Hi everyone, Because of bugs in v0.12.0 I'd like to deploy new DataHub v0.11.0 and clone all metadata from DataHub v0.12.0 there. Do you know what is the best and preferable way to do it?

can’t you just run the ingestion pipelines again? i’ve been treating DataHub as a stateless thing.. We aren’t even allowing people to document their datasets/dashboards/dbt models there yet.. precisely because we don’t have a stable process for this restoration in case of failures. But I’m curious to see solutions from the community.

Hi @gifted-rocket-73005, thank you for the reply. The idea was to transfer datasets fields descriptions also. As I understand, it will not be restored if I run ingestion pipelines on the new instance

@gray-shoe-75895 do you have any suggestions here?

I suspect the datahub ingestion source might be of interest for the datahub -> datahub migration use case https://datahubproject.io/docs/next/generated/ingestion/sources/datahub/ @gifted-rocket-73005 I'd love to understand your reservations around the datahub backup/restore process. We do have docs on it https://datahubproject.io/docs/next/how/backup-datahub/, although I acknowledge that it's not written in the most user-friendly way right now

Its not like I have reservations around it. What I meant about “stable process” was more like a “crash plan” or an automated way of backing up the PVC (or exporting the pgdump somewhere) and recovering the backup when necessary. We haven’t developed one for DataHub yet.. and therefore my team, as a platform team, says it won’t support it like this for now. But do I have time to work on this these days? Absolutely not haha 😅

Fair enough. Acryl offers managed datahub, which might be worth considering to avoid those operational concerns