I configured a new remote host and a new code. The computer test was successful and the code (pw.x) runs well on the remote machine. I am also able to retrieve quantum espresso output files with the command verdi calcjob outputcat <pk>. However, the calculation fails with error message [322]. When I type (aiida) [user]$ verdi process report 285755, I get:
*** 285755: None
*** (empty scheduler output file)
*** Scheduler errors:
Warning: ieee_divide_by_zero is signaling
Warning: ieee_inexact is signaling
FORTRAN STOP
Warning: ieee_inexact is signaling
FORTRAN STOP
*** 3 LOG MESSAGES:
+-> WARNING at 2025-02-18 16:23:43.596624+01:00
| key 'symmetries' is not present in raw output dictionary
+-> ERROR at 2025-02-18 16:23:43.634840+01:00
| The XML output file has an unsupported format.
+-> WARNING at 2025-02-18 16:23:43.638372+01:00
| output parser returned exit code<322>: The XML output file has an unsupported format.
I am working on Alma Linux, am using aiida-core 2.4.0, aiida-quantumespresso 4.4.0 and aiida-vibroscopy 1.1.1
I configured the pw.x code with the following yml file entrances:
---
label: xx
description: xx
default_calc_job_plugin: quantumespresso.pw
computer: xx
filepath_executable: /xx..xx/pw.x
prepend_text: |+
xx...
append_text: " "
Thank you @t-reents, I updated the aiida-quantumespresso plugin to v4.9.0. I am now running the same workflow than previously, but get a MissingEntryPointError: Entry point 'quantumespresso.hubbard_structure' not found in group 'aiida.data' error message which I did not get before. What is this due to and how can I solve it?
We already did some further debugging locally. She installed it from PyPI using pip install aiida-quantuemespresso --upgrade (she also did pip install aiida-phonopy --upgrade, but this shouldn’t cause any problems, at least at first glance).
The issue is indeed that quantumespresso.hubbard_structure also doesn’t appear when running verdi plugin list aiida.data, i.e., the entry point is indeed missing. A few things that we’ve tried/checked:
Since she is using aiida-core 2.4, explicitly restarting/resetting the daemon didn’t change anything
According to our compatibility matrix, aiida-quantumespresso 4.9 requires aiida-core 2.5, due to the changes related to pydantic. However, I wouldn’t really expect that this is the issue. Moreover, I locally reproduced her environment, i.e., the versions of aiida-qe, aiida-core, aiida-phonopy and independent of the versions, I don’t encounter any entry point related error
Very strange! Not immediately sure what I’d try next then… I suppose indeed using aiida-core==2.4 is fine, in hindsight I think aiida-core didn’t have a pydantic dependency before v2.5, so there is no conflict (should probably correct the compatibility matrix again ).
One further question I’d have is what kind of environment it is? I’ve had some hard times in the past troubleshooting conda environments because some packages were installed with both pip and conda.
I’m using a conda environment (conda 23.3.1) and pip installation indeed
The easiest would probably be to downgrade the QE version, as well as the versions of the aiida plugins. This does not frontally solve the bug though. I’m available for a zoom call if you want more information
Update:
I downgraded the QE version to 7.3.1. After a few trial runs with different plugin versions, I found a combination that works well (no complains) with aiida-quantumespresso==4.3, aiida-phonopy==1.1.4, aiida-vibroscopy==1.1.1 and aiida-core=2.4.0.
).