In any case, let me know when you swap the image back to phpmyadmin, and I can take a closer look :)
You're right, I tested locally and the container runs in port 80 internally and exposes it in port 8080
I just change the image to phpmyadmin in kubernetes
My guess is something is preventing phpmyadmin from starting - maybe a misconfiguration?
Since you're getting it working locally, my guess is its the PMA_HOST environment variable you're using, but it's hard to say. Perhaps there is a debug mode you could enable? I don't know much about phpmyadmin, sorry :(
looks like the apache2 inside the container is stopped, I ran
service apache2 start and now is working
really very strange, I will continue to look into it, again thank you very much!
Hi @erulabs FYI I just created an issue in phpmyadmin github project in case you want to follow phpmyadmin/docker#294
Not sure if this is related to the DNS issues mentioned by Stanley Nguyen, but I can't connect anywhere from pods in kubesail. I cannot connect to any other pods within the cluster or to anywhere else on the internet. The same configuration works locally on my machine though. I had one pod that had internet connection but after restarting it with 'rollout restart deployment', it couldn't connect anymore.
The final TLDR on the networking issues is that it turns out we were not restricting in all cases how users could trigger pods to be spawned. A very clever user figured out how to spawn 30k jobs that clogged our system something awful :( We've put some limits in place to prevent this from happening again, as well as major optimizations to the clusters. Once again I apologize if you noticed any outages or issues - please let us know if we can do anything to help
Its possible that our DNS system cached the result prior to it being added
@DavidCamelo looks like, at a high level, our ClusterIssuer would need to support the DNS solver in order to enable wildcard certificates. (https://cert-manager.io/docs/configuration/acme/#adding-multiple-solver-types). Right now we only have the HTTP solver enabled for our ClusterIssuer. I'll double check with @erulabs, but it seems very possible that we can enable that for you.
Hey @DavidCamelo, I did a bit of digging, and adding DNS support is a bit more complex than I initially thought. The issue is that, because cert-manager is non-interactive, you yourself need to generate the the token that Lets Encrypt verifies, set the DNS record, and then generate the certificate request (https://cert-manager.io/docs/configuration/acme/dns01/acme-dns/ -> JSON credentials). We can make a UI to help do this, but that will have to be further down on our roadmap.
In the meantime, you could use certbot to generate the wildcard cert yourself, and then upload it to a Secret in KubeSail. Please do let me know if you run into issues / limits generating certificates for each hostname... Lets Encrypt is very generous with their limits and you shouldn't hit them under normal usage.
Hi @PastuDan, there is no problem generating the certificates, I was just asking because I thought it would be more convenient to have a single certificate for each domain and all subdomains, for example I have these two certificates for a single wordpress deploy: www.davidcamelo.com which redirects to davidcamelo.com
Hey @DavidCamelo - That does happen currently - we're planning on having our system automatically bundle certificates but the problem is it would be hard to know if one day you wanted that cert to be separate... Gets kind of tricky :) Rest assured the certificates will auto-renew automatically, so hopefully it's not much of a hassle how is it!
You can probably generate your own
Certificate object manually (https://cert-manager.io/docs/concepts/certificate/) and get what you want, but our UI will create one certificate per hostname