In the section replication mechanics here I see below quote
The follower index automatically retrieves some updates applied to the leader index, while other updates are retrieved as needed
I don't sure what is is "as needed" means. I think you should elaborate more in the documentation. And more interesting how this "as needed" works.
There are various settings, how I can understand which exactly properties are replicated and which no?
My use case is this
I have leader cluster, application works with it. I am configuring follower - it acts ONLY as a backup cluster in case if I will my leader failed and I decided to switch to follower.
- Can I be sure that this cluster settings are total copy of settings that were on my leader cluster?
- How I can understand list of dynamic properties that are not replicated (for example num_of_replicas is not replicated)? I need it for the case of switch.
I tried to update mappings on leader cluster and saw that when I perform GET host:port/index/_mappings - it immediately replicate this to follower.
If I update this again, without performing below get request - it is also replicated, but with some lag 20-30s.
- Is it normal behavior? What is this lag or where it configured?