error about'couldnt find input file'

Discussion about tangent linear and adjoint models, variational data assimilation, and other related issues.

Moderators: arango, robertson

Post Reply
Message
Author
WY0116
Posts: 8
Joined: Sat Oct 10, 2015 1:04 am
Location: Southampton

error about'couldnt find input file'

#1 Post by WY0116 » Thu Oct 22, 2015 8:50 am

hi everyone,
I am a new leaner of Roms.I run the oceanM after,and I got the problem as flow:

Output/Input Files:

Output Restart File: ./SCS_results/SCS_rst.nc
Prefix for History Files: ./SCS_results/SCS_his
Output Averages File: ./SCS_results/SCS_avg.nc
Input Grid File: ./SCS_forcing/SCS_BIG-grd.nc
Input Nonlinear Initial File: ./SCS_forcing/SCS_TIDE28-init.nc

READ_PHYPAR - could not find input file: ^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@

Elapsed CPU time (seconds):


ROMS/TOMS - Output NetCDF summary for Grid 01:

ROMS/TOMS - I/O error ............... exit_flag: 4


ERROR: I/O related problem.
Job /public/lsf/7.0/linux2.6-glibc2.3-x86_64/bin/openmpi_wrapper ./oceanM ./ocean_SCS.in



anyone has ever got the same problem?
Any suggestions will be appreciated.

User avatar
kate
Posts: 3718
Joined: Wed Jul 02, 2003 5:29 pm
Location: IMS/UAF, USA

Re: error about'couldnt find input file'

#2 Post by kate » Thu Oct 22, 2015 4:48 pm

The problem is in your ocean_SCS.in file. After listing the grid file, my job output lists the forcing files. What do you have for NFFILES? Did you keep the FRCNAME section? It's good to have NFFILES at least 1 and to have something for FRCNAME. There are times when ROMS wants a field from the forcing file even when you don't expect that and the error messages make more sense if you keep FRCNAME. Pointing to an existing file will trigger messages more like "can't file variable stflx in file myfile" (one hopes).

I hope you are trying to run the forward model before taking on the adjoint. Why is this message in the adjoint section?

WY0116
Posts: 8
Joined: Sat Oct 10, 2015 1:04 am
Location: Southampton

Re: error about'couldnt find input file'

#3 Post by WY0116 » Sat Oct 24, 2015 12:53 pm

kate wrote:The problem is in your ocean_SCS.in file. After listing the grid file, my job output lists the forcing files. What do you have for NFFILES? Did you keep the FRCNAME section? It's good to have NFFILES at least 1 and to have something for FRCNAME. There are times when ROMS wants a field from the forcing file even when you don't expect that and the error messages make more sense if you keep FRCNAME. Pointing to an existing file will trigger messages more like "can't file variable stflx in file myfile" (one hopes).

I hope you are trying to run the forward model before taking on the adjoint. Why is this message in the adjoint section?
hi kate,
thanks for your reply.
I have
NFFILES == 1
FRCNAME == ./SCS_forcing/SCS_TIDE28-forc.nc
as NFFILES.
I tried to run the case by the old version(2008) of Roms,and it could run well. But the new version(2015) still doesn't work.Could you give me some more suggestions? The attchment is my ocean_SCS.in file.

Thanks again for your help.
Attachments
ocean_SCS.in
(106.47 KiB) Downloaded 85 times

User avatar
kate
Posts: 3718
Joined: Wed Jul 02, 2003 5:29 pm
Location: IMS/UAF, USA

Re: error about'couldnt find input file'

#4 Post by kate » Sun Oct 25, 2015 4:31 pm

You have:

Code: Select all

     FRCNAME == ./SCS_forcing/SCS_TIDE28-forc.nc                      ! forcing file 1, grid 1 
Try getting rid of the comment after the filename.

Also, the date on your ocean.in file is 2013. You need to use a 2015 ocean.in with a 2015 ROMS. Your ocean.in has NBCFILES, which isn't in the trunk ROMS - are you using trunk ROMS?

WY0116
Posts: 8
Joined: Sat Oct 10, 2015 1:04 am
Location: Southampton

Re: error about'couldnt find input file'

#5 Post by WY0116 » Mon Oct 26, 2015 8:51 am

kate wrote:You have:

Code: Select all

     FRCNAME == ./SCS_forcing/SCS_TIDE28-forc.nc                     
Try getting rid of the comment after the filename.

Also, the date on your ocean.in file is 2013. You need to use a 2015 ocean.in with a 2015 ROMS. Your ocean.in has NBCFILES, which isn't in the trunk ROMS - are you using trunk ROMS?
Hi kate,
I deleted the comment '! forcing file 1, grid 1',and I I modified the newer ocean_SCS.in(the attachment).but it still doesn't work.
I got the code by 'svn checkout --username bruce https://www.myroms.org/svn/src/trunk src'. Is it a trunk version?

User avatar
kate
Posts: 3718
Joined: Wed Jul 02, 2003 5:29 pm
Location: IMS/UAF, USA

Re: error about'couldnt find input file'

#6 Post by kate » Mon Oct 26, 2015 4:55 pm

Yes, that is the trunk version of the code.

Are you still getting the same error? Do you have a file listed for each filename in an unmodified ocean.in or did you comment some out?

WY0116
Posts: 8
Joined: Sat Oct 10, 2015 1:04 am
Location: Southampton

Re: error about'couldnt find input file'

#7 Post by WY0116 » Tue Oct 27, 2015 12:44 am

kate wrote:Yes, that is the trunk version of the code.

Are you still getting the same error? Do you have a file listed for each filename in an unmodified ocean.in or did you comment some out?
Yes,the error is still the same.I am sorry I forgot to attach my new ocean.in file.I cant find the file which listed the each filename in the ocean.in file. And I am puzzled that I could run the application upwelling well.
Attachments
ocean_SCS.in
(117.5 KiB) Downloaded 87 times

WY0116
Posts: 8
Joined: Sat Oct 10, 2015 1:04 am
Location: Southampton

Re: error about'couldnt find input file'

#8 Post by WY0116 » Tue Oct 27, 2015 7:25 am

WY0116 wrote:
kate wrote:Yes, that is the trunk version of the code.

Are you still getting the same error? Do you have a file listed for each filename in an unmodified ocean.in or did you comment some out?
Yes,the error is still the same.I am sorry I forgot to attach my new ocean.in file.I cant find the file which listed the each filename in the ocean.in file. And I am puzzled that I could run the application upwelling well.
Hi kate,
I use the 'svn update -r 700'just now.I guess the version has been downgraded to 2013.And it could run without error(I didnt modify the ocean.in file).I dont know the difference between the 2015 version and the old versions.While the 2015 version still has the error.

User avatar
kate
Posts: 3718
Joined: Wed Jul 02, 2003 5:29 pm
Location: IMS/UAF, USA

Re: error about'couldnt find input file'

#9 Post by kate » Tue Oct 27, 2015 4:50 pm

I think your problem is this combo:

Code: Select all

  LtracerSrc == T T 

Code: Select all

!     SSFNAME == ocean_rivers.nc
You are saying that you want river sources for the tracers and yet you commented out the river filename.

WY0116
Posts: 8
Joined: Sat Oct 10, 2015 1:04 am
Location: Southampton

Re: error about'couldnt find input file'

#10 Post by WY0116 » Sun Nov 01, 2015 12:20 pm

kate wrote:I think your problem is this combo:

Code: Select all

  LtracerSrc == T T 

Code: Select all

!     SSFNAME == ocean_rivers.nc
You are saying that you want river sources for the tracers and yet you commented out the river filename.

Hi kate,
I tried to change the LtracerSrc= F F and the errors occured as follow:
CHECKDEFS - cannot use obsolete option: SPLINES
See following trac ticket for details:
https://www.myroms.org/projects/src/ticket/681
please change header file accordingly.

Elapsed CPU time (seconds):


ROMS/TOMS - Output NetCDF summary for Grid 01:

ROMS/TOMS - Configuration error ..... exit_flag: 5


ERROR: Illegal model configuration.

Then my fellow shared his Rom3.7 to me. While with the same files(.in file and .h file),it could work. It makes me confused.

User avatar
kate
Posts: 3718
Joined: Wed Jul 02, 2003 5:29 pm
Location: IMS/UAF, USA

Re: error about'couldnt find input file'

#11 Post by kate » Sun Nov 01, 2015 5:15 pm

Oh, yeah, SPLINES. That changed pretty recently. It has become three options:

Code: Select all

SPLINES_VVISC       use if splines reconstruction of vertical viscosity
SPLINES_VDIFF       use if splines reconstruction of vertical diffusion
RI_SPLINES          use if splines reconstruction for vertical sheer

Post Reply