Error on MN during initialization due to invalid default path

[alexbarcelo]

The current codebase, once deployed on Mare Nostrum, doesn't behave nicely on filesystem matters.

The traceback given is:

Traceback (most recent call last):
  File "/apps/PYTHON/3.6.4_ML/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/apps/PYTHON/3.6.4_ML/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/apps/DATACLAY/2.6/pyclay/src/dataclay/executionenv/server/__main__.py", line 30, in <module>
    run_main()
  File "/apps/DATACLAY/2.6/pyclay/src/dataclay/executionenv/server/__main__.py", line 26, in run_main
    exec_env_srv.start()
  File "/apps/DATACLAY/2.6/pyclay/src/dataclay/executionenv/server/ExecutionEnvironmentSrv.py", line 208, in start
    self.execution_environment = ExecutionEnvironment(settings.dataservice_name)
  File "/apps/DATACLAY/2.6/pyclay/src/dataclay/executionenv/ExecutionEnvironment.py", line 60, in __init__
    self.store_ee_info()
  File "/apps/DATACLAY/2.6/pyclay/src/dataclay/executionenv/ExecutionEnvironment.py", line 89, in store_ee_info
    fh = open(info_file, 'w')
OSError: [Errno 30] Read-only file system: '/dataclay/metadata//python_ee_dataservice1.info'

AFAICT, the /dataclay path is used by default in containerized environment and so on, but that won't work on Mare Nostrum or other cluster deployments. I don't know if singularity is involved in all that. I think so. I don't know if it has any impact.

The 2.5 deployment is currently working, but I cannot assure if there is some regression problem or there were some manual fine-tuning on the 2.5 deployment that have not been done on 2.6.

[dgasull]

Absolutely right, it was a bug in 2.6 version. Current version stores logicmodule metadata databases like classes, operations, contracts,... in a different path than object metadata and sessions databases. So, we have /dataclay/storage/ for object data and /dataclay/metadata for class-model data. This was done to add fault tolerance support in case we loose the storage database. This new path was not added in orchestration scripts, adding it now.

Fixed in bsc-dom/dataclay-orchestration@cbd6476 .Fix was deployed to MN.

@alexbarcelo can you retry it and confirm that it works for you?