namelist read error for CAM branch run

[brianpm]

Hi everybody,

I am trying to do a branch run using CAM and I keep getting an error saying:

...
READ_NAMELIST: Namelist read returns 88
...
Traceback:
Offset 0x00000130 in procedure __abortutils_NMOD_endrun, near line 45 in file /home/bluevista/brianpm/cam1/models/atm/cam/src/control/abortutils.F90
Offset 0x00008bf4 in procedure __runtime_opts_NMOD_read_namelist, near line 1071 in file /home/bluevista/brianpm/cam1/models/atm/cam/src/control/runtime_opts.F90
Offset 0x00000010 in procedure __runtime_opts_NMOD_runtime_options, near line 2167 in file /home/bluevista/brianpm/cam1/models/atm/cam/src/control/runtime_opts.F90
Offset 0x000004e4 in procedure cam, near line 181 in file /home/bluevista/brianpm/cam1/models/atm/cam/src/control/cam.F90
--- End of call chain ---

I'm confused by this because I think I've followed the instructions for doing a branch run correctly.

In the end, I tried not modifying anything, not nelapse, not the histories, not anything. The only change was that I inserted "nsrevsn" in my namelist, for both cam and clm, both are paths on the NCAR MSS and come from a previous CAM run (which I can run as a restart). I am supplying paths for files that look like caseid.cam2.r.yyy-mm-dd-00000 and caseid.clm2.r.yyy-mm-dd-00000

At this point my namelist is exactly the same between my restart run (which works) and my branch run (which compiles but won't run) except for the nsrevsn lines.

Any ideas?

Thanks, .brian

 

[rneale]

Have you included the namelist variable BRNCH_RETAIN_CASENAME?

 

[brianpm]

No, I didn't. Actually, I lied in the above post. I also changed the case name (in my run script and my namelist), so the other line of my namelist that is different is the caseid.

Here is the output of a diff between my restart namelist and my branch namelist:

9c9
< caseid = 'vista_cam_t42_branch001'
---
> caseid = 'vista_cam_t42'
28d27
< nsrevsn = '/ptmp/brianpm/vista_cam_t42_branch001/vista_cam_t42.cam2.r.003-06-01-00000'
29a29,30
> start_tod = 0
> start_ymd = 000101
35d35
< nrevsn = '/ptmp/brianpm/vista_cam_t42_branch001/vista_cam_t42.clm2.r.0003-06-01-00000'

I have tried both with and without the "start_tod" and "start_ymd" lines with the same result. And here I actually copied the restart files from the MSS to my ptmp directory, just in case that was the problem. (it isn't)


I think I must be missing something obvious somewhere, but I can't figure it out.

 

[brianpm]

FWIW, I tried putting brnch_retain_casename = .false. (changing from casename vista_cam_t42 to vista_cam_t42_branch001) and I still got that same error.

Frustrated, I started trying lots more things... only to realize that my namelist had "nsrevsn = ... "

No, it is supposed to be "nrevsn = ..."

That, combined with a proper path for the restart files, and whatever brnch_retain_casename = .false. provides (I'm pretty sure it is optional, since the default is false), actually gets a branch case going properly.

The trouble was that there was a clear, obvious, easily diagnosable problem in my namelist, but instead of the output of build-namelist simply printing out something like "nsrevsn is not a recognized namelist variable," or "no nrevsn path supplied," it told me namelist read returns 88, and 88 does not seem to mean anything. That said, I still should have found that typo faster than I did.