Learners do not implement GRPC health checks

[elias-dbx]

Bug report criteria

• This bug report is not security related, security issues should be disclosed privately via etcd maintainers.
• This is not a support request or question, support requests or questions should be raised in the etcd discussion forums.
• You have read the etcd bug reporting guidelines.
• Existing open issues along with etcd frequently asked questions have been checked and this is not a duplicate.

What happened?

I was testing out the new experimental health checks: #17914

What I found is that if there was a learner in the cluster, the RPCs would fail with:

{"level":"warn","ts":"2024-10-05T01:35:11.309621Z","logger":"etcd-client","caller":"io_etcd_go_etcd_client_v3/retry_interceptor.go:63","msg":"retrying of unary invoker failed","target":"etcd-endpoints://0xc0011bc780/10.247.133.150:5000","attempt":0,"error":"rpc error: code = DeadlineExceeded desc = latest balancer error: last connection error: connection active but received health check RPC error: rpc error: code = Unavailable desc = etcdserver: rpc not supported for learner"}

I believe this comes from here where we modify what RPCs learners implement:

etcd/server/etcdserver/api/v3rpc/util.go

Line 142 in 7cded36

func isRPCSupportedForLearner(req any) bool {

Should we allow learners to receive grpc health checks?

What did you expect to happen?

Learners do not cause RPCs to fail when experimental-stop-grpc-service-on-defrag is enabled.

How can we reproduce it (as minimally and precisely as possible)?

Enable experimental-stop-grpc-service-on-defrag, make a learner, create an etcd client with the following grpc service config:

`{"loadBalancingPolicy": "round_robin", "healthCheckConfig": {"serviceName": ""}}`

Anything else we need to know?

No response

Etcd version (please run commands below)

$ etcd --version
# paste output here
v3.5.14
$ etcdctl version
# paste output here

Etcd configuration (command line flags or environment variables)

paste your configuration here

Etcd debug information (please run commands below, feel free to obfuscate the IP address or FQDN in the output)

$ etcdctl member list -w table
# paste output here

$ etcdctl --endpoints=<member list> endpoint status -w table
# paste output here

Relevant log output

No response

[ahrtr]

cc @tjungblu @chaochn47 do you have time to take a look at this ticket? thx

[tjungblu]

Learners do not cause RPCs to fail when experimental-stop-grpc-service-on-defrag is enabled.

are you sure that is related? I would assume you just have a learner member in your client and you hit it with a linearized read/write call.

[ahrtr]

Thanks @tjungblu for the comment.

@elias-dbx if you still have any questions, please raise them in discussions.