hdf5 file is not be saved

[NosCC]

When I submit the job, I simply drag a .py file into the job manager. During this process, it causes error like this

"Traceback (most recent call last):
  File "zipdir/NL/GUI/Tools/Tool.py", line 915, in dropEvent
  File "zipdir/NL/GUI/Tools/JobManager/JobManager.py", line 772, in addObjects
  File "zipdir/NL/GUI/Tools/JobManager/JobManager.py", line 886, in addScripts
  File "build/lib/python3.8/contextlib.py", line 120, in __exit__
  File "zipdir/NL/GUI/Tools/JobManager/JobModel.py", line 103, in batching
  File "zipdir/NL/GUI/Tools/JobManager/JobManager.py", line 562, in saveSettings
  File "zipdir/NL/IO/NLSaveUtilities.py", line 453, in nlsave
  File "zipdir/NL/IO/HDF5.py", line 409, in __init__
  File "zipdir/NL/IO/HDF5.py", line 186, in __init__
  File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
  File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
  File "build/lib/python3.8/site-packages/h5py/_hl/group.py", line 167, in __getitem__
  File "h5py/_objects.pyx", line 54, in h5py._objects.with_phil.wrapper
  File "h5py/_objects.pyx", line 55, in h5py._objects.with_phil.wrapper
  File "h5py/h5o.pyx", line 190, in h5py.h5o.open
KeyError: 'Unable to open object (addr overflow, addr = 31749486, size=168, eoa=30745876)'
 ".

Even though the job can still be run in the cluster, the job manager of QuantumATK does not save the .py file as a history, so the .hdf5 file after finishing calculation will not be saved into the software. How can I solve this issue?

I tried to delete the .hdf5 in .vnl folder as suggested in https://forum.quantumatk.com/index.php?topic=6437.msg27030#msg27030. Unfortunately, it does not work. Note that I use version 2021.06

[Anders Blom]

I suggest resetting things and start fresh.
1) Close NanoLab
2) Rename the folder .vnl to something else (don't delete it, just rename it)
3) Start NanoLab and try again

If that works, you are good.

Most likely you will not lose anything by this, except machine settings and old jobs from the Job Manager but that is where the problem seems to be located (I suspect maybe a file permission error) so it's unavoidable. But if you do have some special settings or whatever, you can refer to the backup folder and compare and copy things over to the new .vnl folder that has been created.

[NosCC]

Thank you so much. From your suggestion, it works now and the error has disappeared!

Just to better understand, could you tell me how this issue comes from so that I can be aware of this issue?

[Anders Blom]

I can only guess some file got corrupted, maybe a crash during an exit operation or similar. It is not common so hopefully you should never see this error again