in src/slurm_plugin/instance_manager.py [0:0]
def _parse_slurm_resume(self, slurm_resume: Dict[str, any]) -> List[SlurmResumeJob]:
slurm_resume_jobs = []
for job in slurm_resume.get("jobs", {}):
try:
slurm_resume_jobs.append(SlurmResumeJob(**job))
except InvalidNodenameError:
nodes_resume = job.get("nodes_resume", "")
nodes_alloc = job.get("nodes_alloc", "")
job_id = job.get("job_id", "")
logger.warning(
"Discarding NodeNames with invalid format for Job Id (%s): nodes_alloc (%s), nodes_resume (%s)",
job_id,
nodes_alloc,
nodes_resume,
)
self._update_failed_nodes(
# if NodeNames in nodes_resume cannot be parsed, try to get info directly from Slurm
set([node.name for node in get_nodes_info(nodes_resume)]),
"InvalidNodenameError",
)
return slurm_resume_jobs