Connection failed to active manager after second manager failed to join

Description

Steps to reproduce:
1. Start cluster from manager m1
2. Join cluster from manager m2. at the middle of the join process stop the process with keyboard interrupt.
3. Wait until the managers are stable
4. Set-active m2 to become the leader-manager in the cluster
5. Try to connect back to m1 or m2

The cli fails to connect any of the managers after the join-failure process. I would expect that even after join-cluster failure, m1 will still be the active manager, which I can connect to

Activity

Show:
Sivan Barzily
January 11, 2018, 5:01 PM

it sounds like the option to make a manager active should be disabled if it failed to join the cluster

Guy Offer
January 23, 2018, 12:59 PM

Omer, please take with Lukasz to see if this is a regression and whether we want to fix.

Omer Duskin
February 11, 2018, 2:25 PM

since the action was interrupted, need to update the cli profile using one of the following:

  • `cfy cluster update-profile`

  • `cfy cluster nodes list`

Fixed

Assignee

Łukasz Maksymczuk

Reporter

Uri Wygodny

Labels

Bug Type

regression bug

Target Version

4.3

Severity

None

Epic Link

Sprint

None

Fix versions

Affects versions

Configure