Opened 3 years ago

Closed 3 years ago

#2074 closed help (fixed)

Using MOOSE archiving

Reported by: nfreychet Owned by: ros
Component: UM Model Keywords:
Cc: Platform: MONSooN
UM Version: 8.5

Description

Hello,

I'm trying to save output from a run using MOOSE. I switch the Archiving system to MOOSE, and added the name of my project (lotus)
The path to the archiving script is: $UMDIR/archiving/bin
And the name is: um_archiving
And I don't delete any superseded file for now.
I just run for 1 day.

I have all the expected output on MONSooN (/projects/lotus/nifre/um/xnbrb/).
However when I connect on mass from JASMIN, and check the directory of my project (moo ls moose:/adhoc/projects/lotus/), it's empty.

Change History (10)

comment:1 Changed 3 years ago by ros

Hi Nicolas,

You are only running for 1 day so there will be nothing to archive at the end of it. You will need to run for much longer. For example; your job is set up to archive restart dumps every month and your PP files are re-initialised either every 30 or 90 days.

Regards,
Ros.

comment:2 Changed 3 years ago by nfreychet

Hi Ros,

Oh OK, I thought it would archive any output that are not explicitly deleted.
I'm gonna try with a longer run then.

Thanks,
Nicolas

comment:3 Changed 3 years ago by nfreychet

Hi again,

So I did 1 month run, but now I have an error in the archive.leave :

Reading archiving commands from file /projects/lotus/nifre/um/xnbrb/archive_171956.do
mkset command:  ['moo mkset -v moose:crum/xnbrb']
Mkset: system error, return code 255
stdout: Unable to load Moose credentials. Please check credential file permissions. Only file owner should have read and write access
; stderr: ls: cannot access /home/nifre/.moosedir/moose: No such file or directory

TMPDIR=JOBTEMP= /scratch/jtmp/pbs.2251763.xcm00.x8z
The command to archive is : moo put -f -vv /projects/lotus/nifre/um/xnbrb/xnbrba.da19811001_00 moose:crum/xnbrb/ada.file
Error, return code = 255
stdout: Unable to load Moose credentials. Please check credential file permissions. Only file owner should have read and write access
; stderr: ls: cannot access /home/nifre/.moosedir/moose: No such file or directory
Last edited 3 years ago by ros (previous) (diff)

comment:4 Changed 3 years ago by ros

  • Owner changed from um_support to ros
  • Status changed from new to accepted

Hi Nicolas,

In order to archive files to MASS you have to have been granted access by the Met Office at which point you will be sent some instructions along with some moose credential files which you need to put in your ~/.moosedir. Have you ever been sent these files? If not please contact the MONSooN team on: monsoon@….

Regards,
Ros.

comment:5 Changed 3 years ago by nfreychet

Hi Ros,

Yes I did apply for a MASS access, but it was for JASMIN (and I think I set it up on JASMIN properly).
I cannot fidn any ~/.moosedir on MONSooN so I suppose I missed a step. I will contact the monsoon helpdesk then.

Thanks,
Nicolas

comment:6 Changed 3 years ago by nfreychet

Hi Ros,

So after solving the problem of moose installation on MONSooN, I still have an error when archiving data, but it seems to be related to the parameters now:


ILP!!! fileAndPath /projects/lotus/nifre/um/xnbrb/*.do
ILP!!! files ['/projects/lotus/nifre/um/xnbrb/archive_130733.do', '/projects/lotus/nifre/um/xnbrb/archive_171956.do']
Reading archiving commands from file /projects/lotus/nifre/um/xnbrb/archive_130733.do
mkset command: moo mkset -v moose:crum/xnbrb?
Mkset: system error, return code 2
stdout: ; stderr: mkset command-id=325917608 failed: (SSC_TASK_REJECTION) one or more tasks are rejected.

moose:/crum/xnbrb: (TSSC_PROJECT_NAME_REQUIRED) A project name must be specified.

mkset: failed (2)

TMPDIR=JOBTEMP= /scratch/jtmp/pbs.2268286.xcm00.x8z
The command to archive is : moo put -f -vv /projects/lotus/nifre/um/xnbrb/xnbrba.da19811001_00 moose:crum/xnbrb/ada.file
Moose error: user-error (see Moose docs). Return code = 2
stdout: ; stderr: put command-id=325917615 failed: (SSC_TASK_REJECTION) one or more tasks are rejected.

/projects/lotus/nifre/um/xnbrb/xnbrba.da19811001_00 → moose:/crum/xnbrb/ada.file/xnbrba.da19811001_00: (TSSC_SET_DOES_NOT_EXIST) no such data set.

put: failed (2)


As I understand, the command moo mkset needs a project name. But in the UM configuration, I already put a project name for mass archiving (Post Processing >> Main Switch >> Monsoon project group name: lotus)
Do I have to specify it somewhere else?

Nicolas

comment:7 Changed 3 years ago by nfreychet

(I think I found the answer in this ticket: http://cms.ncas.ac.uk/ticket/1930 )

But so, each time I have to run a new job I need to manually create the directories first or is there a way to do it automatically from the UM configuration?

Cheers,
Nicolas

comment:8 Changed 3 years ago by ros

Hi Nicolas,

Yes you have to manually run the mkset command before running a new job. It was done this way on MONSooN to make it a conscious decision to archive, and thus stop people from archiving to MASS unintentionally.

Cheers,
Ros.

comment:9 Changed 3 years ago by nfreychet

Hi Ros,

Is there a directory on MONSooN for the input data for nudging (ERA Interim…)?

And same question for the fcm branches for UM8.5 (something like fcm:um_br/dev/ptelford/vn7.3_nudge_updated/src and fcm:um_br/dev/mdalvi/VN7.3_GregorianCal_fixes/src )

Nicolas

comment:10 Changed 3 years ago by ros

  • Resolution set to fixed
  • Status changed from accepted to closed

See duplicate ticket #2080

Note: See TracTickets for help on using tickets.