Open
Description
launch a cluster with these --starter.port
's :
9528
9628
9728
and stop all instances.
go into the respective directories, and re-launch the starter without any parameters again. The cluster will now bind:
tcp6 0 0 :::8528 :::* LISTEN 378399/arangodb
tcp6 0 0 :::9529 :::* LISTEN 378598/arangod
tcp6 0 0 :::9530 :::* LISTEN 378588/arangod
tcp6 0 0 :::9531 :::* LISTEN 378422/arangod
tcp6 0 0 :::9629 :::* LISTEN 378801/arangod
tcp6 0 0 :::9630 :::* LISTEN 378790/arangod
tcp6 0 0 :::9631 :::* LISTEN 378634/arangod
tcp6 0 0 :::9729 :::* LISTEN 379007/arangod
tcp6 0 0 :::9730 :::* LISTEN 378997/arangod
tcp6 0 0 :::9731 :::* LISTEN 378843/arangod
So, starter instance 1 will not bind port 9528
but port 8528
instead.
Starter instance 2 & 3 will continue to groan about not being able to connect the first starter:
2020-10-05T16:42:32+02:00 |WARN| Failed to load cluster configuration from http://localhost:9528/ component=arangodb error="Get http://localhost:9528/hello?update=1: dial tcp [::1]:9528: connect: connection refused"
The processes will re-launch and work as expected.
This happenes, though the setup.json
contains the proper port to bind:
{
"version": "0.2.2",
"id": "98c52ca1",
"peers": {
"Peers": [
{
"ID": "98c52ca1",
"Address": "localhost",
"Port": 9528,
"PortOffset": 0,
"DataDir": "/tmp/CLUSTER/node1",
"HasAgent": true,
"IsSecure": false
},
Metadata
Metadata
Assignees
Labels
No labels