by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Jul 18 03:14
    dependabot[bot] labeled #2
  • Jul 18 03:14
    dependabot[bot] opened #2
  • Jul 18 03:14

    dependabot[bot] on npm_and_yarn

    Bump lodash from 4.17.15 to 4.1… (compare)

  • Jul 15 09:44

    j8r on gh-pages

    Update docs (compare)

  • Jul 15 09:42

    j8r on master

    Remove redundant next Update dependencies (compare)

  • Jul 15 09:09

    j8r on master

    Change CI image to official Alp… (compare)

  • Jul 10 09:57

    j8r on master

    Update dependencies; Update cod… (compare)

  • Jul 10 09:57
    j8r closed #36
  • Jul 09 22:19
    dscottboggs opened #36
  • Jul 09 20:56

    j8r on gh-pages

    Update docs (compare)

  • Jul 09 20:42

    j8r on master

    Format (compare)

  • Jul 09 20:41

    j8r on master

    Change deprecated usages and fi… (compare)

  • Jul 09 20:41
    j8r closed #7
  • Jul 09 20:41
    j8r commented #7
  • Jul 09 20:41

    j8r on master

    Change CI image to official Alp… (compare)

  • Jul 09 18:43
    dscottboggs opened #7
  • Mar 22 19:34

    j8r on gh-pages

    Update docs (compare)

  • Mar 22 19:33

    j8r on master

    Fix base package error for self… (compare)

  • Mar 22 19:29

    j8r on master

    Fix base package error for self… (compare)

  • Mar 22 19:24
    j8r commented #6
Julien Reichardt
@j8r
then -p 8081:80
and so on
D. Scott Boggs
@dscottboggs_gitlab
true
still prefer to just use the full url with SSL for consistency's sake, even on the same machine
Julien Reichardt
@j8r
Traefik is more suitable (when it works)
D. Scott Boggs
@dscottboggs_gitlab
thanks for the input
Julien Reichardt
@j8r
full url?
D. Scott Boggs
@dscottboggs_gitlab
I'm taking a look at caddy
Julien Reichardt
@j8r
I meant having then a reverse proxy on top of this
D. Scott Boggs
@dscottboggs_gitlab
on top of what?
Julien Reichardt
@j8r
there will be caddy -> localhost:8080 -> myapp:80
as an example
D. Scott Boggs
@dscottboggs_gitlab
what for?
two reverse-proxies??
Julien Reichardt
@j8r
no, just one
because the container port is bound to the host one
D. Scott Boggs
@dscottboggs_gitlab
oh I see
Julien Reichardt
@j8r
try for instance:
docker run -p 8080:80 --rm -it alpine nc -lv -p 80
curl http://localhost:8080
D. Scott Boggs
@dscottboggs_gitlab
yes, I'm aware of this feature, I just don't often use it
well, that's not true
I use docker for two different things -- a consistent dev environment and a reproducable deployment environment. In the dev env I use -p all the time, but not so much with deployments
Julien Reichardt
@j8r
isn't just simpler to fix the Traefik conf?
D. Scott Boggs
@dscottboggs_gitlab
It would be simplest to just revert to the old 1.7 config, but I've tried for a while to get the v2 config working and asked for help on their forum and I want a solution that's more solid and maintainable in the longrun. That's why I was thinking of something that generates nginx configs from docker labels and hooks into certbot -- it would be a thin layer over industry standards instead of a full monolithic solution.
Julien Reichardt
@j8r
what would be wrong to bind to local ports then?
then nginx will be able to redirect traffic to this ones
D. Scott Boggs
@dscottboggs_gitlab
you mean with docker -p? nothing, I just don't usually
nginx can redirect to docker IPs on the same network as it as well
Julien Reichardt
@j8r
right, but they will need to be statically set in case of reboots
(is that possible?)
D. Scott Boggs
@dscottboggs_gitlab
there could just be a service which rebuilds the configs on boot and every so often (using systemd timers?), checking the hash and rewriting if necessary
which was the part I was planning to build this weekend but idk it seems like way too much for how much time I have rn
Julien Reichardt
@j8r
imo it complicates all the thing
D. Scott Boggs
@dscottboggs_gitlab
hm?
Julien Reichardt
@j8r
for me more robut meant simpler under the ground logic
personally, I would go with caddy (simple configuration), then binding to port (or IP/port if you prefer)
I mean, this isn't like you have dozen of regular deployments, right?
D. Scott Boggs
@dscottboggs_gitlab
no, usually less than 10
Adrien D.
@dtrckd
@j8r thanks. How are we supposed to upgrade. I have :
$ dppm  a u dppm
INFO initializing: upgrade
INFO cache up-to-date: /srv/dppm/src/default
ERR! Not a semantic version: "2020-01-10"
Julien Reichardt
@j8r
dppm has not a semantic version :/
Julien Reichardt
@j8r
not installed with one I mean
Julien Reichardt
@j8r
thanks, fixed
replaced the dashes of git show -s --format=%ci by dots
Julien Reichardt
@j8r
excluding that, the upgrade works :)
the major challenge would be to upgrade properly bug apps like Nextcloud, which will require database migrations, plugins upgrade, cache
Julien Reichardt
@j8r
Got an idea - the environment variables can be in a file, instead of being inside the service
it is also more secure if there are credentials
Sam Squire
@samsquire
You can source environment variables nicely with
set -a
source variablesToExport
Julien Reichardt
@j8r
hum, no really with OpenRC, because I need to pass it to --env VAR=val VAR2=val2 etc
I've done --env $(egrep -v '^([ ]+|)# /some/file)
Julien Reichardt
@j8r
(missed a quote before the file name)