This one is truly baffling me. Can someone please please point me in the direction of what i've done wrong.
Top and bottom is - Nothing has changed (on my side), but now its broken.
Gitlab CI deploy stage looks like this (deploying to heroku via heroku cli).
heroku-deploy-staging:
extends: .heroku-deploy
needs: ['test-heroku-env', 'build-staging-frontend']
script:
- heroku builds:create --app my-app-name # WORKS FINE
- cat ./build/flushall | heroku redis:cli --app my-app-name --confirm my-app-name
... and so on.
Entire CI passes, the app deploys to heroku, heroku confirms app deploy. Final stage is to flush redis.
Response... Error: No Redis instances found
This is the strange part.
Doesn't work β
(flushall file π)
FLUSHALL
QUIT
cat ./build/flushall | heroku redis:cli --app my-app-name --confirm my-app-name
Doesn't work β
echo "FLUSHALL\r\nQUIT" | heroku redis:cli --app my-app-name --confirm my-app-name
Connects to redis β
(but with no commands, so hangs infinitely)
heroku redis:cli --app my-app-name --confirm my-app-name
As a final sense check... I added this command BEFORE redis, just to confirm it's there.
heroku addons --app my-app-name
Yep - REDIS is there, created and available.
So why on earth, does it fail only when I pipe a command to it??
I can connect with no command piped. Cannot be found without the command piped.
Any help on this would be hugely appreciated. One entire morning has been wasted so far...