- a wall mount 6U rack which has:
- 1U patch panel
- 1U switch
- 2U UPS
- 1U server, intel atom 4G ram, debian (used to be main. now just has prometheus)
- a mini ryzon 16 core 64G ram, fedora (new main)
- multiple docker services hosted.
- synology nas with 4 2TB drives
- turris omnia WRT router -> fiber uplink
network is a mix of wireguard, zerotier.
- wireguard to my external vms hosted in various global regions.
- this allows me ingress since my ISP has me behind CG-NAT
- zerotier is more for devices for transparent vpn into my network
i use ssh and remote desktop to get in and about. typically via zerotier vpn. I have one of my VMs with ssh on a backup port for break glass to get back into the network if needed.
everything has ipv6 though my ISP does not provide it. I have to tunnel it in from my VMs.
- a wall mount 6U rack which has:
- 1U patch panel
- 1U switch
- 2U UPS
- 1U server, intel atom 4G ram, debian (used to be main. now just has prometheus)
- a mini ryzon 16 core 64G ram, fedora (new main)
- multiple docker services hosted.
- synology nas with 4 2TB drives
- turris omnia WRT router -> fiber uplink
network is a mix of wireguard, zerotier.
- wireguard to my external vms hosted in various global regions.
- this allows me ingress since my ISP has me behind CG-NAT
- zerotier is more for devices for transparent vpn into my network
i use ssh and remote desktop to get in and about. typically via zerotier vpn. I have one of my VMs with ssh on a backup port for break glass to get back into the network if needed.
everything has ipv6 though my ISP does not provide it. I have to tunnel it in from my VMs.
> the keyboard was amazing
Was it? It’s quite a bit “mushy” in my opinion. 😅
It’s funny how tiny the touchpad is by today’s standards.
> the keyboard was amazing
Was it? It’s quite a bit “mushy” in my opinion. 😅
It’s funny how tiny the touchpad is by today’s standards.
> the keyboard was amazing
Was it? It’s quite a bit “mushy” in my opinion. 😅
It’s funny how tiny the touchpad is by today’s standards.
(In the case of the Rust installer, I still wonder why they go through the trouble of having a shell script (POSIX, portable, even runs on Windows apparently), when all it does is download a binary and run that. Is that super useful to people, yeah? I’m sure there’s some reason, I just don’t see it.)
(In the case of the Rust installer, I still wonder why they go through the trouble of having a shell script (POSIX, portable, even runs on Windows apparently), when all it does is download a binary and run that. Is that super useful to people, yeah? I’m sure there’s some reason, I just don’t see it.)
(In the case of the Rust installer, I still wonder why they go through the trouble of having a shell script (POSIX, portable, even runs on Windows apparently), when all it does is download a binary and run that. Is that super useful to people, yeah? I’m sure there’s some reason, I just don’t see it.)
I wondered how a client would figure out the endpoint where to POST to.
1. In a call with exactly two participants the "View" menu doesn't do anything anymore. I cannot focus on the content of the screenshare and always have the silly screen space wasted on the right with a giant, useless other person's profile picture. As soon as a third participant is in the call, the "View" menu works again. For months now. You can't even make it the default in the settings.
2. Over the last couple of weeks screenshares seem to get delayed for up to 20 seconds sometimes. I never experienced that before. This makes pair programming or diagnosing stuff very hard and way more time consuming than it should be.
3. I somehow never find the chat box. With the old Linux client that was no problem, but since they moved it to the top, it always takes me several seconds to open it.
4. Sometimes the first call in the morning ends up in total silence so I have to restart Chromium. It then works.
5. On live events I have to completely remove all the cookies and login again, because I get the error message that I have to accept third party cookies. Even if the ten domains or so are explicitly whitelisted or *all* third party cookies are accepted. Always get the error. Each and every time.
[](https://movq.de/v/f7b7391d37/helvete%2D1.jpg)
[](https://movq.de/v/f7b7391d37/helvete%2D1.jpg)
[](https://movq.de/v/f7b7391d37/helvete%2D1.jpg)
@movq I just don't want to run such crapware. Browser, mail client and video player aside, I think I don't do too bad on that regard with my private stuff. Yeah, definitely ignoring the situation at the dayjob.
@prologic Only for Rust. Otherwise I stay away from that for sure.
But yeah, I absolutely get your point as well, @bender. I also do not mind long messages over here. So I support you in increasing message length limits. :-)
Gonna buy more of them later on as well.
It's just a hobby for me, something to do, and I always enjoy getting various hardware related things. Especially open source stuff.
http://darch.dk/wf-lookup.php
http://darch.dk/wf-lookup.php
http://darch.dk/wf-lookup.php
http://darch.dk/wf-lookup.php
curl foo | sh is basically equivalent to running precompiled binaries or the huge dependency mess that we have these days (simple programs pulling in 47289 libraries). We run completely untrusted code all the time and nobody cares anymore. The idea of eliminating distributions (which at least provide *some* layer of quality control) pops up again and again. A curl foo | sh is probably the *least* harmful thing these days, because it’s the easiest issue to fix.(Meh: Rust’s
curl https://sh.rustup.rs | sh downloads a 15 MB binary that does god-knows-what.)Or am I missing the point? 🤔
curl foo | sh is basically equivalent to running precompiled binaries or the huge dependency mess that we have these days (simple programs pulling in 47289 libraries). We run completely untrusted code all the time and nobody cares anymore. The idea of eliminating distributions (which at least provide *some* layer of quality control) pops up again and again. A curl foo | sh is probably the *least* harmful thing these days, because it’s the easiest issue to fix.(Meh: Rust’s
curl https://sh.rustup.rs | sh downloads a 15 MB binary that does god-knows-what.)Or am I missing the point? 🤔
curl foo | sh is basically equivalent to running precompiled binaries or the huge dependency mess that we have these days (simple programs pulling in 47289 libraries). We run completely untrusted code all the time and nobody cares anymore. The idea of eliminating distributions (which at least provide *some* layer of quality control) pops up again and again. A curl foo | sh is probably the *least* harmful thing these days, because it’s the easiest issue to fix.(Meh: Rust’s
curl https://sh.rustup.rs | sh downloads a 15 MB binary that does god-knows-what.)Or am I missing the point? 🤔
I truly shouldn't complain. I don't participate often enough to ask for anything. :-D Instead, I will adapt.
- I maintain a small "Mini DC" comprised of 22RU cabinet 600mm deep.
- This houses 3x 1RU Xeon machines + 1 RU 10 3.5" + 4 2.5" NAS + 4RU UPS + 1RU 24-port Gbps Switch/Router + 1RU Tray in the middle + 1RU patch panel at the top.
- This is now hooked up to 250/100 Mbps Fibre 😅
- I run Proxmov VE on the 3x Hypervisor machines. They run a dozen or so Virtual Machines.
- I run a couple of Docker Swarm clusters on those machines, running BurmillaOS (a fork of RancherOS).
- I just use the local LAN network to SSH into machines, but each physical machine also has an IPMI management interface too for when things go wrong (rarely).
- I run so many services I can't being to list them here. But it's in the order of ~50-60 unique services. Some of which you're familiar with as many are public facing, some are internal and others are locked down behind auth.~
- I maintain a small "Mini DC" comprised of 22RU cabinet 600mm deep.
- This houses 3x 1RU Xeon machines + 1 RU 10 3.5" + 4 2.5" NAS + 4RU UPS + 1RU 24-port Gbps Switch/Router + 1RU Tray in the middle + 1RU patch panel at the top.
- This is now hooked up to 250/100 Mbps Fibre 😅
- I run Proxmov VE on the 3x Hypervisor machines. They run a dozen or so Virtual Machines.
- I run a couple of Docker Swarm clusters on those machines, running BurmillaOS (a fork of RancherOS).
- I just use the local LAN network to SSH into machines, but each physical machine also has an IPMI management interface too for when things go wrong (rarely).
- I run so many services I can't being to list them here. But it's in the order of ~50-60 unique services. Some of which you're familiar with as many are public facing, some are internal and others are locked down behind auth.~
- I maintain a small "Mini DC" comprised of 22RU cabinet 600mm deep.
- This houses 3x 1RU Xeon machines + 1 RU 10 3.5" + 4 2.5" NAS + 4RU UPS + 1RU 24-port Gbps Switch/Router + 1RU Tray in the middle + 1RU patch panel at the top.
- This is now hooked up to 250/100 Mbps Fibre 😅
- I run Proxmov VE on the 3x Hypervisor machines. They run a dozen or so Virtual Machines.
- I run a couple of Docker Swarm clusters on those machines, running BurmillaOS (a fork of RancherOS).
- I just use the local LAN network to SSH into machines, but each physical machine also has an IPMI management interface too for when things go wrong (rarely).
- I run so many services I can't being to list them here. But it's in the order of ~50-60 unique services. Some of which you're familiar with as many are public facing, some are internal and others are locked down behind auth.~
I was actually positively surprised that after the outlined rustup upgrade oneliner above, running
make in Newsboat again worked flawlessly. Nothing else required. I delayed rebuilding for quite some time because I thought getting this Rust toolchain sorted out is going to be a major endeavor. Luckily, I was wrong. :-)I just don't know if I now have two Rust installations in parallel or not. Or how much disk space I waste with all this. At least the script didn't tell me it found an old installation. It printed heaps of stuff, but skimming over it, I didn't see anything like that. I then simply selected the regular install. Whatever that meant. Researching this topic will be a project for another day if I'm really bored.
rustup. 🙄 I guess the same is true for Newsboat?Hmm, no, not really. Newsboat wants “Rust Edition 2021”, which is supported since Version 1.56. That’s already “ancient”. 🤣
rustup. 🙄 I guess the same is true for Newsboat?Hmm, no, not really. Newsboat wants “Rust Edition 2021”, which is supported since Version 1.56. That’s already “ancient”. 🤣
rustup. 🙄 I guess the same is true for Newsboat?Hmm, no, not really. Newsboat wants “Rust Edition 2021”, which is supported since Version 1.56. That’s already “ancient”. 🤣
https://www.rust-lang.org/tools/install recommends this very dangerous and fishy thing:
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
TLS 1.2 certainly fits the rusty motto.=