I connected from a Linux box to my repo server with kopia repository connect server --url https://kopia:443 and it worked without any issues.
So here are my 2 questions:
why would I have to specify --server-cert-fingerprint, if it worked without it? so when is this fingerprint needed?
after rebooting, the repository was still connected. how? I never created a systemd unit file, nor did I add a script that runs at startup. this is great. I just don’t understand how this is possible.
P.S.: The documentation is also wrong: It states:
$ kopia repository status
kopia: error: operation supported only on direct repository, try --help
This is not the case. I am getting the following on my client:
# kopia repository status
Config file: /root/.config/kopia/repository.config
Description: API Server: https://kopia:443
Hostname: cator01ps
Username: root
Read-only: false
Format blob cache: 15m0s
I think that using a secured connection is the default with Kopia and you must specifically chose to not use a secure connection on the repo server.
As for your 2nd question, always remember, that Kopia (remote client) doesn’t pertain a persistent connection to the remote repo. Being “connected” means, that there is a valid repository config in place. The KC usually stores this in ~/.config/kopia on UNIX-like systems, unless you specifiy it otherwise. A reboot won’t remove these files, so Kopia Client can still connect to the remote repo after a reboot. If you want to invalidate the connection, user kopia repo disconnect, or remove the config files.
Thanks for the reply, but I don’t understand what your statement has to do with my question:
I am creating a secure connection, since I am connecting to https:// so I am still not sure, why or when I would need the fingerprint. Because I didn’t specify it, despite it always been mentioned in all docs, but it worked without it.
The fingerprint is only needed for self-signed certs, whose CAs are not in the operating system’s trust store.
Kopia is a great piece of software and I am impressed how nicely it works. But the documentation is probably one of the worst I have ever seen. It’s outdated, inconsistent, the command line reference only shows a verbatim copy of the --help page, but no explanations whatsoever. (e.g. what values can I use for an argument)
The only way to figure out how Kopia really works is to spend hours and hours for trial and error.
Yep, I think this project might be dead. The dev doesn’t merge any PRs and the only commits in the past months are dependency updates.
I will monitor this project for another year. If nothing changes, I’ll switch to something that is actively developed.
The same - thought initially that it can be my fav backup software - but since then I changed mind. However good initial design and implementation without ongoing dev it rots. For backups I need something solid - does not even have to be funky.
I would add it to the doc, if there was any indication that this project wasn’t dead. I am not going to waste my time beating a dead horse. (see my previous comment)
I think the community needs assurance what is happening with this project. I am already looking for alternatives, even though I really like kopia. But there is no use in entrusting my data to an unmaintained product.