living-atlases / la-toolkit Goto Github PK

I you install the toolkit using an user with an id different that the default one (1000) the volumes are mounted with ids different that (ubuntu the user of the container also has the id 1000).

Workaround:

docker exec -u ubuntu -i -t la-toolkit bash
sudo chown ubuntu:ubuntu /home/ubuntu/ansible/la-inventories/ /home/ubuntu/ansible/logs /home/ubuntu/.ssh  /home/ubuntu/ansible/ala-install

The backend should verify fix and autocorrect the permissions.

Very strangely, we cannot use the SSH gateway for the FIRST server defined.

In the screenshot below, gateway (recol2) is used to connect to recol1.

I had to re-create my configuration, starting by the VM recol2 and then recol1, because if recol1 was defined first, it wasn't saved (when you go back to this screen, the gateway part is empty)

settings.baseFooterUrl in branding is not configured correctly (the la-toolkit configures the main home url)

thanks @guillemfrancisco for the feedback

In Ubuntu 20.04 LTS, running ./dockerTas.sh compose, fails with a message:
Opening site......................../dockerTask.sh: line 104: xdg-open: command not found
That command is provided by xdg-utils, which can be installed with the command:
apt install xgd-utils

I seems that new portals are created with /cas as auth context instead of empty:

instead of:

Also for detecting branding ssl cert issues

Reported by Johannes Eberenz :

When I set the ssh users to match the user name i set for "default user in your servers", this worked. After removing the user name again from the advanced ssh config, (but not changing the "default user name in your servers"), it seems la-toolkit used again ubuntu user for ssh. So I again set the ssh user names.

When we zoom in the browser, we cannot click anymore on button or switch

Naama Arkin reported that Ecodata and ecodata-reporting software version cannot be configurable in the Tune tool (insteasd biocollect is selectable).

I see that the backend responds correctly the versions that retrieves from nexus:

But for some reason is not processed:

During ssh config generation the assh.d directory should be created if it's not present for some reason.

Thanks @sylmorin-gbif for the feedback.

Verify this

Branding deployment is failing because some npm libs are missing.

Executing:
npm install npm update

is fixing the issue.

It would be great to have it executed by default

Docker compose starts 4 containers not 3 - la-toolkit-mongo-db-backup is missing from README (see https://github.com/living-atlases/la-toolkit#or-directly-using-docker-compose).