Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

first problem on Configure;

Hi, My new machine is a "generic_linux" which is listed as "supported" in the CCSM create_newcase, so I though I got the newcase established, I have the 3 files: env_conf, env_run, and env_mach.$MACH. So far I followed exactly what's said in the README , so then I tried the following:
> ./configure -mach $MACH
as said, I got the 3 sub-directories (Buildexe, Buildlib, Buildnml_Prestage), and "one" script "$CASE.$MACH.build" ! That is, I did NOT get the other 2 scripts: "$CASE.$MACH.run, and $CASE.$MACH.l_archive, WHY?
In the mean time I got error messages:
".../env_mach.anchorage: No such file or directory.
configure generate_batch error"
(the file "env_mach.anchorage"??, what is it?, I have NO idea about it)

It looks like because of this error, the script $CASE.$MACH.run is NOT generated, and I can NOT go any further from here!! BUT I have exactly followed what's said in the README, so what can I do now? Please get back to me when you can, thanks.
Pontzsong 3/15/05
 

njn01

Member
It looks as if you've discovered a bug in the ccsm3 scripts.

I suggest you try the following:

1) In your downloaded version of ccsm3,
cd csm3/scripts/ccsm_utils/Machines

2) use your editor to edit the script batch.linux.generic_linux
change the line that says,
set mach = anchorage
to say instead,
set mach = generic_linux

3) remove the directory created by your first attempt at "create_newcase"

4) re-issue your create_newcase command

5) reissue the configure -mach command

This should at least get you past the problem you have encountered.

-- Nancy
 
Thanks very much Nancy, following your direction I re-created my $CASE and got the both scripts (the "build" and the "run"). Now, from README the next step is to build, I typed in ./$CASE.$MACH.build, it apparently did NOT work. Let me attach the screen-error-messages here:

(1) message within the $CASE dir:
master:/rstor/asong/CCSM/ccsm3_0/scripts/Test2> ./Test2.generic_linux.build
## Warning: SCRATCH not defined in system environment. Set SCRATCH to be /rstor/asong
-------------------------------------------------------------------------
Preparing T42_gx1v3 component models for execution
-------------------------------------------------------------------------
- Create execution directories for atm,cpl,lnd,ice,ocn
mkdir: cannot create directory `/ptmp': Permission denied
cannot make /ptmp/asong/archive/Test2: Command not found.
- If a restart run then copy restart files into executable directory
ccsm_getrestart: get /rstor/asong/Test2 restarts from /ptmp/asong/archive/Test2/restart
- Check validity of configuration
- Determine if build must happen (env variable BLDTYPE)
- Build flag (BLDTYPE) is TRUE
- Build Libraries: esmf, mph, mct
Fri Mar 18 09:12:06 CST 2005 esmf.buildlib.050318-091206
Fri Mar 18 09:12:11 CST 2005 mph.buildlib.050318-091206
ERROR: mph.buildlib failed, see mph.buildlib.050318-091206
ERROR: cat /rstor/asong/Test2/mph/mph.buildlib.050318-091206
master:/rstor/asong/CCSM/ccsm3_0/scripts/Test2> vi msg_build

(2) message within the Scratch dir:
Fri Mar 18 09:12:11 CST 2005 mph.buildlib.050318-091206
## Warning: SCRATCH not defined in system environment. Set SCRATCH to be /rstor/asong
cp: cannot stat `mph_conf': No such file or directory
diff: mph_conf.old: No such file or directory
mpif90 -I. -I -I/rstor/asong/Test2/lib/include -I -c -r8 -i4 -Kieee -Mrecursive -Mdalign -Mextend -DLINUX -DPGF90 -DNO_SHR_VMATH -DLINUX mph.F
PGF90-S-0012-Value required for command line switch: -idir
PGF90-S-0011-Unrecognized command line switch: -I/rstor/asong/Test2/lib/include
PGF90-S-0012-Value required for command line switch: -idir
PGF90/any Linux/x86 5.0-2: compilation completed with severe errors
gmake: *** [mph.o] Error 2


There seem 2 clues: one about "ptmp not be created", and the other about the "mph". Both are clearly out of my ability to try to find out what's wrong. Please get back to me when you can.

Thanks in advance,

Aaron 3/18/05
 
pontzsong said:
(2) message within the Scratch dir:
Fri Mar 18 09:12:11 CST 2005 mph.buildlib.050318-091206
## Warning: SCRATCH not defined in system environment. Set SCRATCH to be /rstor/asong
cp: cannot stat `mph_conf': No such file or directory
diff: mph_conf.old: No such file or directory
mpif90 -I. -I -I/rstor/asong/Test2/lib/include -I -c -r8 -i4 -Kieee -Mrecursive -Mdalign -Mextend -DLINUX -DPGF90 -DNO_SHR_VMATH -DLINUX mph.F
PGF90-S-0012-Value required for command line switch: -idir
PGF90-S-0011-Unrecognized command line switch: -I/rstor/asong/Test2/lib/include
PGF90-S-0012-Value required for command line switch: -idir
PGF90/any Linux/x86 5.0-2: compilation completed with severe errors
gmake: *** [mph.o] Error 2

There seem 2 clues: one about "ptmp not be created", and the other about the "mph".

Although rather long time has passed since this post, and its author hopefully resolved these problems, I would like to write my comments about them, in order other users could get some extra help.

The solution of "ptmp" problem is simple and obvious. One needs to edit the env_mach.xxx file and enter new paths. One needs to look at lines 89-92, 114-124 and around. This is also mentioned in the model documentation.

The solution of the "mph" problem is not such obvious.
When installing the model on a new computer, I was recalling of what I did previously, for a considerable time.

In order this error doesn't occur, the environment variables
INC_MPI, INC_NETCDF, LIB_MPI, and LIB_NETCDF
must be set pointing to the locations of the respective libraries.

As far as I remember, this is mentioned only in (a) compilation script(s), not in the documentation, and the MPI and NetCDF installation scripts don't set these variables.

The compiler gets empty -I switch, if these variables aren't set, and the PGI compiler doesn't like it. These switches are highlighted in the quote above.
 
Top