def hostnames(output_dir, job_id, verbose):
"""Show the hostnames of active nodes participating in the batch."""
level = logging.DEBUG if verbose else logging.INFO
setup_logging(__name__, None, console_level=level)
try:
cluster, _ = Cluster.deserialize(output_dir, deserialize_jobs=True)
except InvalidConfiguration:
print(
f"{output_dir} is not a JADE output directory used in cluster mode",
file=sys.stderr)
sys.exit(1)
if cluster.is_complete():
print("All jobs are already complete.")
sys.exit()
groups = make_submission_group_lookup(
[cluster.config.submission_groups[0]])
hpc_mgr = HpcManager(groups, output_dir)
nodes = []
for _job_id in cluster.job_status.hpc_job_ids:
if job_id is not None and _job_id != job_id:
continue
nodes += hpc_mgr.list_active_nodes(_job_id)
if not nodes:
print("No nodes were detected.", file=sys.stderr)
sys.exit(1)
print(" ".join(nodes))
def manager_node(output_dir, job_id, verbose):
"""Print the name of the manager node to the console. Requires a single job in the batch."""
level = logging.DEBUG if verbose else logging.INFO
setup_logging(__name__, None, console_level=level)
try:
cluster, _ = Cluster.deserialize(output_dir, deserialize_jobs=True)
except InvalidConfiguration:
print(
f"{output_dir} is not a JADE output directory used in cluster mode",
file=sys.stderr)
sys.exit(1)
if cluster.is_complete():
print("All jobs are already complete.")
sys.exit()
groups = make_submission_group_lookup(
[cluster.config.submission_groups[0]])
hpc_mgr = HpcManager(groups, output_dir)
if job_id not in cluster.job_status.hpc_job_ids:
print(f"job_id={job_id} is not active", file=sys.stderr)
sys.exit(1)
node = hpc_mgr.list_active_nodes(job_id)[0]
print(node)