Hello Susan,

> On 26. Mar 2019, at 18:00, Susan James <[log in to unmask]> wrote:
> 
> 
> [resend with correction in the .cshrc snippet]
> 
> Hello All,
> 
> Loading software modules in a startup script is broke on SL-7.3 but works with no issue on Centos-7.x OS.  See below.  If anyone has clues as to how to fix this, please share.

loading software modules (or modifying environment variables like PATH, LD_LIBRARY_PATH etc. in any other way ) in "startup scripts" (and in particular ~/.*shrc, which is the worst way of doing this) is the root of all evil.

To fix it, just don't do it.

Move deviations from the default and supported environment into scripts *not* run during login. Way easier to debug, way less surprising.

- Stephan


> 
> [ snippet from .cshrc ]
> module load jdk/1.8.0_05
> module load python/3.5.1
> module load matlab/2018b
> module load gcc/6.3.0
> 
> 
> [ broke SL-7.3 ]
> apps5.als% id
> uid=40013(alsoper2) gid=31(controls) groups=31(controls)
> apps5.als% tcsh
> .user_aliases
> ModuleCmd_Load.c(213):ERROR:105: Unable to locate a modulefile for 'jdk/1.8.0_05'
> .epics.local - linux-x86_64
> .epics.local - EPICS_CA_ADDR_LIST set to controls network
> ModuleCmd_Load.c(213):ERROR:105: Unable to locate a modulefile for 'matlab/2018b'
> ModuleCmd_Load.c(213):ERROR:105: Unable to locate a modulefile for 'gcc/6.3.0'
> ModuleCmd_Load.c(213):ERROR:105: Unable to locate a modulefile for 'python/3.5.1'
> apps5.als% module list
> No Modulefiles Currently Loaded.
> 
> 
> [works Centos-7.x]
> sioc02.als% id
> uid=40013(alsoper2) gid=31(controls) groups=31(controls)
> sioc02.als% tcsh
> .user_aliases
> .epics.local - linux-x86_64
> .epics.local - EPICS_CA_ADDR_LIST set to controls network
> sioc02.als% module list
> Currently Loaded Modulefiles:
>   1) jdk/1.8.0_05   2) matlab/2018b   3) gcc/6.3.0      4) python/3.5.1
> sioc02.als% 
>