emqx/emqx-operator

Multiple node emqx-ee cluster will be not available for use, even if just one node error.

wooonderful opened this issue · 5 comments

wooonderful commented

Describe the bug
I have a three node emqx-ee cluster that created by emqx-opeartor which 1.2.7 version;
When just one node running failed for some reason, for example, file direction permission, My whole emqx-ee cluster will be inaccessible， for the 1883\8883 ports will be disappear in the default service.

image

To Reproduce

  1. create a multiple node emqx-ee cluster;
  2. get into one node, change the data direction permission to 400;
  3. watch the default service's change.

Expected behavior
just part of node running failed, the emqx-ee cluster should be also accessible.
the service also should be like as the blow example:
image

Anything else we need to know?:
the latest emqx-operator version also has this same problem.

Environment details::

  • Kubernetes version: 1.22.3
  • emqx-operator version: 1.2.7-ecp.1
  • Install method: e.g. helm/static manifests helm
Rory-Z commented

Hi, @wooonderful Sorry for delay, the EMQX Operator 1.2.7 is too old, could you please try 2.2.0?

wooonderful commented

Hi, @wooonderful Sorry for delay, the EMQX Operator 1.2.7 is too old, could you please try 2.2.0?

ok

Rory-Z commented

Hi @wooonderful Is your issue still there?

wooonderful commented

Hi @wooonderful Is your issue still there?

yes, it is still a problem for us until we upgrade to 2.x. But, the upgrade maybe will be done in the Q4.

Rory-Z commented

Hi @wooonderful the EMQX Operator 1.2.7-ec.2 is fixed this issue, could you please try it.
https://github.com/emqx/emqx-operator/releases/tag/1.2.7-ecp.2

I will close this issue, if you have any question, can reopen it in anytime