production update from 2.0 to 2.1 errors

[ezra.j.rice]

Following the instructions found in upgrade to 2.1.0 I've run into some errors.

Environment Info:

Ubuntu 16.04

Installing from installation script with the following options:

bash install_tethys.sh -b release --configure-selinux --production -S ezrajrice -P 'mypass' --db-password 'dbpass'

Error #1:

ERROR: tethys.tethys_apps.harvester:App myapp not loaded because of an issue with loading urls

App import errors when import starts with tethysapp such as 'from tethysapp.myapp.model import x, y, z'

Results in ImportError: No module named 'tethysapp'

Error #2:

Install script could not link uwsgi nor link nginx config files because they already exist.

Error #3:

After installation file completed, 'tuo' and 'tso' fail with

chown: cannot access '': no such file or directory

[Sarva Pulla]

Hi Ezra, 

The following work around fixed the issue temporarily for Error #3. You will have to change the tethys-activate.sh file in the following directory: ~/tethys/miniconda/envs/tethys/etc/conda/activate.d/

Currently, there is no variable for TETHYS_SRC. That's why it's unable to run the following aliases tuo,tso or tsr. Insert the following into the file right after the first block of export variables:

export TETHYS_SRC='/home/prod/tethys/src'

Then run the following command to update the bashrc:

. ~/.bashrc

This helped me get past it. Hopefully, it will work for you. We do need to open an issue to patch it up. Thanks!

[swainn]

We just issued a patch release, 2.1.1, that should address this issue. A fresh install of 2.1 will be required to get these changes.