I am trying to follow the steps of the following StackOverflow answer:
Running ChromeDriver with Python selenium on Heroku
But I can't figure out how to do this part.
Question
How do I create an environment variable GOOGLE_CHROME_BIN, with the path of chrome on heroku: /app/.apt/usr/bin/google-chrome and an environment variable called CHROMEDRIVER_PATH with the path of chromedriver on heroku: /app/.chromedriver/bin/chromedriver?
Using Heroku CLI for your app:
$ heroku config:set GOOGLE_CHROME_BIN=/app/.apt/usr/bin/google-chrome
$ heroku config:set CHROMEDRIVER_PATH=/app/.chromedriver/bin/chromedriver
Or you can do this by editing the config vars on your app’s settings tab on Dashboard
Related
I created a Hugo application which I wanted to deploy through Heroku.
Following Hugo's instructions on their github readme, it says to create a Heroku app with the command:
$ heroku create --buildpack https://github.com/roperzh/heroku-buildpack-hugo.git
followed by:
$ git push heroku master
$ heroku open
Upon doing so, however, I keep getting an error:
remote: -----> Failed to detect app matching https://codon-buildpacks.s3.amazonaws.com/buildpacks/heroku/go.tgz buildpack
and the push ultimately fails.
Most of what I've looked up tells me that it needs a specification of what language is being used, which in this case is Go.
But even when doing:
heroku buildpacks:set -a myproject heroku/go
It tells me:
! The buildpack heroku/go is already set on your app.
So, if it's already set on my app, then why can't it detect an app matching the buildpack?
Check your hugo version using
hugo version
and then set your hugo version on heroku
for example 0.40
heroku config:set HUGO_VERSION=0.40 -a <your app name>
I have installed heroku with homebrew in my OSX El Capitan(10.11.5) and installed the heroku-container-tools with:
heroku plugins:install heroku-container-tools
The installation goes successfully:
Installing plugin heroku-container-tools... done
When I run a heroku version, I got this:
heroku-toolbelt/3.42.22 (universal.x86_64-darwin15) ruby/2.0.0
heroku-cli/5.2.20-9d094b0 (darwin-amd64) go1.6.2
=== Installed Plugins
heroku-container-tools#3.0.0
But when I run a heroku container:init I got this:
! `container:init` is not a heroku command.
! See `heroku help` for a list of available commands.
Running heroku help container I got this:
Usage: heroku container
Use Docker to build and deploy Heroku apps
Use Docker to build and deploy Heroku apps
Additional commands, type "heroku help COMMAND" for more details:
container:login # Logs in to the Heroku Docker registry
container:push [PROCESS] # Builds, then pushes a Docker image to deploy your Heroku app
I am following this heroku tutorial: https://devcenter.heroku.com/articles/local-development-with-docker
In the tutorial is said to run heroku container:init
This command was replaced but some other and heroku did not updated their documentation or I have some problem installing the plugin?
Yes, they restricted access to their container registry, just read the warning on top of the tutorial you're following https://devcenter.heroku.com/articles/local-development-with-docker
heroku container:release, which creates a Heroku-compatible slug and
deploys it to Heroku, has been deprecated. For access to our container
registry (available in private beta), please contact
docker-feedback#heroku.com
now there are only login and push commands. I hope you can easily ask for the access.
When I use the command heroku addons create:graphenedb --version v195. It gives error "No app specified , Run this command from an app folder or specify which app to use." I am new to Heroku and I do not understand which app folder it is talking about.
If you're not in your project folder or if it doesn't have a git remote to point it to Heroku you need to use the --app flag for the heroku command to specify the name of your app on Heroku
I want to call heroku postgres backup/restore commands within a scheduled heroku task, but the heroku toolbelt isn't available from bash prompt, so I can't call heroku commands:
$ heroku run bash --app myapp
Running `bash` attached to terminal... up, run.4805
~ $ heroku --version
bash: heroku: command not found
How can I get heroku commands available in my scheduled bash script?
I don't know anything about Ruby or Ruby Gems.
None of the other solutions worked for me since Heroku locks the filesystem after the buildpacks finish running.
There's a third-party buildpack that installs the CLI for you. First you set your auth key as an ENV variable on your app:
heroku config:set HEROKU_API_KEY=`heroku auth:token` -a myapp
Then add the buildpack:
heroku buildpacks:add https://github.com/heroku/heroku-buildpack-cli -a myapp
Redeploy your app and it should be able to access your apps via the CLI. If you have https://devcenter.heroku.com/articles/dyno-metadata enabled, you can even access your current app's name with $HEROKU_APP_NAME.
Put this at the top of your scheduled bash script:
# install heroku toolbelt
# inspired by https://toolbelt.heroku.com/install.sh
curl -s https://s3.amazonaws.com/assets.heroku.com/heroku-client/heroku-client.tgz | tar xz
mv heroku-client/* .
rmdir heroku-client
PATH="bin:$PATH"
then, for example, you can call:
heroku pg:reset HEROKU_POSTGRESQL_YELLOW_URL --app myapp-staging --confirm myapp-staging
heroku pgbackups:restore HEROKU_POSTGRESQL_YELLOW_URL `heroku pgbackups:url --app myapp-production` --app myapp-staging --confirm myapp-staging
and poof! Staging database is updated from production database.
As Heather Piwowar said, you can indeed download and untar the heroku tollbelt yourself, but no need to move files around after this. Here is a shorter version:
curl -s https://s3.amazonaws.com/assets.heroku.com/heroku-client/heroku-client.tgz | tar xz
PATH="/app/heroku-client/bin:$PATH"
You can now use the heroku command as you wish.
For authentification, you may want to set the HEROKU_API_KEY env var (using heroku config:set HEROKU_API_KEY=1234567890 -a your-app-name).
Also, note that the first use of the heroku command will run longer than expected because it will try to install the latest version, dependencies and core plugins.
You can use Heroku Scheduler and configure the following command (as an example) to create a database backup:
curl -s https://s3.amazonaws.com/assets.heroku.com/heroku-client/heroku-client.tgz \
| tar xz && ./heroku-client/bin/heroku pg:backups:capture -a you-app-name-here
For this to work, you need to add a Config Variable named HEROKU_API_KEY and set its value to the "API Key" value from your Accounts page.
so I have building management app, which locally uses the pdftk to generate pdf forms (prefilled with tenant/expense data). Works like magic on my local machine.
However, on Heroku I get the error:
pdftk executable /usr/local/bin/pdftk not found
in the logs when I try to generate the pdf file. I realize that I need to install pdftk on my heroku app using a buildpack. I've tried following some tutorials with Vulcan, but vulcan is deprecated and they say use heroku run, however I can't find much documentation of how to install the pdftk-source: https://github.com/millie/pdftk-source using heroku run.
I'm going to try https://github.com/millie/heroku-buildpack-ruby-pdftk, but if there is an easier/less messy way let me know, thanks!
EDIT:
Tried the above method, and now my heroku logs say:
RuntimeError (pdftk executable /app/vendor/pdftk/bin not found)
So I'm thinking the buildpack didn't include pdftk to begin with, which doesn't make sense.
I must be doing something wrong, but I followed the instructions exactly, only difference is I used dropbox instead of S3 to store the tar.gz file (the pdftk source)
EDIT:
OK, I figured out how to include pdftk executable in the heroku buildpack and upload it successfully as part of the app environment. HOWEVER, for some very strange reason, in the heroku bash console, when I cd into pdftk directory and try to run the executable, heroku bash says pdftk executable not found.
It works on my local machine, when cd into the pdftk directory and run pdftk, it runs the executable, so its not the executable..so why isn't it working inside the heroku bash directory?
SOLUTION:
Was missing setting LD_LIBRARY_PATH in my config vars on heroku, because pdftk relies on a library file. Also, remember to tar the tar.gz to the root directory and set a PATH to /bin/pdftk. Just check out #heroku on IRC, that's where I got my answer.
Was missing setting LD_LIBRARY_PATH in my config vars on heroku, because pdftk relies on a library file. Also, remember to tar the tar.gz to the root directory and set a PATH to /bin/pdftk. Just check out #heroku on IRC, that's where I got my answer.
This is how i setup pdftk in nodejs app in heroku
Create heroku app
heroku create
Set buildpack for pdftk
BUILDPACK_URL=https://github.com/ddollar/heroku-buildpack-apt
Set buildpack for nodejs
heroku buildpacks:add --index 1 heroku/nodejs
Add the libgcj.so.* to your search path:
heroku config:set LD_LIBRARY_PATH=/app/bin
Turn on at least one dyno
heroku ps:scale web=1
Create a Procfile in the root of your project and define the following:
web: node server.js
Push changes in heroku
git push heroku master
Adding answer that worked as of July 2019
Note: Thanks to shake-apps for creating the Buildpack https://elements.heroku.com/buildpacks/shake-apps/heroku-buildpack-pdftk
If you need to have PDFtk installed on Heroku for a Node JS Application, follow these instructions:
Heroku Installation
You'll need to set the base nodejs buildpack, and the pdftk buildpack by shake-apps.
Set them by running the following:
heroku buildpacks:set heroku/nodejs;
heroku buildpacks:add --index 1 https://github.com/shake-apps/heroku-buildpack-pdftk.git;
After build packs are set, the server can be deployed normally with
git push heroku master
If you have any issues with the deploy it could be your previous buildpack. You can clear it to start fresh with:
heroku buildpacks:clear