There’s probably a bunch of permissions errors, filesystems warnings for cross-filesystem mounts or links, etc. all going to stderr. Linux output streams are a bit odd, 1 is stdout and 2 is stderr. So the command is redirecting the “noise” to null and just printing the actual command output. That would be my assessment, but OP could probably give a more correct answer…!

A wireless router consumes single digit watts of power. A tablet or mobile phone consumes 1/10 or less of that. Making a slice of toast and a cup of tea uses more power than both of those devices all day combined. This is another silly attempt to shift blame from corporations and billionaires down to consumers and everyday people. Don’t buy it.

It’s not perfect, but the new (2019+) mazda system is very nice. It’s all controlled by buttons and dials, zero requirement to ever touch a screen. It all feels quite thoughtfully done, especially when you compare it to fords or teslas with a big dumb laggy iPad stuck to the dash.

Local options are always better. The Mexican joint sells you a massive breakfast burrito for $6. Nepalese takeout will feed you for days for $16. Hot dog truck will fill you up with delicious processed meat for $4.

Subway? Subpar lunch made out of cardboard and ground up yoga mats for almost $20.

I wish I was the right kind of creative, greedy, and dull to come up with this kind of crap. I could scam so many bald billionaires.

Not sure I’d want to take risk management advice from the mortgage-backed securities ghouls who crashed the economy 15 years ago, but okay I guess.

I don’t need artificial intelligence in my terminal. Do you know how many times some troll has posted about “rm -fr /” on Reddit and other shitty forums, which then gets gobbled up and laundered by LLMs? Not letting that anywhere near my prod servers with valuable data.

Anecdotal… we drove through rural Ohio a few weeks ago. In several hours of travel we only saw ONE trump sign. The same place in 2016 or 2020 would have been full of them. Regardless of the impact of this, the enthusiasm is dead. There might be “maga guys” on Twitter but they’re largely disengaged in real life.

It’s crazy how the US gov basically handed him a monopoly on EV charging infrastructure, something Rockefeller could have only dreamed of, and the guy throws it away less than two weeks later in some ketamine fuelled stupor. Then has to backtrack at the cost of reputation, confidence, and sentiment. Truly another great stable genius.

The bastards can never take away your shell script full of arcane and unreadable curl commands parsed by incomprehensible awk scripts!

In my opinion it points to a more dangerous thing, “continuous delivery” software mindset seeping into safety critical systems.

It’s fine, good even, that web developers can push updates to “prod” in minutes. But imagine if some dork could push largely untested control system updates to your car’s ECU… it’s one thing for a website site to get a couple errors, but it’s a very bad thing if it makes your steering wheel stop working.

Unfinished products make more money, and it’s high time a consumer protection law clamped down on this.

It’s fine. RAID is not a backup. I’ve been running simple mirrors for many years and never lost data because I have multiple backups. Focus on offsite and resilient backups, not how many drives can fail in your primary storage device.

Not sure how to do that in docker, I’ve run mine as a plain old PHP-FPM site for years and years. It might be something that can be tweaked using config files or environment variables, or might require building a custom image.

ClamAV is slow and doesn’t catch the nastiest of malware. Its entire approach is stuck in 2008. It’s better than nothing for screening emails, but for a private file store it won’t help much considering that you’ll already have the files on your system somewhere. And most importantly, it slows down file uploads 10x and increases CPU load substantially. The only good reason to use ClamAV for nextcloud is if you will be sued if you don’t!

It needs some tweaks to be snappy. The defaults are really bad.

• change database from SQLite to a proper database like MySQL or Postgres, and configure the database server to use your memory fully
• increase the PHP memory limit from the default (128M on many distros) to >1G, the more the better
• install APCu in-memory cache for PHP
• add Redis as additional cache
• turn off the antivirus extension, if installed (ClamAV is useless)
• use http/2 on Apache/nginx to increase performance with multiple connections

https://docbot.onetwoseven.one/services/nextcloud/

They’re not going to jail for you. Never assume a service provider will put themselves at risk on your behalf.

This was my setup from about four years ago. Other than moving suricata elsewhere, it’s largely the same. Worth a shot if it’s something you’re into!

https://nbailey.ca/post/linux-firewall-ids/

OpenBSD is also great, I’m just more familiar with the Linux tools. All the required tools are in the base image, and they have a great official guide:

https://www.openbsd.org/faq/pf/example1.html

Yep. Firewall, routing, dhcp, dns, everything you’d expect from a gateway device. Plain Debian (or really any distro) can do it all. With a 1gbps bi-directional connection fully saturated it will run at about 10% cpu on my very crappy low power Celeron CPU.

Plus, there’s no web UI full of janky and insecure CGI scripts to exploit, and software updates are forever (well, until x64 is deprecated, so basically forever).

IPtables on Debian because I like my life to be boring and unchanging.

For about a year I was running a full out of band IPS on my network. My core switch was set up with port mirroring to spit out a copy of all traffic on one port so that my Suricata server could analyze it. Then, this was fed into ElasticSearch and a bunch of big data crap looked for anomalies.

It was cool. Basically useless because all it did was complain about the same IP crawler bots as my nginx logs. But fun to setup and ultimately good for my career lol.

Not an arch user, but it’s possible they moved dbus to a user scoped unit now. Might be possible to start it like this (or something similar)

systemctl —user start dbus.service