SeaDAS 9.0.1 errors on launch in RHEL 8 HPC ssh -X

[robbiethek]

Using a Mac, running ssh -X into a HPC cluster then starting an interactive sessions via srun -X and XQuartz installed on a Mac, all of these errors appear:

Via the CLI:

Code: Select all

WARNING: An illegal reflective access operation has occurred
WARNING: Illegal reflective access by org.netbeans.ProxyURLStreamHandlerFactory (file:/opt/opt/SeaDAS/9.0.1/platform/lib/boot.jar) to field java.net.___URL.handler
WARNING: Please consider reporting this to the maintainers of org.netbeans.ProxyURLStreamHandlerFactory
WARNING: Use --illegal-access=warn to enable warnings of further illegal reflective access operations
WARNING: All illegal access operations will be denied in a future release

Via pop up alerts

Code: Select all

Unable to install auxdata/color_schemes
Unable to install auxdata/rgb_profiles
Unable to install auxdata/color_palettes
Unable to initialize WWWorldMapToolView: Profiles [GL4bc, Gl3bc, GL2, GLES1] not available on device null

[treble]

I run SeaDAS sometimes on a CentOS7 server, through ssh. I get the same CLI errors, so wouldn't worry too much about those. I don't get the pop-ups.

It looks like it can't write to the auxdata directory. That is by default located at ~/.seadas9/auxdata. Mine looks like this:

Code: Select all

[server:~/.seadas9/auxdata 1013]$ pwd
/home/xxxxxx/.seadas9/auxdata
[server:~/.seadas9/auxdata 1013]$ ll
total 24K
drwx------. 12 xxxxxx xxxxxx 4.0K Nov  1 10:20 .
drwx------.  9 xxxxxx xxxxxx 4.0K Nov  4 12:01 ..
drwxrwxr-x.  2 xxxxxx xxxxxx 4.0K Oct 23 14:05 color_palettes
drwxrwxr-x.  2 xxxxxx xxxxxx 4.0K Oct 23 14:05 color_schemes
drwxrwxr-x. 12 xxxxxx xxxxxx 4.0K Nov  1 10:20 dem
drwxrwxr-x.  3 xxxxxx xxxxxx   27 Oct 23 14:05 hdf_natives
drwxrwxr-x.  7 xxxxxx xxxxxx   85 Oct 23 14:05 meris-operators
drwxrwxr-x.  3 xxxxxx xxxxxx   29 Oct 23 14:05 netcdf_natives
drwxrwxr-x.  3 xxxxxx xxxxxx   29 Oct 23 14:05 openjpeg
drwxrwxr-x.  4 xxxxxx xxxxxx 4.0K Oct 23 14:05 rgb_profiles
drwxrwxr-x.  3 xxxxxx xxxxxx   32 Oct 23 14:05 s2tbx
drwxrwxr-x.  2 xxxxxx xxxxxx   10 Oct 23 14:05 tool-adapters

[robbiethek]

It looks like it can't write to the auxdata directory.

Which error indicates that? Is this something that should be changed in the source code?

Comparing mine, there is no write for the group:

Code: Select all

ls -l ~/.seadas9/auxdata
total 36
drwxr-xr-x 2 myuser user 4096 Nov  4 11:04 color_palettes
drwxr-xr-x 2 myuser user 4096 Nov  4 11:00 color_schemes
drwxr-xr-x 3 myuser user 4096 Nov  4 10:47 hdf_natives
drwxr-xr-x 3 myuser user 4096 Nov  4 10:47 meris-operators
drwxr-xr-x 3 myuser user 4096 Nov  4 10:47 netcdf_natives
drwxr-xr-x 3 myuser user 4096 Nov  4 10:47 openjpeg
drwxr-xr-x 4 myuser user 4096 Nov  4 11:00 rgb_profiles
drwxr-xr-x 3 myuser user 4096 Nov  4 10:47 s2tbx
drwxr-xr-x 2 myuser user 4096 Nov  4 10:47 tool-adapters

FWIW the warning does say:

Code: Select all

WARNING: Please consider reporting this to the maintainers of org.netbeans.ProxyURLStreamHandlerFactory

For other cluster users I found a way to work around this error

Code: Select all

unable to create directory '/run/user/####/dconf': Permission denied. dconf will not work properly

In the module file add this:

Code: Select all

prepend-path	 XDG_RUNTIME_DIR      /local/$::env(USER)

I know these errors can be ignored I've never found a true solution:

Code: Select all

libEGL warning: MESA-LOADER: failed to open swrast (search paths /usr/lib64/dri)
libEGL warning: MESA-LOADER: failed to open swrast (search paths /usr/lib64/dri)
MESA-LOADER: failed to open swrast (search paths /usr/lib64/dri)
libGL error: failed to load driver: swrast
libEGL warning: MESA-LOADER: failed to open swrast (search paths /usr/lib64/dri)
libEGL warning: MESA-LOADER: failed to open swrast (search paths /usr/lib64/dri)

[treble]

I think the error 'unable to install' indicates that it cannot write to that directory, but there might be another cause as well. The directorly permissions look good to me though.

For what it is worth I use the following syntax to request an interactive bash shell on our cluster (4 cpus, 6 hours, x11 forwarded, not sure what the pty does):

Code: Select all

srun --pty --x11=all -c 4 -t 6:00:00 bash

Then I load these 2 modules:

Code: Select all

ml system x11

The cluster is running CentOS 7.