[Bug] rpc error: code = AlreadyExists desc = banyandb: resource already exists"

[lifehacking]

Search before asking

• I had searched in the issues and found no similar issues.

Apache SkyWalking Component

BanyanDB (apache/skywalking-banyandb)

What happened

banyandb version :0.8.0
banyandb type: cluster mode
banyand data、banyand liaison deploy on diff hosts。when restart banyand data，log warn "error":"key /banyandb/nodes/192.168.1.6:17912 ， rpc error: code = AlreadyExists desc = banyandb: resource already exists。

banyand data start cmd：banyand data
--etcd-endpoints=192.168.1.2:2379,192.168.1.3:2379,192.168.1.4:2379
--etcd-username=root
--etcd-password=Root123
--node-host-provider=ip
--measure-root-path=/opt/banyandb/data/measure
--stream-root-path=/opt/banyandb/data/stream

What you expected to happen

after restart data node，banyand can listend

How to reproduce

etcd del /banyandb/nodes/192.168.1.6:17912

Anything else

No response

Are you willing to submit a pull request to fix on your own?

• Yes I am willing to submit a pull request on my own!

Code of Conduct

• I agree to follow this project's Code of Conduct

[hanahmily]

Thank you for bringing this up.

The lease for node information has a 5-second TTL. After this period, etcd will clean up the node. The data node will wait until the previous node's info is removed. However, during this waiting loop, the errors you mentioned may occur.

[lifehacking]

"etcd will clean up the node"，need etcd config some option？when restart data node，after 5-second TTL node's info in etcd not removed，need manual del ，after del node info，data node process can listen on 17912 and 17913 port

[loxehate]

I encountered the same issue. I started a single node using only banyand data, and it successfully registered with etcd, with the lease also being generated correctly. After stopping banyand data, the lease was deleted, but the information in etcd still remained. Even after waiting for 15 seconds, it was still there.