Worker node can't access SSL cert from the server.

[danilang]

Howdy.  

I've come across an interesting one.  We have dual server environment with scheduler/worker and another worker node.  The worker node has this error message in the logs when the service restarts.

2025-02-05 18:23:22.517000,ERROR,5488,AlteryxService,Main,,svcacctalteryxdev,,,,,,"Failed to retreive SSL certificates with error code <2>. Please contact your System Administrator to ensure a port <443> is configured with an SSL Certificate: Win32 Error: The operation completed successfully. (0)"

During the configuration phase, the test button comes back compatible

The certificate on the scheduling server is valid and we can https:// connect to the gallery through the browser, even on the worker node!

Any thoughts?

Dan

[gawa]

@danilang It's mysterious issue but can be routing problem by any chance?

If your scheduling server is public to internet through reverse proxy, client will refer to SSL certificate issued to reverse proxy by authorized CA.

Issue might occur when worker node host tries to communicate with scheduling server directly inside private network, since usually self-signed  certificate or certificate issued by private CA(untrusted as default) is applied. Unless those CA is trusted intentionally, communication fails.

Workaround is to add issuer of certificate for scheduling server host as trusted CA.

[danilang]

Thanks @gawa 

I'll look into it.

Dan

[danilang · Accepted solution]

So I finally solved it

S is the scheduler.  W is the original worker that could not connect to S.  T is a new server to run some tests

1. I set T in worker mode
2. I connected T to S with no problem
3. I convert the W to a full server and connected T to it as a worker with no problem.
4. On a hunch, I converted W back to worker and connected to S.  Now the issue was gone. 

Who knew?

Dan