Adding Multiple PD Endpoints Does Not Provide Redundancy

[ProjectInitiative]

I have a TiKv cluster running in 3 test VMs: 10.10.0.111:2739, 10.10.0.112:2739, 10.10.0.113:2739

Starting a basic mount with the following:

./target/release/tifs -f tifs:10.10.0.111:2379,tifs:10.10.0.112:2379,tifs:10.10.0.113:2379 /mnt/tifs/

Dec 18 10:00:35.994 INFO connect to tikv endpoint: "10.10.0.113:20162"

If I pause, stop or kill VM 10.10.0.113, there is no failover, the file system doesn't try one of the other endpoints. Is this expected?

Are there plans to add redundancy?

Thanks

[ProjectInitiative]

Doing more research, looks like the TiKV_client library might not be working correctly as the docs states is should be working. https://docs.rs/tikv-client/latest/tikv_client/struct.TransactionClient.html#method.new_with_config

Going to have to dig a bit deeper to see why the client library is not connecting to multiple PD nodes