Best practice in 2024 for simple RGW failover
Hi, The requirements are actually not high: 1. there should be a generally known address for access. 2. it should be possible to reboot or shut down a server without the RGW connections being down the entire time. A downtime of a few seconds is OK. Constant load balancing would be nice, but is not necessary. I have found various approaches on the Internet - what is currently recommended for a current Ceph installation? Thanks,
The requirements are actually not high: 1. there should be a generally known address for access. 2. it should be possible to reboot or shut down a server without the RGW connections being down the entire time. A downtime of a few seconds is OK.
Constant load balancing would be nice, but is not necessary. I have found various approaches on the Internet - what is currently recommended for a current Ceph installation?
I am having haproxy, I think I saw this most here. When I would scale rgw in the orchestrator it would automatically load balance. Advantage is also that you can do a little more with security, url filtering and url rewriting.
participants (2)
-
E Taka
-
Marc