Puma2 submit-failed

Hi CMS

I have a new puma2 account and i have worked through the setup docs.
i think i can ssh between puma2 and archer2 and svn to mosrs

however, when i try to run lfric_atm tests with cylc8 rose stem & cylc play
my run fails,
i’m getting

state sumbit-failed from EXPORT-SOURCE remote-init_ex

when i try to run a job from puma2

i can’t see any evidence that this has communicated with archer2, there is no
~/cylc-run
in my archer2 homespace

there’s almost no log information at
/home/n02/n02/marqh/cylc-run/lfric_apps_r4036/run1

my colleagues have run these lfric_atm tests, so i think the problem is with me

I’m not sure where to look to try and progress this.
are you able to help me?
thank you
mark

Hi Mark

Please allow us read permissions on your ARCHER2 home and work spaces, and on your puma2 space.

Grenville

hi Grenville
I think I’ve done this
m

Mark

/home/n02/n02/marqh/cylc-run/lfric_apps_r4036/run1/log/scheduler/01-start-01.log

says

2024-10-23T10:34:42Z WARNING - platform: ln03 - Could not connect to ln03.
...
2024-10-23T10:34:43Z WARNING - platform: ln04 - Could not connect to ln04.
  ...
2024-10-23T10:34:44Z WARNING - platform: ln02 - Could not connect to ln02.
  ...
2024-10-23T10:34:45Z WARNING - platform: ln01 - Could not connect to ln01.
 

which sounds like a setup problem. What happens if on puma2, you type

rose host-select archer2

Grenville

[WARN] ln04: ssh failed
[WARN] ln03: ssh failed
[WARN] ln01: ssh failed
[WARN] ln02: ssh failed
[FAIL] No hosts selected.

okay, i’ve tried to update some settings
this was working previously but perhaps i broke something
i can now

$ rose host-select archer2
ln02

but, with a new run:

/home/n02/n02/marqh/cylc-run/lfric_apps_r4036/run2/log/scheduler/01-start-01.log

i get

    STDERR:
        cylc: line 0: exec: cylc: not found

Mark

create a .bash_profile on ARCHER2 with


. /work/y07/shared/umshared/bin/rose-um-env-puma2

Grenville

thanks Grenville

i hadn’t seen that instruction in the docs, so hadn’t implemented it

I also had some issues with ssh access, which i’ve got to the bottom of now, I think

I appear to have a test run running properly now
thank you for your help
mark