Is able Redshift to restore leader node upon fail as it would do with compute nodes?

Question

Documentation states that upon node fail, Redshift provisions new nodes, restore data from nodes/S3, etc..will this be done as well upon leader node fail o will the whole cluster go down, hence needing a whole cluster restore?

Accepted Answer

Hi,

A failed leader node will be replaced and the meta-data will be replicated from the other nodes/S3 to restore functionality of the cluster. Since the meta-data is typically very small the recovery time is also low once the hardware has been replaced.

No restore from snapshot should be required.

Is able Redshift to restore leader node upon fail as it would do with compute nodes?

Conteúdo relevante