Main menu

Navigation

errors with validating a cesm port

12 posts / 0 new
Last post
xgao304@...
errors with validating a cesm port

I am able to create_newcase with cesm2. However, when I try to validate the porting (http://esmci.github.io/cime/users_guide/porting-cime.html), I got errors in several steps.

1. xmllint --noout --schema /net/fs05/d1/xgao/cesm/cesm2/cime/config/xml_schemas/config_machines.xsd config_machines.xml

config_machines.xml validates

2.when I run scripts_regression_tests.py, I get the following error:

xgao@c083 tests]$ ./scripts_regression_tests.py
No handlers could be found for logger "CIME.XML.machines"
ERROR: Could not initialize machine object from /net/fs05/d1/xgao/cesm/cesm2/cime/config/cesm/machines/config_machines.xml or /home/xgao/.cime/config_machines.xml

3. Perform the cheyenne "prealpha" tests on my machine, I got so many failed cases (see the attached cs.status.20180626_200457_au3vht)

./create_test --xml-category prealpha --xml-machine cheyenne --xml-compiler intel --machine svante --compiler pgi

 

I have not carried out ensemble consistency tests yet.

I am wondering if I should clear all these errors first before I proceed with my case setup, build, and submit. Any information is appreciated.

Thanks,

Xiang

 

 

 

jedwards

In scripts_regression_tests.py it is not recognizing your machine based on the NODENAME_REGEX field you provided.  You can correct that field or use the --machine option

to scripts_regression_tests.py to solve that problem. 

The file that you provided does not contain output from cs.status - please try again. 

CESM Software Engineer

xgao304@...

There are so many files from the new scripts_regression_tests.py (see the attached list of output files). Which file should I send? Thanks.

xgao304@...

And I rerun "./create_test --xml-category prealpha --xml-machine cheyenne --xml-compiler intel --machine svante --compiler pgi", but the resulting cs.status file is similar to previous one - nothing is output expect for a single line

#! /bin/bash

/net/fs05/d1/xgao/cesm/cesm2/cime/scripts/Tools/cs.status "$@" /net/fs05/d1/xgao/cesm/cases/*.20180703_134836_asmyzb/TestStatus

xgao304@...

I folloewed your instruction to rerun the cripts_regression_tests.py, here is the summary of the results:

Ran 129 tests in 3171.120s

FAILED (failures=9, errors=1, skipped=11)

There are a totla of 12 cs.status.* files in the test root. I attached the results of executing these cs.status.* files (cs_status.txt).

Please let me know if the porting of cesm on our machine is successful.

Attachment: 
jedwards

Please run scripts_regression_tests.py --machine svante 1> srt.txt 2>&1 and send the resulting srt.txt file.

CESM Software Engineer

xgao304@...

Here is the resulting srt.txt file. Thanks.

Attachment: 
jedwards

You do not have inputdata configured correctly.    The relevenet error message is:

 Client protocol gftp not enabled
        
        Could not connect to svn repo 'https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata'
        This is most likely either a credential, proxy, or network issue .
        To check connection and store your credential run 'svn ls https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata' and permanently store your password
         

CESM Software Engineer

xgao304@...

I am a bit confused. If I run "svn ls" as follows, I am not asked to input password or userid. Does it mean I already have my password permanently stored? The only thing is this is done in the head node. But my program is run on cluster node. However, if I run "svan ls .." in the cluter node, I will get error message. But I assume once the password is stored, it should be transparent across all the cluter system?

[xgao@svante-login ~]$ svn ls https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata
MaxTSIInput
atm/
ccsm4_init/
cesm2_init/
cpl/
dx7/
glc/
hiresFiles
ice/
import/
lnd/
ocn/
rof/
share/
template.readme
validation/
wav/

Thanks.

jedwards

Try running:

create_test SMS.f19_g16.X.svante_pgi 

If it fails try running ./check_input_data --download from the login node - does that work?

CESM Software Engineer

xgao304@...

I am running it.

I know ./check_input_data works from the login node because I use that to download the data for my test case (compset I2000Clm50SpGs and res hcru_hcru). I am able to run for three months, then got the error message (see my another post in "general discussion") - need your help with that post as well

Thanks,

 

xgao304@...

The attaed is the resulting file by running "create_test SMS.f19_g16.X.svante_pgi". I don't want to download the input data for this case because I know "./check_input_data"

is working fine for my other test case. Please let me know if I should do other things to validate the porting.

 

Thanks.

Attachment: 
Log in or register to post comments

Who's new

  • kamal.tewari1@...
  • rchemke
  • abdulla.sakalli@...
  • mehmetugurgucel@...
  • borst