After setting up the docker container based on the docs as a test system, I added a reverse proxy to add HTTPS as well as use a decent name, but when I changed the BASEROW_PUBLIC_URL environment variable, upon the next login the system asked me to create the admin account, with all data already in the system gone … !? How is it that the base url is linked to the whole database? When I changed it back to the original URL, it once again asked to create the admin login …
Is the data that was in the system gone? I checked a pg_dump and it appears as if my data is indeed gone from a cursory browse through the SQL data … !?
Update: Seems like I made a mistake in the docker config, instead of having the /baserow/data directory linked into the master system’s filesystem, I had /opt/baserow/data linked, so on a rebuild, the data directory also got rebuilt … now just got to see how to restore the pg dump into the real system from the backup …
Edit: got the whole directory restored, only lost today’s changes … phew …
3 Likes
@GarryG just to check, did you also manually set the DISABLE_VOLUME_CHECK env variable? We added a specific check on startup in the baserow/baserow:XYZ images to make sure that /baserow/data was linked to a volume so users didn’t run into this problem. I’m asking as I’d like to make sure that this check is still actually working but was just perhaps disabled in this situation.
No, I did not … as I said, by my own stupidity the table data wasn’t kept outside of the container, which caused the loss of the data … at least in my case, it wasn’t any fault or error of the Docker container …