You are using Torque and not OpenPBS Pro. Hence, the command line arguments to qstat / qsub would differ.
qstat --version
Version: 6.1.1.1
Commit: 22f28343b8ee83b1234479b20224353f6c2db317
qstat -Bf
Server: node01
server_state = Active
scheduling = True
total_jobs = 4
state_count = Transit:0 Queued:0 Held:0 Waiting:0 Running:4 Exiting:0 Comp
lete:0
acl_hosts = node01
managers = haichao@node01
operators = haichao@node01
default_queue = batch
log_events = 2047
mail_from = adm
resources_assigned.nodect = 6
node_check_rate = 150
tcp_timeout = 300
job_stat_rate = 300
poll_jobs = True
down_on_error = True
mom_job_sync = True
pbs_version = 6.1.1.1
keep_completed = 300
next_job_number = 875
net_counter = 0 0 0
moab_array_compatible = True
nppcu = 1
timeout_for_job_delete = 120
timeout_for_job_requeue = 120
note_append_on_error = True
Thanks for the history, could you help me to modify my scripts?
The suggestion made above are for OpenPBS and would not work for the work load manager that you have on your system. You would need to refer the admin/user manual for the work load manager that is deployed on your system or you would need to switch to OpenPBS if there is an opportunity.
Thanks for your suggestions! I will refer the admin/user manual of torque.
For this matter, I need to make a correction: when using the -koe
parameter, I can find the standard output and error files, but they are located in the home
directory instead of the directory where I submitted the script to run.
I think you should work on the first issue you found: why your output ends up in undelivered. Most likely, something is wrong with your ssh setup. What happens if you login to node03 and run
pwd
ssh node01 pwd
If you get errors from these, you need to fix them and then see if the original script now works.