Closed fordprefect480 closed 2 years ago
Hi,
I haven't tried using it without a domain name as the setup uses acme-companion to fetch a Let's Encrypt certificate, and I'm not sure if it even works without.
Things to try:
docker-compose logs grampsweb
, docker-compose logs proxy
)cat /opt/grampsweb/letsencrypt.env
if you don't remember)Resources seem ok. 70-80% memory usage is possibly a bit high.
Hmm there's nothing at cat /opt/grampsweb/letsencrypt.env
, definitely the correct path?
root@gramps-web:/opt/grampsweb# cat /opt/grampsweb/letsencrypt.env
cat: /opt/grampsweb/letsencrypt.env: No such file or directory
root@gramps-web:/opt/grampsweb# ls
docker-compose.yml firstlogin.py firstlogin.sh media nginx_proxy.conf
I ran firstlogin.sh
to regenerate the letsencrypt.env file and here's what's in it:
root@gramps-web:/opt/grampsweb# cat letsencrypt.env
VIRTUAL_HOST=familytree.owen.nz
LETSENCRYPT_HOST=familytree.owen.nz
LETSENCRYPT_EMAIL=owen.symes@gmail.com
Hi, the memory usage is indeed a bit high. You could try with a larger droplet and/or changing the GUNICORN_NUM_WORKERS
environment variable in /opt/grampsweb/docker-compose.yaml
to 2.
Have you configured your subdomain correctly, i.e. added an A record pointing to your DO IP address?
It would be great if we could get this to work and update the documentation if we find something that doesn't work the way it's currently documented.
I ran firstlogin.sh to regenerate the letsencrypt.env file and here's what's in it:
Have you restarted the containers after generating the file? (Or even easier restarted the droplet.)
OK here's what I've done:
GUNICORN_NUM_WORKERS
to 2Will wait a few hours for the DNS change to propagate.
Were there no clues in the logs in my previous comment? I thought the "Unable to init server: Could not connect: Connection refused" was a bit suspicious but I don't know much about nginx.
Oh wow I'm in:
So I can get to it only via my familytree.owen.nz address - browsing directly to https://157.230.247.174/ still doesn't seem to work.
Trouble is I'm not sure which one of the above fixed it. Perhaps it was the reboot after generating the letsencrypt.env file?
Great!
I just created a fresh droplet and I'm also in, but I had the issue that I SSHed in too early, the first run script was already shown, but docker compose was not installed yet, and so it failed to start and I had to manually do docker-compose up -d
. So the installation script needs some tuning.
Please let me know if there are further issues.
I encountered the same issue the first time I created my droplet. I blew it away and created a new one, this time following your explicit instructions to wait a few minutes before SSHing in and it worked.
Not sure if we got any learnings out of this but I'll start filling out my tree now and keep you posted if I encounter any issues. Perhaps I'll even brush up on Lit and send some PRs your way.
Cheers mate
I created a Gramps Web droplet (Version 1.0 / OS Ubuntu 20.04), waited for the droplet to be created, then waited an additional 5 minutes before SSHing into the machine. I then completed the 1-click app setup. Once that completed I opened up a browser hoping to see the Gramps GUI-based setup process however I received a 503 instead:
I've followed the instructions here and even though I haven't configured my domain to point to the droplet's public IP yet, I would still have expected to get a happy webpage here. I may well be missing a step that is obvious to devs who have completed the dev environment setup but is not mentioned in the Digital Ocean instructions?
Disclaimer: I am a dev however I'm quite new to docker and configuring linux distros. I'll do my best to get you what you need to help me figure out what's going on :)
SSH session:
I then restarted the droplet. Still received the 503.
Attempted to confirm that things are indeed running correctly: