ingress-nginx: Health check error

Hello,

My ingress controller suddenly stopped working. This is the message that I get. I have deployed it in the past following exactly the instructions here: https://kubernetes.github.io/ingress-nginx/deploy/ Everything was working, but after I restarted kubernetes and docker it doesn’t work anymore. I tried to redeploy it but still. I am running on CentOS 7.

healthcheck error: Get http+unix://nginx-status/healthz: dial unix /tmp/nginx-status-server.sock: connect: no such file or directory

About this issue

Original URL
State: closed
Created 5 years ago
Reactions: 8
Comments: 37 (9 by maintainers)

Most upvoted comments

I’ve been facing this issue too and it got fixed after solving a couple of issues in my ingress resources. In most of the cases, I could see how there were ingress resources deployed in my cluster which had no any endpoints available or even deployed. After deleting those useless and problematic ingress resources, nginx started to start up normally.

@eljefedelrodeodeljefe Regarding your comment about how one can use this in production, I have to say that we run this component in production since 3 years and so far it hasn’t cause any outage.

mmingorance-dh on Aug 26, 2019

I am also getting this error :

healthcheck error: Get http+unix://nginx-status/healthz: dial unix /tmp/nginx-status-server.sock: i/o timeout

And after some time, it works again and issue is intermittent, and without making any change , it started working.

Any solution to overcome this?

umar1201 on Sep 9, 2019

@OGKevin yes, you can check this using the image quay.io/kubernetes-ingress-controller/nginx-ingress-controller:dev

aledbf on Sep 20, 2019

I think this issue is not being treated with enough tenacity. Honestly I wonder how one can use this in production, as even our staging tests for months keep failing on issues like this.

Restarting the controller is not a fix; It’s sysadmin patch work.

eljefedelrodeodeljefe on Aug 25, 2019

Anybody solved it? I still suffer this issue.

kwladyka on Aug 20, 2019

@ZzzJing check https://github.com/kubernetes/ingress-nginx/issues/4735

kwladyka on Apr 27, 2020

I have the same behavior with AKS 1.14.8 and nginx-controller 0.27.1 + HPA. @kwladyka

aasier on Jan 25, 2020

I still have this issue:

kubectl --context=etingroup-production get node
NAME                                  STATUS   ROLES    AGE   VERSION
gke-production-pool-1-ce587bf0-rxwq   Ready    <none>   31m   v1.14.7-gke.10


ingress version
`tag: "0.26.1"`

Is it possible it fail, because third party pod which has nginx-ingress service fail? Will nginx-ingress fail, because third party app fail?

kwladyka on Nov 1, 2019

In my case I will wait. I don’t see other choice.

kwladyka on Sep 12, 2019