Cloudron Forum

Sorry, this is a native Ubuntu install. I suppose this is the wrong venue. Thanks for the heads up.

I'm setting up a private demo Element Synapse server following the instructions "Using the Installer in an Air-Gapped Environment" (https://ems-docs.element.io/books/element-server-suite-documentation-lts-2310/page/using-the-installer-in-an-air-gapped-environment](https://ems-docs.element.io/books/element-server-suite-documentation-lts-2310/page/using-the-installer-in-an-air-gapped-environment).

Following these instructions, I ended up with an installation where 19 of the pods were still trying to access the Internet for regular updates and failing. I was able to fix 18 of them using conventional tactics of modifying the replicaset, deployment, or CRD.

The problem is that I can’t get one of the pods, the one associated with synapse-haproxy, to keep from trying to access the Internet for updates and then failing. This pod has, as far as I can figure, three levels of configuration. The replicaset policy is given by the deployment. The deployment policy is given by the CRD image. In the CRD documentation, a parameter haproxyPullPolicy is said to affect this pod’s desire to download, but setting it to IfNotPresent has no apparent effect since the deployment and the replicaset continue to revert to Always no matter what (from all I’ve tried). Thus this pod ends up in an ErrImagePull state. Incidentally, setting the pullPolicy in the same CRD image also has no effect.

Any ideas?

Since this is intended as a private demo server, I can share full yaml files or anything you might like to see; just let me know.