Hi all,
I’m posting this here as I was not sure if there is a custodian-related group. I was running some high-throughput defect calculations on Mendel, and whereas some jobs completed without any errors, others stopped after a few ionic steps with the following error in FW_job.error,
mpirun: killing job…
ERROR:custodian.custodian:{u’handler’: VaspErrorHandler, u’errors’: [u’eddrmm’], u’actions’: [{u’action’: {u’_set’: {u’ALGO’: u’Normal’}}, u’dict’: u’INCAR’}, {u’action’: {u’_file_delete’: {u’mode’: u’actual’}}, u’file’: u’CHGCAR’}, {u’action’: {u’_file_delete’: {u’mode’: u’actual’}}, u’file’: u’WAVECAR’}]}
ERROR:custodian.custodian:MaxErrors
Traceback (most recent call last):
File “/global/u1/s/sbajaj/sb_vasp/codes/fireworks/fireworks/core/rocket.py”, line 202, in run
m_action = t.run_task(my_spec)
File “/global/u1/s/sbajaj/sb_vasp/codes/MPWorks/mpworks/examples/firetasks_ex.py”, line 59, in run_task
c.run()
File “/global/u1/s/sbajaj/sb_vasp/codes/custodian/custodian/custodian.py”, line 221, in run
.format(self.total_errors, ex))
RuntimeError: 1 errors reached: MaxErrors. Exited…
INFO:rocket.launcher:Rocket finished
Also, in the output file vasp.out, I noticed the following warning,
WARNING in EDDRMM: call to ZHEGV failed, returncode = 6 3 11
I am using ALGO = Normal, and tried using both vasp/5.2 and vasp/5.3.3_vtst.matgen with the same issue. Any suggestions as to what might be causing this?
Thanks
Saurabh