Trying to restore data from docker desktop to a docker engine

Hi

I’m trying to transfer the data from a docker desktop to a docker engine.
We’re using the weaviate 1.25.0

Firts of all we’re taking all the data from the volume of the docker desktop container, creating a .tar file.

Then we create the containers inside the docker engine using the docker compose up.

Then we restore the data from the old volume to the new one using wsl commands.

Once we’re done, we look into the volume and there is all the schemas and data stored, but when I try to access to the data using the url v1/objects, its empty.

What are we doing wrong? How can I transfer all the data to my docker engine container.

Regards

hi @garcia.e !!

Welcome to our community :hugs:

Here are suggested steps to move backups around:

  1. spin a target cluster, with the very same version used in source cluster
  2. configure the same s3/gcp/azure bucket on both servers, as per the backup module docs.
  3. create a backup in source cluster
  4. check if the backup was indeed generated in the bucket
  5. restore the same backup in target cluster
  6. upgrade target cluster. Don’t skip releases: upgrade to 1.26.latest, restart. Then 1.27.latest, restart. And finally 1.28.latest

This is important to make sure that any migrations that were created between 1.25.current and 1.28.latest are run.

because those clusters share the backup bucket, they can backup and restore between each other. You need to only use the same id when creating the backup.

Note that if you have a single node cluster, and want to restore the backup on a multi node cluster, this approach will not work.

In that case, you need to migrate your data. Migrating your data is different, because when you insert the objects, all the calculations will be computed again and the data will be replicated as per the new configuration.

For that case, you can spin a target cluster (and here it can already be in the latest version ), and use the migration guide to move your data over:

Note that, migrating your data will consume some resource, so depending on the size of your dataset, you will need to allocate the resources accordingly.

Let me know if that helps!