That did the trick. Thanks!

Error installing app-lets-encrypt: no installation candidate for certbot

I am finally getting around to getting a non-self-signed cert so I tried to install the Let's Encrypt app. Both the marketplace app and yum throw an error when attempting to install, about there being no way to resolve the certbot dependency:



This was extremely confusing and my only thought is the SSL error for the marketplace repo. In response I reset the system clock via ntp, cleared the yum cache, and tried again. Same error. If anyone has any idea what's going on, I'm all ears.

Nick Howitt wrote:

Have you got a third part repo enabled?

That seems to be it. The error was pretty misleading, as you said the issue was one of the third-party repos had stopped resolving and broke the whole thing.

Thanks!

Yum fails with curl error: "Peer's certificate issuer has been marked as not trusted by the user."

So based on the guidance for updates that was sent out, I went and looked at my system's update history and found the last update was in January. That isn't necessarily a problem but I would like antispam updates, so I opened a session on the server and checked for antispam updates. Yum failed immediately:



The error suggests disabling the repo, but I never got to the repo, it failed out looking for a mirror. Not sure what could even be causing this.

Re: Server seemingly ignoring openvpn connections

Figured it out. Networkmanager requires the auth mode to be TLS with user, not just TLS and key password. It will not prompt the user for a login (at least in gnome shell) but instead just time out. This is contra the official clearos documentation, so someone might want to look into that. The doc also has instructions for a deprecated version of the windows client, which uses a different import and setup method.

Figured it out. Networkmanager requires the auth mode to be TLS with user, not just TLS and key password. It will not prompt the user for a login (at least in gnome shell) but instead just time out. This is contra the official clearos documentation, so someone might want to look into that. The doc also has instructions for a deprecated version of the windows client, which uses a different import and setup method.

Road Warrior setup. The syslog is from the client, the tcpdump is from the server, as is the (empty) message log.

I got my hands on a windows machine and was able to connect just fine via the openvpn client on Windows 10. This seems to be a networkmanager issue.

Server seemingly ignoring connections

I have been trying to set up openvpn through clearos's tools but every attempted connection times out:

syslog:


Connectivity is fine with everything else (SSH, HTTPS) and I checked, both UDP and TCP were automatically added to the incoming firewall. I tried connecting with the client both with the firewall on and allow rules and with the firewall (temporarily, obviously) down. Still nothing. So my next thought was dropped packets

tcpdump -ni enp3s0 udp and port 1194


Nothing there. Considering this is a stock clearos openvpn setup and the client is using the imported .ovpn I am at a complete loss.

Re: Zarafa throwing sql errors

Figured it out; there were a bunch of core dumps sitting around from the last time zarafa acted up, and the system slowly strangled itself with the reduced space.

Incidentally, while SQL isn't on root, it does require space on root for transactions.