountered error while trying to install package.
â°â> pycuda
note: This is an issue with the package mentioned above, not pip.
hint: See above for output from the failure.
check_install_deps.sh: 66: ERROR: installing python failed.
[cryosparc_user@xingcryoem2 cryosparc_user] cd cryosparc2_worker/
[cryosparc_user@xingcryoem2 cryosparc2_worker] bin/cryosparcw gpulist
Traceback (most recent call last):
File “”, line 1, in
File “/mnt/ssd/cryosparc_user/cryosparc2_worker/bin/connect.py”, line 23, in print_gpu_list
import pycuda.driver as cudrv
ModuleNotFoundError: No module named ‘pycuda’
If your CryoSPARC instance is currently not functional due to a broken worker configuration, I recommend:
cryosparcm update (should update the master to CryoSPARC v4.2, released yesterday).
identify or create a CUDA 11 toolkit installation that is independent from system updates, redefined symbolic links or alternatives. For the following steps, let’s assume such an installation exists at /usr/local/cuda-11.5.
move your existing cryosparc_worker installation “out of the way”: mv cryosparc_worker cryosparc_worker_obsolete_20230228
if a new cryosparc_worker package wasn’t downloaded as part of cryosparcm update, download it.
unpack the cryosparc_worker archive such that it occupies the original path of the cryosparc_worker directory that you just moved.
run
cd cryosparc_worker
./install.sh --cudapath /usr/local/cuda-11.5 --license "<your-license-id>"
In order to run processing jobs, you will need to connect this
worker to a cryoSPARC master.
But how do I connect the worker to the master now?
[cryosparc_user@xingcryoem2 cryosparc2_worker]$ ./bin/cryosparcw connect --worker xingcryoem2.oncology.wisc.edu --master xingcryoem2.oncology.wisc.edu --port 39000
Attempting to register worker xingcryoem2.oncology.wisc.edu to command xingcryoem2.oncology.wisc.edu:39002
Connecting as unix user cryosparc_user
Will register using ssh string: cryosparc_user@xingcryoem2.oncology.wisc.edu
If this is incorrect, you should re-run this command with the flag --sshstr
Traceback (most recent call last):
File “bin/connect.py”, line 76, in
assert cli.test_connection(), “Could not connect to cryoSPARC master at %s:%d” % (master_hostname, command_core_port)
File “/mnt/ssd/cryosparc_user/cryosparc2_worker/cryosparc_tools/cryosparc/command.py”, line 112, in func
assert “error” not in res, f’Error for “{key}” with params {params}:\n’ + format_server_error(res[“error”])
AssertionError: Error for “test_connection” with params ():
ServerError: Authentication failed - License-ID mismatch.
Please ensure cryosparc_master/config.sh and cryoparc_worker/config.sh have the same CRYOSPARC_LICENSE_ID entry
or CRYOSPARC_LICENSE_ID is set correctly in the current environment.
See CryoSPARC Architecture and System Requirements - CryoSPARC Guide for more details.
Attempting to register worker xingcryoem2.oncology.wisc.edu to command xingcryoem2.oncology.wisc.edu:39002
Connecting as unix user cryosparc_user
Will register using ssh string: cryosparc_user@xingcryoem2.oncology.wisc.edu
If this is incorrect, you should re-run this command with the flag --sshstr
Connected to master.
Current connected workers:
xingcryoem2.oncology.wisc.edu
Autodetecting available GPUs…
Detected 4 CUDA devices.
id pci-bus name
0 0000:1B:00.0 NVIDIA GeForce RTX 2080 Ti
1 0000:3E:00.0 NVIDIA GeForce RTX 2080 Ti
2 0000:88:00.0 NVIDIA GeForce RTX 2080 Ti
3 0000:B1:00.0 NVIDIA GeForce RTX 2080 Ti
All devices will be enabled now.
This can be changed later using --update
Traceback (most recent call last):
File “bin/connect.py”, line 225, in
assert args.ssdpath is not None or args.nossd, “Either provide --ssdpath or --nossd”
AssertionError: Either provide --ssdpath or --nossd