ESP startup error

[pratap....@slalom.com]

Hi,

While starting an ESP container with the following startup options I see the following errors via the container logs.

--service=echo-api.endpoints.myproject-id.cloud.goog 

--rollout_strategy=managed 

--backend=echo:8080 

--healthz=/

I am also trying to configure ESP httpHealthCheck to Internal Load Balancer on port 80 to "/" PATH using healthz startup option. It doesn't work whereas tcpHealthCheck works totally fine.

Help/support very much appreciated.

[pratap....@slalom.com]

Anyone facing similar issues? Appreciate if anyone can share their experience. 

[qiwz...@google.com]

Unfortunately,  "/" is not supported for --healthz path.   Root "/" path is used for routing normal requests. 

You have to specify a path not conflicting with your normal requests. 

[qiwz...@google.com]

https://github.com/cloudendpoints/esp/pull/540 will update the doc.

[Pratap Sisodia]

Thanks for the reply and clarification about –healthz options.

 

Looks like our ESP that sits on a VM with private IP ONLY is not able to reach out to Google Service Management (see errors in the screenshot).

 

All the required APIs are enabled, Networking egress rules opened and Private Google Access enabled too.

 

What else it could we that I am missing?

 

 

 

--
You received this message because you are subscribed to the Google Groups "Google Cloud Endpoints" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-cloud-endp...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-cloud-endpoints/ef30ca03-6300-42c4-ba81-7affc90775e5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Wayne Zhang]

It is DSN server issue.  By default, ESP is using Google DNS sever 8.8.8.8

In some cases, it could not resolve these domain names.  You can try to use metadata server 169.254.169.254

By using --dns=169.254.169.254 in the start esp flag

I am not sure if it will work

[pratap....@slalom.com]

On Sunday, February 10, 2019 at 5:09:55 PM UTC-6, pratap....@slalom.com wrote:

[pratap....@slalom.com]

For ESP health check startup options, I am using the following actual command lines. This is deployed on a Managed Instance Group (MIG)  frontend with an Internal Load Balancer (private IP). When I use HTTP type health check, the MIG backend never shows as healthy however with TCP type health check, it does.

Why does HTTP type health check to ESP /healthz end point NOT work on private IP ILB?

docker network create --driver bridge esp_network

docker run --detach --name=echo --net=esp_network gcr.io/google-samples/echo-java:1.0

docker run --name=esp --detach --publish=80:8080 --net=esp_network gcr.io/endpoints-release/endpoints-runtime:1 --service=echo-api.endpoints.sdrs-server.cloud.goog --rollout_strategy=managed --healthz=/healthz --backend=echo:8080

To unsubscribe from this group and stop receiving emails from it, send an email to google-cloud-endpoints+unsub...@googlegroups.com.

[pratap....@slalom.com]

On Sunday, February 10, 2019 at 5:09:55 PM UTC-6, pratap....@slalom.com wrote:

[qiwz...@google.com]

You are debugging a network issue. Sorry,  I could not help.  

Thanks

-Wayne