4 ms·
If you need to connect two servers together semi-permanently, and haven't tried spiped, stop reading and do it now. It's the single most useful tool I have on
by aidenn0 6y ago
If you need to connect two servers together semi-permanently, and haven't tried spiped, stop reading and do it now. It's the single most useful tool I have on the backend.
Since Colin thinks kivaloo is more useful, but I've not heard of it, in going to look at that right now...
[edit]
I don't really have a use for a high performance, durable, KV store right now, so not convinced on kivaloo
- cperciva 6y agoHey, I didn't say that kivaloo was useful for everyone. ;-)
- zurfer 6y agoI could not find you on https://dbdb.io https://dbdb.io. Maybe you want to submit it at least there, so people will find it. How does kivaloo compare to other databases like Redis? Why should I use it? PS: fully understand that you might have just written it for fun, but if you want it to grow, this is the kind of information I am looking for as a database user. :)
- asymmetric 6y agoWhat’s the advantage over a WireGuard point-to-point connection?
- OJFord 6y agoThe most obvious thing to say is 'it's been there for the last decade'. I don't say that to preclude it from continuing to have an advantage over WireGuard, but I haven't used it (spiped) so couldn't comment.
- aidenn0 6y agoI've been using it long before wire-guard was generally available. I would guess that it's far less configuration to just "make port X on host Y look like port Z on localhost" with spiped, but I can't be sure. Things I like about it off the top of my head: - I will often start out with a single box, then say "well the DB should really be on a different machine" and just setup an spiped between them and I'm done. The actual application configuration changes not at all since it's still connecting to localhost. - My dev machines can be configured to spiped to the non-production database and I can run tests just like I would in production, since everything internal connects to ports on localhost; I literally can't accidentally mess up the production DB in a dev environment because the only place the hostname for the DB is is in the service file that starts spiped. [edit] Totally forgot the best feature: it can proxy between unix domain socket and TCP, which is something I think it had before SSH (it certainly had it before the ssh on my machine could do it). It also will re-lookup hostnames at a configurable interval, so new connections can be sent to a different machine just by configuring DNS.