3 ms·
> ...with a replication factor of 10 RF 10 is insane overkill. There is probably someone out there doing it, but it is completely unnecessary for %99.999 real
by helper 12y ago
> ...with a replication factor of 10
RF 10 is insane overkill. There is probably someone out there doing it, but it is completely unnecessary for %99.999 real world deployments.
Getting back to your original comment, Azure persistent storage (or AWS, or Google) isn't giving you anything near RF10 * 2 DCs.
- tracker1 12y agoAzure storage, each write is to two local and a third (optionally geo redundant) location... This is data that will still be there if my VM reboots... the local/scratch disk is gone if my VM reboots... if a significant number of those VMs reboot, you will lose data. I do hope that you are at least backing up those nodes regularly to persistant options.
- helper 12y agoI don't know how Azure works but AWS reboots are fine with ephemeral disks. We backup all sstables to s3. We've never needed to restore a node from an s3 backup in 5 years of running in production.