Hi Everyone, What are you doing for backups for on-premises deployments (C* 3.11)? We don’t have the option of EBS snapshots, tablesnap to S3, etc. but we have lots of NFS-mounted file server space.
We’re currently doing a snapshot and then copying the full snapshot to a file server every night. This is tremendously time consuming & wasteful of disk space. I’m investigating Medusa. It looks like as easy way to transition to an incremental backup strategy…but after playing with it for the last day or two, it seems like it has a lot of rough edges, which gives me some hesitancy. Does anyone run Medusa in production? What has your experience been? Is it reliable, functionally correct, etc? … or is there some other option I should be looking at? I suppose I could also write my own script based on “incremental_backups: true” in cassandra.yaml; but I’d like to avoid creating a DIY tool if at all possible; given all of the nuance (saving token ranges, schemas, etc). Maybe best option is to submit PRs to help smooth the Medusa "rough edges"??? :-) Thanks everyone. - Max
