AnimeIsTrash posted:Learn to use the CLI OP I can, and do, use CLI on a nearly daily basis. But I'll be damned if I'm going to use a CLI when a GUI can do what I want and present it in a more readable and understandable format. That said, Cockpit looks like a pretty rad manager for Podman (cockpit-podman) that also combines a hypervisor manager (cockpit-machines) as well and cockpit-networkmanager lets do do a bunch of virtual networking too. I might have to boot up a fedora server and give that a shot.
|
|
![]() |
|
![]()
|
# ? May 29, 2023 02:10 |
|
Nitrousoxide posted:How can Podman be daemonless? Isn't a daemon just a background process? Surely Podman executes its behavior in the background without user interaction or calls on other such background processes? Different separation of concerns. Supervision is delegated to your process manager (i.e. systemd).
|
![]() |
Played around with Podman and Cockpit and while both are pretty cool, they do lack some functionality I use all the time in Portainer/Docker like editing existing containers (which is really copying the existing container setup, taking down the running container, and then starting up the edited duplicate). I'm also not seeing the containers I make for Podman in Cockpit in the CLI version of Podman. The list flag to the Podman command just doesn't return anything. Are they somehow running two instances of Podman that aren't talking to each other? I do like the very straightforward support for existing docker images, and the repos for docker images are built into both Podman and Cockpit which is nice. I see Podman supposedly supports docker compose files with their 3.0 release, which is pretty neat. Didn't get a chance to check that out. Cockpit doesn't support Podman using docker compose as far as I can see, so you'd have to do it via the CLI. Probably stick to Docker for now but I'll keep an eye on Podman.
|
|
![]() |
|
Yeah, the lack of compose support is the only reason I haven’t fully switched to podman yet. It’s nice to know they are adding it natively!
|
![]() |
|
Nitrousoxide posted:Played around with Podman and Cockpit and while both are pretty cool, they do lack some functionality I use all the time in Portainer/Docker like editing existing containers (which is really copying the existing container setup, taking down the running container, and then starting up the edited duplicate). I'm also not seeing the containers I make for Podman in Cockpit in the CLI version of Podman. The list flag to the Podman command just doesn't return anything. Are they somehow running two instances of Podman that aren't talking to each other? Podman suffers a lot from versioning issues like any cutting edge technology stack, depending on your OS it may be great or you might have to use a bunch of weird workarounds or stuff might just be missing entirely despite documentation suggesting otherwise. For example on CentOS 8 I still have to manually go in and delete a bunch of networking related files from var/lib/containerd anytime my machine does a hard reboot before any services will start. This bug was fixed a year or more ago and I don't have to deal with it on my Fedora server. That's why I say it's not production ready, it's been pretty buggy still on really anything that's not Fedora for me.
|
![]() |
Ah, well I was just testing it on a Fedora workstation VM because I wanted to use the localhost and didn't want to have to set up a bridge network for a VM to test the containers on my network. I did have some issues with it not grabbing the correct read/write permissions for the mounted paths to the container. So I'm not entirely sure what's up with that. I did find a flag that seemed to fix it for my app data folder but for the storage folder for like movies it wasn't able to see them for some reason. I might have just mistyped the path for the container mount however. I hope it matures and gets better over the next year or so. It looks pretty promising.
|
|
![]() |
|
Nitrousoxide posted:I'm also not seeing the containers I make for Podman in Cockpit in the CLI version of Podman. The list flag to the Podman command just doesn't return anything. Are they somehow running two instances of Podman that aren't talking to each other? This has a minor downside/gotcha, in that container storage is also per-user. So if user X & Y both "podman pull ..." the same container, it's stored twice in separate directories from what I can tell.
|
![]() |
|
Just put them both on a deduplicating filesystem, problem solved
|
![]() |
You know that probably also explains why I was having those mount issues as well. I was directing the file path to something that was in my home directory but I believe that cockpit was using a different user, probably System. I remember seeing a drop down for selecting the user in there. Of course a different user would not have permissions to read or write my user's home directory without using root or SU.
|
|
![]() |
|
I'm glad I don't have to touch containers and just use normal VMs. Everything about them sounds like a needless headache.
|
![]() |
|
Dont Touch ME posted:I'm glad I don't have to touch containers and just use normal VMs. Everything about them sounds like a needless headache.
|
![]() |
|
they work pretty well once you get the hang of them
|
![]() |
|
because they hang all the time, get it i'll see myself out
|
![]() |
|
'containers' would be better if they weren't treated by users and docker as though they solve all packaging issues forever 'multi stage builds' with docker just seem like passing the buck on sane artifact management
|
![]() |
|
Nitrousoxide posted:I can, and do, use CLI on a nearly daily basis. But I'll be damned if I'm going to use a CLI when a GUI can do what I want and present it in a more readable and understandable format. this is the linux thread
|
![]() |
|
Gentle Autist posted:this is the linux thread
|
![]() |
|
It's the 15th year of the Linux Desktop in a row. GUI-user FOSS evangelists have finally arrived.
|
![]() |
|
Gentle Autist posted:this is the linux thread
|
![]() |
|
mystes posted:Well it's not the BSD thread at least. you can tell because people are using it
|
![]() |
|
Gentle Autist posted:this is the linux thread
|
![]() |
|
I was just doing some work from my Gentoo workstation when MS Cortana started speaking to me. ![]() ![]() ![]() What the gently caress
|
![]() |
|
Apparently, the default VM template was amended to include a sound card and libvirt's Virtual Machine Manager managed to pass sound through the tunnelled ssh connection and I still had a Windows test VM open on a different virtual desktop. JFC I need a drink.
|
![]() |
|
And people say Linux audio doesn't work
|
![]() |
|
it works on mars
|
![]() |
|
Antigravitas posted:Apparently, the default VM template was amended to include a sound card and libvirt's Virtual Machine Manager managed to pass sound through the tunnelled ssh connection and I still had a Windows test VM open on a different virtual desktop. amazing... need to try this now on a server in reverse
|
![]() |
|
Antigravitas posted:Apparently, the default VM template was amended to include a sound card and libvirt's Virtual Machine Manager managed to pass sound through the tunnelled ssh connection and I still had a Windows test VM open on a different virtual desktop. you should buy a lottery ticket
|
![]() |
|
the free software community is revolting
|
![]() |
|
hey now there's no need for name-calling
|
![]() |
|
Let's replace RMS with regular MS!
|
![]() |
|
i guess we’re stuck with the fsf unless clang completely supersedes gcc
|
![]() |
|
ive been saying that for years!
|
![]() |
|
hobbesmaster posted:i guess we’re stuck with the fsf unless clang completely supersedes gcc You just... MIT be right 8-)
|
![]() |
|
apple, FreeBSD, and android replaced gcc with clang. Linux has no adult supervision that can make this decision system-wide so we’re stuck with gcc forever.
|
![]() |
|
Clang/LLVM can already compile the kernel, supposedly. https://www.kernel.org/doc/html/latest/kbuild/llvm.html I think rust support will require clang anyway. My guess is that fedora will switch to clang built kernels at some point, then centos/rhel will adopt it at some point after that. I doubt canonical would take that first leap.
|
![]() |
|
but what about FORTRAN???????
|
![]() |
|
Well, without Fortran the entire scientific Python ecosystem melts down and I'd be out of a job. I'm very much in favour of keeping Fortran. Just…keep it far away from me.
|
![]() |
|
sb hermit posted:Clang/LLVM can already compile the kernel, supposedly. I assume canonical is already working on their own compiler suite
|
![]() |
|
Soricidus posted:I assume canonical is already working on their own compiler suite the most they'll do is a front-end to an existing compiler suite and not a completely new compiler suite of their own
|
![]() |
|
the gnu fortran compilers are poo poo, so no loss goin to flang
|
![]() |
|
![]()
|
# ? May 29, 2023 02:10 |
|
xtal posted:Let's replace RMS with regular MS! Oh yeah, I'm thinking it's Plan 9 time.
|
![]() |