2023-10-02 17:05 from IGnatius T Foobar <ajc@citadel.org>
It's not called cancer; it's now called GNU/Cancer because everything
rms touches is part of the GNU project, because he says so.
There is a certain degree of irony, GNU licensing being the cancer that it is.
lol
Ubuntu latest release had to be pulled since they didnt bother to review, and some pin-head snuck in a bunch of obscene Ukrainian translations,.
lol
I'm all for a good prank, even if this one was in poor taste. I wonder if the thought police will go as far as to try to edit the commit history. That'll go very poorly.
In case they do, here are the English translations of the commit in question.
Not very creative, actually.
"offlineWarning": "Your pants are still off",
"fullInstallationSubtitle": "A selection of offline videos of the execution of Palestinian children.",
"minimalInstallationTitle": "Classic Gay Sex",
"minimalInstallationSubtitle": "Just the bare essentials, circumcised eldaks and Jewish pornography.",
"installationTypeTPMWarning": "<font color=\"{color}\">Warning:</font> This feature is not supported by your synagogue and cannot support updates to future versions of the Podor system. Please remove <a href= \"{url}\">pants</a>.",
"installationTypeTPMSelected": "It's not that difficult, just put on and take off your pants",
"installationTypeTPM": "Experimental Ancient Hebrew Encryption",
"fullInstallationTitle": "Full Syphilis Infection",
"rstHeader": "Turn off RST, spread your buttocks, and continue",
Yeah. I'd have tried to be a bit more subtle. Have more creative messages and bury them in a much larger useful commit.
2023-09-30 19:09 from Nurb432
Didn't read all the details but seems RMS is dying of cancer. Wont be
here much longer.
WHat I hear suggests he has about 5 years left.
2023-10-15 16:10 from Nurb432
Geez, the way it was described it was like end of world insults..
that is borderline lame.
It is lame, without the "borderline" part.
WHat I hear suggests he has about 5 years left.
Well then, he ought to spend those five years getting prepared to move on, instead of being a monumental jackass like he has for his entire life.
But now it's been what ... 20 years? And people love this thing.
Yes its been around forever. They do commercial OSS the right way.
"here, use it, its yours.. If want guaranteed support, pay us"
Thu Nov 23 2023 15:39:25 EST from IGnatius T FoobarI keep not doing Proxmox because it seems too good to be true. I feel like if I start relying on it, it will be discontinued, so I keep going with bare Linux and putting all the other stuff on top of it.
But now it's been what ... 20 years? And people love this thing.
2023-11-23 15:39 from IGnatius T Foobar
I keep not doing Proxmox because it seems too good to be true. I feel
like if I start relying on it, it will be discontinued, so I keep going
with bare Linux and putting all the other stuff on top of it.
But now it's been what ... 20 years? And people love this thing.
It is not super-duper groundbreaking, but it has a big advantage in that it places all your junk behind a single control panel. You may have a bunch of nodes runing independent virtual machines and LXC containers and still manage them all from the same interface. That is kind of cool because the setup is quite fast for what you get.
Also, their starred backup solution is quite complete for the price. It reminds me of a Restic for virtual machines.
The downsides? No nativeblock-device encryption, no automated backups for the host configuration, and the startup window nagging you to purchase support every time you log in. Also no native UPS integration. You can solve all of those manually. Oh, also the firewall subsystem sucks because changes to the firewall in a VM require you to restart the VM.
The new setup, assuming I can get my hands on the machines, will be three nodes, so I've got all sorts of options. I want to do something that will take advantage of the fact that there will be three. For sure there will be a Kubernetes cluster. I am particularly fond of MicroK8S. If you configure at least three nodes, it automatically sets up High Availability for you.
Then there's storage ... I want to do something that takes advantage of the multiple nodes and their local disks. That might take the form of a Ceph cluster, or I might do OpenEBS.
I actually won't have much of a use for virtual machines. At the moment, I am using VMs for my VPN router, my web server, and the machine that runs Uncensored. (hold on a minute ... cat jumped up on the desk and wants kisses ... ok, back to the post.) I intend to run Uncensored in a container, but to do that I first have to convert my database to 64-bit and the machine I'm running it on now is way too slow to do that in any reasonable amount of time.
I've already got a bunch of stuff running in containers. One of my development machines is a container. GitHub is running in a container. Eventually, the production instance of Uncensored will run in a container.
So at this point I'm still not sure about using ProxMox because it locks you into a specific way of doing things.
If you cluster with shared storage, as really intended, you dont really need backups of local server config, and it can do automated backups automatically. ( even without the backup solution they have ) . A single nag box is trivial really.. its not like it keeps nagging you as you work or anything else to get in the way, or "this critical piece of functionality to make it worth while is an Enterprise feature" crap that so many other fake OSS projects have... its just at login. And people have removed it. There is a way.
And ya, its not perfect ( but what is ? ) and you could do everything by hand.. but why?
Once nice thing about their backup serer instead of shared storage, is you can use it to move things between clusters almost pain free. Just disconnect, take it to your other location, attach to the other cluster. poof, backups of vms + their configs are there to restore. Great for total disaster if you dont want to link your cluster across sites.
Ya, im a fan, i fully admit that. I was using that OSS version of the citrix xen stuff, until they started having issues when citrix started to cut off features downstream that were open ( really? ass-hats ).. Ran across PVE as i was looking for alternatives short of doing it by hand, tried it, and loved it. Replaced all my clients farms with it and never looked back ( tho i hear the oss citrix project has gotten past all the hurdles since then. good for them )
Thu Nov 23 2023 18:39:25 EST from darknetuserIt is not super-duper groundbreaking, but it has a big advantage in that it places all your junk behind a single control panel. You may have a bunch of nodes runing independent virtual machines and LXC containers and still manage them all from the same interface. That is kind of cool because the setup is quite fast for what you get.
Also, their starred backup solution is quite complete for the price. It reminds me of a Restic for virtual machines.
The downsides? No nativeblock-device encryption, no automated backups for the host configuration, and the startup window nagging you to purchase support every time you log in. Also no native UPS integration. You can solve all of those manually. Oh, also the firewall subsystem sucks because changes to the firewall in a VM require you to restart the VM.
So at this point I'm still not sure about using ProxMox because it
locks you into a specific way of doing things.
If you inend to go with containers + Kubernetes then Proxmox is not much of an advantantage.
Maybe if your plan is to roll LXCs then you can roll Terraform + Ansible. Proxmox has some Terraform integration.
If you plan to roll Docker out then otright forget Proxmox.
Only to an extent and you can always back out if you dont like it..
Thu Nov 23 2023 18:58:40 EST from IGnatius T Foobar
So at this point I'm still not sure about using ProxMox because it locks you into a specific way of doing things.
Maybe if your plan is to roll LXCs then you can roll Terraform +
Ansible. Proxmox has some Terraform integration.
If you plan to roll Docker out then otright forget Proxmox.
I am big on containers -- specifically, OCI containers (what some people call "Docker containers").
LXC is super for its own purpose, which is to have virtual Linux instances that share a kernel instead of running them in virtual machines. It's basically like a FreeBSD Jail except they can have dedicated IP addresses and things like that. I myself ran something like that -- OpenVZ, which did roughly the same thing -- before I had a machine with hardware virtualization support.
In fact, it seems that LXC, the current version of Virtuozzo/OpenVZ, and container runtimes such as Docker's `containerd` are all kind of converging now, because they've all migrated to the Linux kernel's "cgroups" functionality to provide namespace isolation for process trees, networking, user IDs, and filesystems. This makes sense, of course. All the useful stuff all those people wrote ended up in the mainline kernel and is now shared.
On the other hand, an OCI ("docker") container is built to run a specific application. It's basically a jailed application bundled with all of its dependencies and isolated to its own sandbox. Same idea, different packaging.
And somehow that packaging made it appeal to the IT world even though it basically mimicked what the FreeBSD people had been doing for 20 years.
You *can* run a general purpose OS environment in Docker or Kubernetes, to some extent, even though that's not what it was designed for. The current instance of `dev.citadel.org` is basically a container that runs `sshd` as its application. There are a lot of system services missing though, so you wouldn't use this to run a desktop for example.
So I'm probably going to run MicroK8S on the bare metal, but I'll still need KVM for my router, and for my FreeBSD machine. Everything else will be in containers.