I have a Slurm cluster without GPUs. I’ve installed both master and worker on my headnode, as it is equal to my compute nodes. Now I’m trying to log into host:39000, and I’m only able to see a progress bar very briefly, then I’m presented with a light blue page, with nothing more. Wasn’t it supposed to be a login page?
By inspecting cryosparcm log command_core, I see this message being flooded every 5 seconds:
[JSONRPC ERROR 2020-04-29 08:38:32.638243 at get_config_var ]
-----------------------------------------------------
Traceback (most recent call last):
File "cryosparc2_command/command_core/__init__.py", line 114, in wrapper
res = func(*args, **kwargs)
File "cryosparc2_command/command_core/__init__.py", line 438, in get_config_var
assert res is not None, "Variable not found."
AssertionError: Variable not found.
-----------------------------------------------------
[JSONRPC ERROR 2020-04-29 08:38:32.638319 at get_running_version ]
-----------------------------------------------------
Traceback (most recent call last):
File "cryosparc2_command/command_core/__init__.py", line 114, in wrapper
res = func(*args, **kwargs)
File "cryosparc2_command/command_core/__init__.py", line 397, in get_running_version
return get_config_var('running_version')
File "cryosparc2_command/command_core/__init__.py", line 123, in wrapper
raise e
AssertionError: Variable not found.
-----------------------------------------------------
[JSONRPC ERROR 2020-04-29 08:38:32.638357 at get_system_info ]
-----------------------------------------------------
Traceback (most recent call last):
File "cryosparc2_command/command_core/__init__.py", line 114, in wrapper
res = func(*args, **kwargs)
File "cryosparc2_command/command_core/__init__.py", line 417, in get_system_info
'version' : get_running_version(),
File "cryosparc2_command/command_core/__init__.py", line 123, in wrapper
raise e
AssertionError: Variable not found.
-----------------------------------------------------
To install the master, I’ve followed this instructions: https://cryosparc.com/docs/reference/install#master
And to install the worker, I’ve followed this: https://cryosparc.com/docs/reference/install#cluster, but changed the --cudapath <cuda_path> to --nogpu. At first, the worker installation gave me an error when compiling some package that was including cuda.h. A second try returned with success, don’t know why.
Kind regards,
Ian