search

timescale / docs.timescale.com-content

Content pages for TimescaleDB documentation
https://docs.timescale.com
37 stars 51 forks source link

Include restore points in backup/restore section #700

Closed pmwkaa closed 3 years ago

pmwkaa commented 3 years ago

Since restore points are now supported for the multi-node environment, remove previous warning about it and update the section

k-rus commented 3 years ago

@pmwkaa Will the short description how to do multinode backup be in a separate PR?

pmwkaa commented 3 years ago

@k-rus It is not clear to me how we should document it other then already there, since this functionality is PostgreSQL compatible and follows Performing Physical Backups section. What user should do exactly, other then already described? This is much more related on the internal infrastructure of user, rather then specific set of commands to execute.

If you take a look at the backup section using WAL-E https://docs.timescale.com/latest/using-timescaledb/backup#docker-wale, everything is described in terms of docker containers rather then a way to configure PostgreSQL recovery, which makes sense.

I don't think we have the same for multi-node. Feels like this should be a part of a way to configure and maintain/run multi-node cluster using Kubernetes, etc.

I guess it is up for discussion, what people think we should include there? @erimatnor @mfreed @NunoFilipeSantos

k-rus commented 3 years ago

@pmwkaa The description can be simple, like: create the distributed restore point from the access node, and then backup the access node and each data node. During the restore of each node use the restore point. Since the distributed restore point was created on each node to preserve consistency between nodes, the restoring to it will result in the consistent state.

pmwkaa commented 3 years ago

@pmwkaa The description can be simple, like: create the distributed restore point from the access node, and then backup the access node and each data node. During the restore of each node use the restore point. Since the distributed restore point was created on each node to preserve consistency between nodes, the restoring to it will result in the consistent state.

Feels like the same is already said here and in the api description, I guess I can reformulate it and separate into new section

helrabelo commented 3 years ago

Please add the labels for the versions these changes apply? @pmwkaa