Monkey With A Shell

Was the prior drive set in some kind of raid set or just individuals, and what are the old drive capacity vs new?

I guess it depends a lot on what your doing with the server. If it’s pure data store I would just boot off a USB and give yourself all the data space since it’s quite likely all running in ram anyhow.

If you run apps out of it and need the M2 for swap and rapid cache storage the fastest would likely be make a 2 drive zpool, copy a single to it, and repeat as needed until you have it all copied over, then add the 3rd to the zpool

Wikipedia though has a strong reputation for being well cited and due in large part to the huge user base glaring inaccuracies get corrected quickly. I saw a study at one point comparing them to a traditional encyclopedia and they had of course faster shifting errors, but on average where pretty on par for accuracy.

A federated system where I or any other knucklehead could put up an instance isn’t going to have that ‘checked by 1000 eyes’ factor going so much, or if it did ever get to that point then they’ve likely become the defacto ‘real’ federated encyclopedia and the others inherently suspect.

All in all It’s a neat idea, but sounds like it’d be rife with chaotic discord. As a general thing if something is on the standard Wikipedia I can be pretty sure it’s reasonably accurate without having to research who posted it, and I can torrent a copy of the whole thing as I just recently found out.

Generally yeah, the same could be said for Lemmy and their communities. The challenge is when you have federated systems like this you have to largely take it as good faith that instance owners will keep reasonably updated software and good practices so you don’t end up with a pile of spam edits.

With communal wiki type systems as a whole you end up with the question of credibility. Some people would cite only well researched and validated studies, and some people whole heatedly believe that a religious text was written by divine hand and this must be true. How do you reconcile those two without giving weight to things that are patently nonsense, aka you must teach the gospel of the flying spaghetti monster to be fair to all?

Yeah, that’s the part that I was referencing to. If you have a topic called ‘the sky is’ on multiple instances, and they all say a different color, it rather defeats the purpose of an encyclopedia as some sort of source of truth if you have to pick which one is right.

That could be, I was thinking if the pages were more like the communities. I would have to think they need to be for any kind of moderation, otherwise who approves edits or has edit permissions? If someone else doesn’t agree with the vision on the existing page/thread what stops someone from putting up an alternate version?

Missing an /s?

Take a look at the ‘news’ on various instances like hexbear vs world and it’s night and day. An encyclopedia is meant to be factually reliable, but if this works like it does here you would have the equivalent of conservapedia and prolewiki sitting side by side as ‘true’.

Probably a lot more stats on it too if you look at the county recorder/assessor offices. Used to do document recording work a couple decades back and the amount of info you can get off public records for free is kind of nuts.

This sounds like a way to have 13 different ‘realities’ for any given topic, depending on which instance it was hosted on.

The guy beat you to it in 1996

What you might call a stateful NAT is really a 1-1 NAT, anything going out picks up an IP and anything retuned to that IP is routed back to the single address behind the NAT. Most home users a many to one source nat so their internal devices pick up a routable IP and multiple connections to a given dest are tracked by a source port map to route return traffic to the appropriate internal host.

Basically yes to what you said, but a port forward technically is a route map inbound to a mapped IP. You could have an ACL or firewall rule to control access to the NAT but in itself the forward isn’t a true firewall allow.

Same basic result but if you trace a packet into a router without a port forward it’ll be dropped before egress rather than being truly blocked. I think where some of the contention lies is that routing between private nets you have something like:

0.0.0.0/0 > 192.168.1.1 10.0.0.0/8 > 192.168.2.1

The more specific route would send everything for 10.x to the .2 route and it would be relayed as the routing tables dictate from that device. So a NAT in that case isn’t a filter.

From a routable address to non-route 1918 address as most would have from outside in though you can’t make that jump without a map (forward) into the local subnet.

So maybe more appropriate to say a NAT ‘can’ act as a firewall, but only by virtue of losing the route rather than blocking it.

NAT in the sense used when people talk about at home is a source nat, or as we like to call it in the office space a hide address, everyone going to the adjacent net appears to be the same source IP and the system maintains a table of connections to correlate return traffic to.

The other direction though, if you where on that upstream net and tried to target traffic towards the SNAT address above the router has no idea where to send it to unless there’s a map to designate where incoming connections need to be sent on the other side of the NAT so it ends up being dropped. I suppose in theory it could try and send it to everyone in the local side net, but if you get multiple responses everything is going to get hosed up.

So from the perspective of session state initiation it can act as a firewall since without route maps it only will work from one side.

Assuming it’s not a 1-1 NAT it does make for a functional unidirectional firewall. Now, a pure router in the sense of simply offering a gateway to another subnet doesn’t do much, but the typical home router as most people think of it is creating a snat for multiple devices to reach out to the internet and without port forwarding effectively blocks off traffic from the outside in.

The model here needs tuning, it hasn’t managed to mimick coherent human language yet.

This is part of the annoyance of Nix as a desktop though. With windows you have 64bit and (for whatever reason) x86 versions of apps and it’s generally just assumed to work with what your running, unless you have an antique with win98 or something.

With Nix there are a a whole pile of possible variables and ways to install things. Particularly with people getting so used to phone/tablet app stores the need for easy install, use, removal is needed for mass adoption. Nobody wants to create folder structures and set environment variables to use some app.

For server hosting it’s the only way to go.

Gaming has improved significantly, although it’s rather frustrating that it’s by all these compatibility layers and such rather than native run.

For desktop, as a workstation and general purpose it’s ‘ok’ with rough edges. Things like (limited tests with a couple common distros like Ubuntu/Mint/Bazzite) the nextcloud app not supporting virtual files that have been available for a while in Windows and domain auth being twitchy where I’ve tried.

For the end user a big part is being able to just find an app and use it, no compiling or tweaking of settings needed for it to do what’s expected. Package managers help greatly, but with the huge number of distros out there it makes it really hit and miss to say just go for it. The relatively few times you can just download a Linux version of an app from a site (as people are prone to doing if they go read about something on the web) you often would have to go chmod +x it and quite possibly have to run it from a CLI rather than just click the downloaded app.

So usable yes, but in a place where I could just drop it on someone and say go to town less so…

Sweet, will have to take a further look at that. Had set it up to check the state of Nix games and it seems to be pretty decent (last I really looked was when StarCraft was still a big thing). Didn’t see anything in the settings about ldap/ad auth in my messing about though.

Largely because half the services I host are tied to a Univention DC, and so are the current Windows client machines, I’d like to maintain that state.

I know Ubuntu has an AD option when you set it up, but it doesn’t seem to work with any of the ‘AD compatible’ replacements I’ve tried.

Big thing holding me up is domain auth. Someone had decent luck getting something like Bazzite joined to a DC without major manual work?

867-5309 (just a guess)

My most recent one didn’t have a number. Maybe just someone pointing out open sign up instances to banish? Haven’t bothered to look too far.

Yeah, was kind of clunky worded way to say ‘no ID’, /. Has screen names but in chan world everyone is the same ‘anonymous’. When everyone is faceless and nameless people tend to let the ugly out with ease.

edit: I’m not a regular of /. but it looks like there are ‘Anonymous Coward’ responses, but they’re more the exception than the rule. I’m sure the mods there have some measure of trolls to deal with.