etcd service configuration file is not updated after ESXi changing mgmt ip address
search cancel

etcd service configuration file is not updated after ESXi changing mgmt ip address

book

Article ID: 426021

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

  • Due to data center migration, the management IP addresses of all ESXi hosts in the cluster are changed.
  • After mgmt ip changing, etcd service configuration file ./scratch/var/clusterAgent/etcd.yml is not auto updated, the saved ip address is not current ESXi mgmt ip.

name: ########################
data-dir: /var/cache/datafiles/etcdData
initial-cluster:#########################=https://C.C.C.C:2380  <<<<< This ip address is not ESXi current mgmt ip.
initial-cluster-state: new
listen-peer-urls: https://0.0.0.0:2380
initial-advertise-peer-urls: https://C.C.C.C:2380
advertise-client-urls: https://C.C.C.C:2379
listen-client-urls: https://0.0.0.0:2379
...

IP address C.C.C.C is this ESXi's previous mgmt ip address, not current mgmt ip.

  • The etcd cluster still use ESXi old mgmt ip addresses A.A.A.A, B.B.B.B and C.C.C.C. These three IPs are not in use.

clusterAgent.log:
No(5) clusterAgent[2099412]: INFO  Creating etcd client    {"opID": "supervisor-####", "endpoints": ["A.A.A.A:2379", "B.B.B.B:2379", "C.C.C.C:2379"], "tls": true, "username": "C.C.C.C", "timeout": "30s", "incomplete-password-refresh": false}
No(5) clusterAgent[2099412]: WARN  grpc: addrConn.createTransport failed to connect to {C.C.C.C:2379  <nil> 0 <nil>}. Err :connection error: desc = "transport: Error while dialing dial tcpC.C.C.C:2379: i/o timeout". Reconnecting...
No(5) clusterAgent[2099412]: WARN  grpc: addrConn.createTransport failed to connect to {B.B.B.B:2379  <nil> 0 <nil>}. Err :connection error: desc = "transport: Error while dialing dial tcpB.B.B.B:2379: i/o timeout". Reconnecting...
No(5) clusterAgent[2099412]: WARN  grpc: addrConn.createTransport failed to connect to {A.A.A.A:2379  <nil> 0 <nil>}. Err :connection error: desc = "transport: Error while dialing dial tcpA.A.A.A:2379: i/o timeout". Reconnecting...

  • ESXi etcd service still connecting other ESXi hosts' old mgmt ip addresses. 

etcd.log:
Er(3) etcd[2099430]: publish error: etcdserver: request timed out
Wa(4) etcd[2099430]: health check for peer aadba2c1e56f4cae could not connect: dial tcp A.A.A.A:2380: i/o timeout
Wa(4) etcd[2099430]: health check for peer 59f8bc13b00ed505 could not connect: dial tcp B.B.B.B:2380: i/o timeout
Wa(4) etcd[2099430]: health check for peer 59f8bc13b00ed505 could not connect: dial tcp B.B.B.B:2380: i/o timeout

Environment

VMware vSphere ESXi

Resolution

Reinstall ESXi host fixed this issue.

Additional Information

Reference KB: https://knowledge.broadcom.com/external/article?articleNumber=385110