S3 data replication #5899
-
|
Hi. Is it possible to have Mimir duplicate writes to more than one S3 target? This is for data availability purpose. I have been looking at store-gateway configuration, but I'm not sure I fully understand the architecture. Thanks you in advance and please excuse my limited understanding of Mimir - I'm a usually a full time storage guy. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
it's currently not possible to do that. The ingesters and the compactor are the components which write to the S3 bucket and they only support writing to a single endpoint. The store-gateways also need a unified view of the bucket (i.e. you cannot configure different availability zones differently). I think to achieve this you'd need to take care of the bucket sync in the background and reconfigure which bucket mimir uses when one is unavailable. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you very much for your explanation. I'll pursue a different way to archive the desired fault tolerance then. |
Beta Was this translation helpful? Give feedback.
it's currently not possible to do that. The ingesters and the compactor are the components which write to the S3 bucket and they only support writing to a single endpoint. The store-gateways also need a unified view of the bucket (i.e. you cannot configure different availability zones differently).
I think to achieve this you'd need to take care of the bucket sync in the background and reconfigure which bucket mimir uses when one is unavailable.