initially i had a docker stack (using docker-compose) which used the influxdb docker image (version 1.8) and stored the data on a docker-compose managed named volume.
Now, the application layout changed and i need to transfer the data from the previous stack to the new one. So i thought it would be easy and i copy the contents from the old volume to the new one and i can just start my new stack and it will be alright.
I used Guido Diepen’s script (docker-convenience-scripts/docker_clone_volume.sh at master · gdiepen/docker-convenience-scripts · GitHub) for transfering the data.
After copying i compared the content of the volumes and they are exactly the same.
But now when i launch the new stack, influxdb still starts from “scratch” and wants me to create an admin user etc., it does not take the existing data into account. It seems some information is missing.
I also found another way to “solve” this partially using backup and restore (but needed to realise that the backup functionality does not include the user data - see No way to make a full(users+databases) restore from '-portable' backup · Issue #9723 · influxdata/influxdb · GitHub). At least my data is transfered.
This is why i would prefer my initial approach.
Can somebody explain while the existing data is not used? Or if it is possible to make it work like this at all?