Question about deploying in k8s

[darrenxyli]

Tried follow with steps in Documentation, but in primary meta node, always receive following logs

ERROR openraft::replication: error replication to target=2, error: timeout after 1s to replicate 1->2

or

ERROR openraft::replication: error replication to target=2, error: transport error
    at /opt/rust/cargo/git/checkouts/openraft-c47b1872371ccc4e/c6fe29d/openraft/src/replication/mod.rs:669

and in 2nd\3rd nodes, keeping get

Error: timeout after 20s when receive an active leader latest: Metrics{id:2,Learner, term:1, last_log:None, last_applied:None, leader:None, membership:{log_id:0-0 membership:[{2}]}, snapshot:None, replication:

Not sure what the root cause

[BohuTANG]

@drmingdrmer

[wubx]

use v0.9.51 later

#10194

[drmingdrmer]

The last error shows that there is a DNS error

[darrenxyli]

I tried bootstrap with only 1 meta service, but it still fails though. Error message:

Error: fail to join 0 cluster via ["databend-meta-0.databend-meta.databend-meta.svc.cluster.local:28004"], caused by errors:

[drmingdrmer]

I assume the only meta-service is not node-0? This node has not yet joined into the cluster. It needs a working cluster to join into.

[darrenxyli]

it's node-0, it's only instance.

[drmingdrmer]

What are the startup command line arguments of every node?

Normally node-0 should not configured with --join argument and it should out try to join a cluster.

node-0 should start up with argument --single to initialize a cluster. So that other nodes can join to this cluster with argument --join.

https://databend.rs/doc/deploy/metasrv/metasrv-deploy