|
CaptainSarcastic posted:I don't remember exactly but I think it was somewhere around 2009-2010. It was definitely Unity
|
# ? Jan 25, 2021 20:48 |
|
|
# ? Jun 8, 2024 05:44 |
|
Mr. Crow posted:I'm late to the party but going to agree with other posters and nth the Fedora recommendation. You seem like you're getting by and enjoy the tinkering (based on not giving up), so I don't want to discourage you from continuing with Arch, it can be a great OS. Despite that every hardcore nerd I've known, yours truly included, has eventually gotten sick of the constant tinkering with Arch. If you get there consider Fedora, I am a heavy gamer and do a lot of messing around with linux for work and play on most distros and I have found Fedora to be the sweet spot of latest kernel (so mainly hardware compatibility; AMD drivers in your case) and package versions to poo poo breaking arbitrarily. It's very easy to get a gaming fedora rig up and running these days as well with RPMFusion (community repo of non-free or license restricted packages, more pacman less AUR). Pick the spin of Fedora (they default to gnome, please use anything but gnome) you want, install RPMFusion repos, install steam and optionally nvidia drivers from RPMFusion, go hog wild with bleeding edge stability. They also default to BTRFS as of Fedora 33.
|
# ? Jan 25, 2021 21:41 |
|
Cheese Thief posted:I like rolling distros, so Tumbleweed ticks all those boxes for me while also have the Suse name, which is more respected I'd reckon than Arch in an enterprise world. I've run Tumbleweed on a couple secondary laptops and I felt I was spending more time updating than actually using them. Leave the computer off for a couple weeks and next you turn it on there are a thousand packages that need to be updated. I do advocate rolling distros and think more distros should move to them or at least provide such version. But I also think we need to find a better balance with the update frequency. If it's a security update, package it. If the upstream creates a proper changelog entry listing bugfixes and features then probably package that too. But if it's just a git snapshot then skip it. If the upstream does their own binary releases for some distros that could be a good schedule to follow. Has anyone calculated how many supported versions of Linux kernel there are. Kernel.org themselves support half a dozen and most distros also support several versions, often quite not matching with upstream, just something that was out when they were preparing Distro version X. Think of all the duplicated effort.
|
# ? Jan 25, 2021 21:46 |
Do all applications have their config files go into /etc/ or is it something like Windows where they could be wherever? Trying to figure out how to back up then in case I screw up my system?
|
|
# ? Jan 25, 2021 23:39 |
|
Nitrousoxide posted:Do all applications have their config files go into /etc/ or is it something like Windows where they could be wherever? They do by convention, but they can indeed be elsewhere. Other common places could be in /opt or in the home directories for your user or root. If you're only trying to back up certain files, it might be easier to think of which software you're trying to back up and then find where the files are for them specifically.
|
# ? Jan 25, 2021 23:44 |
|
Nitrousoxide posted:Do all applications have their config files go into /etc/ or is it something like Windows where they could be wherever? Most applications do, but its not a hard rule by any means. You may also find stuff under /opt
|
# ? Jan 25, 2021 23:44 |
|
For user configurations, I would start by looking in ~/.config/. That subdirectory is where most user configurations are stored. There is no rule for this and other applications may differ. For instance, Firefox user configurations are contained inside ~/.mozzila/.
|
# ? Jan 26, 2021 01:01 |
|
It's not a hard rule but I don't think I've ever come across a package installed through the package manager repos that didn't put configs in /etc. If you download a package and install it by hand (dpkg -i ./somefile.deb) is generally when it might throw stuff in non-standard locations. Some packages will let you define user specific configs, those are generally all over... Typically either ~/.some_config, ~/.config, or ~/.local. Home directory dotfiles are generally frowned upon, .config is analogous to /etc and .local is to /usr
|
# ? Jan 26, 2021 03:07 |
|
A lot of applications have both user configs and system-wide configs. The user configs are usually within the user's home directory. The system-wide configs can be anywhere, but usually in /etc and /opt as others have mentioned.
The Gadfly fucked around with this message at 03:29 on Jan 26, 2021 |
# ? Jan 26, 2021 03:25 |
|
RFC2324 posted:I'm really starting to think I might want to start using a tmux shell tmux owns but nothing I describe there is part of (or depends on tmux) except the current time (displayed in the tmux status bar) -- the rest is all zsh.
|
# ? Jan 26, 2021 03:58 |
|
terminator for me, the pane management of tmux but with a gui. Or: the overweight python based clone of iTerm 2.
|
# ? Jan 26, 2021 04:04 |
|
Anyone have any experience with multiple AllowUsers in Redhat 7 in sshd? I found this: https://access.redhat.com/discussions/3872831 Basically I have to touch a bunch of servers with ansible to add a bunch of users we deploy and using something like this: code:
I believe I can just add lines to the end of the file and just do multiple entries but I was wondering if anyone in here has done this (not necessarily with ansible)?
|
# ? Jan 26, 2021 04:40 |
|
We've had to implement things like that ("add text to a line but only if it doesn't exist") with puppet and the only manageable solutions we found were to either have a function that did a query for all possible values (in our case, a hiera lookup), build an array of unique values, then paste it as a string or loop it with a template. The other option is augeas, which is a really weird and cool idea that I cannot wish on anyone, it's agony to use. There's also the "add if no such line exists" approach that gets limited use but you're already considering that. It is easy to do though so if it works in your case I'd try it.
|
# ? Jan 26, 2021 04:46 |
|
xzzy posted:We've had to implement things like that ("add text to a line but only if it doesn't exist") with puppet and the only manageable solutions we found were to either have a function that did a query for all possible values (in our case, a hiera lookup), build an array of unique values, then paste it as a string or loop it with a template. The other option is augeas, which is a really weird and cool idea that I cannot wish on anyone, it's agony to use. Yeah, this works: code:
|
# ? Jan 26, 2021 04:55 |
|
RFC2324 posted:Its not real big because of that german origin. for a long time there was no support in english, so you were stuck trying to figure out german language forums if anything suse specific went wrong. It's not real big because people who use Linux in the 1998-2008 era hate big companies in general, and Novell buying SuSE meant the distro was devil magic. Today it's problem is often doing things just differently enough from any other distro to break things. I'm thinking of the thread on the Plex forums about how to install the server on OpenSuSE which basically concluded, "this is just different enough from the nearest distro to require us to make a full repackage made just for it. gently caress." The installer could also use some work, particularly when it comes to the partition stage. YaST is great, though. RFC2324 posted:so you agree that the problem isn't actually arch, its people recommending arch to new users? The problem isn't just installing Arch, it's maintaining it. This happens less in Manjaro, but you still are in situations where it's suggested you browse a web site before you just issue a global update command to everything. "-Syu" is a meme at this point for breaking things, and someone upthread said it's not a problem if you check the homepage first, but that's a nonstarter for a lot of people from polished stability-tested distros who expect to install as many or as little updates are as available and still see things function afterward. Of course it requires active geek-dom and is a lot better with even a minor interest in scripting/programming; the entire philosophical bent of Arch is to consume code from upstream as quickly as it's pushed to GitHub. It's the operating system designed by thousands of people without the team of dozens first putting it through it's paces and making sure all that code doesn't cause conflicts. Craptacular! fucked around with this message at 05:08 on Jan 26, 2021 |
# ? Jan 26, 2021 05:05 |
Craptacular! posted:The problem isn't just installing Arch, it's maintaining it. This happens less in Manjaro, but you still are in situations where it's suggested you browse a web site before you just issue a global update command to everything. "-Syu" is a meme at this point for breaking things, and someone upthread said it's not a problem if you check the homepage first, but that's a nonstarter for a lot of people from polished stability-tested distros who expect to install as many or as little updates are as available and still see things function afterward. Having a filesystem that does snapshotting is a must for a rolling release OS like Arch IMO now that I've played with it. It takes a lot of the risk away of an update breaking everything since you can easily just roll it back and only update what didn't break.
|
|
# ? Jan 26, 2021 05:52 |
|
I think that's why Garuda is getting so much hype, it's Arch/Manjaro with Timeshift installed by default
|
# ? Jan 26, 2021 10:51 |
|
I don't know, I'm running my main desktop as Tumbleweed on ext4 and don't worry about it. The snapshots would be nice, but I figure that if I can't quickly get it to boot then having /home as a separate partition means that reinstalling is pretty trivial for me. This is on a very fast machine and I have other machines to use while it's down, though - if the computer wasn't crazy fast or if I didn't have a stupid level of redundancy in machines then I might not be so lackadaisical about it. I think the extent of my tweaking on a fresh install is just some minor cosmetics and installing a few packages (web browsers, Steam, VLC, etc.) - I run a pretty stock KDE desktop for the most part.
|
# ? Jan 26, 2021 11:42 |
Neato, my entire Garuda install nuked itself overnight, including the snapshots. Won't even load into the GRUB bootloader without complaining. Oh well, a good excuse to distro hop to another version.
|
|
# ? Jan 26, 2021 16:22 |
|
Jesus, that's an horrifying post to come back to. I can only say I hope you had your stuff backed up. Also I can chime in as another Fedora user to say that I've used for about 5 years now and at least it certainly never did that
|
# ? Jan 26, 2021 17:00 |
|
Nitrousoxide posted:Neato, my entire Garuda install nuked itself overnight, including the snapshots. Won't even load into the GRUB bootloader without complaining. Christ on a bike. Can you please let us know if it turns out it was a HDD death rattle or something else? Because it makes me want to stay the hell away from Garuda the next time I install a linux, but I'm basically the target user so it'd be a pity to discard it over a false alarm.
|
# ? Jan 26, 2021 17:07 |
|
Kassad posted:Jesus, that's an horrifying post to come back to. I can only say I hope you had your stuff backed up. I've heard stories about Fedora doing it too, its just pretty rare nowadays. And this is how distro hoppers are born ☺️
|
# ? Jan 26, 2021 17:12 |
|
I've used linux on and off for 20 years and it's never spontaneously nuked itself so I'd be very curious to know how that could happen
|
# ? Jan 26, 2021 17:37 |
|
netcat posted:I've used linux on and off for 20 years and it's never spontaneously nuked itself so I'd be very curious to know how that could happen It usually means you did something the night before but the self-nuked system destroyed all records of what you did. MacOS is the only OS i have never heard of it happening with
|
# ? Jan 26, 2021 17:45 |
NihilCredo posted:Christ on a bike. Can you please let us know if it turns out it was a HDD death rattle or something else? Because it makes me want to stay the hell away from Garuda the next time I install a linux, but I'm basically the target user so it'd be a pity to discard it over a false alarm. Oh it's quite possible I hosed it up. I was messing with the swap disk config before I went to bed. I swapped over to PopOS and this feels much snappier and much more responsive than Garuda. Nitrousoxide fucked around with this message at 18:45 on Jan 26, 2021 |
|
# ? Jan 26, 2021 18:43 |
|
Matt Zerella posted:Anyone have any experience with multiple AllowUsers in Redhat 7 in sshd? I wouldn't want to do that much editing for sshd_config, it would be nice if you keep it default so package updates can do any changes from Red Hat to it. I would rather use a custom group to allow access, it's trivial to add and remove users from a group. At work we have groups for sudo and normal users and then use /etc/security/access.conf to grant them SSH access. rufius posted:Honest question - what’s notable/interesting about running OpenSUSE? One niche SUSE Enterprise has found seems to be SAP. We are running SAP HANA on RHEL7 which is our most common server OS, but HPE which provides us HANA support has said about 90% of their customers run HANA on SUSE, so that's what our next iteration will use. We have hundreds of RHEL and Ubuntu servers and we don't want to touch SUSE, but the few SUSE servers we will have to have are also some of the most important ones.
|
# ? Jan 26, 2021 18:51 |
|
Linux people, may I request some owners of Intel, AMD, or nVidia GPUs on a Linux desktop test this webpage in a browser for at least 20s and see if it can maintain a smooth jank-free animation: http://ahyoomee.miru.hk/test/www/ff.html Basically I'm looking for OS version and GPU specs to add to this report, REF: https://github.com/mrdoob/three.js/issues/21088 Skip if you are running in a VM as it appears neither VirtualBox or Parallels actually support WebGL2 anymore. It looks like ChromeOS and Windows are fine, but MacOS and Android definitely have issues. Cheers.
|
# ? Jan 26, 2021 19:19 |
If only there was a filesystem, such as ZFS, which implemented reverse-delta snapshots in such a way that they could be used as boot environments, such that if you, prior to doing anything administrative, created a boot environment, with, say a command like beadm or bectl, you would then, in the loader, be able to select the boot environment, which would then bring the system back to the state that it was in, and you could then examine it. That would be pretty neat if that was possible.
|
|
# ? Jan 26, 2021 19:30 |
|
No home system is worth that kind of prep work, just back up the poo poo you can't recreate to a separate drive or a cloud service and if the system blows up wipe and reinstall. Which I've never had happen in 20+ years but it's obviously possible. Unless one considers testing ZFS at home self training for work, that's kind of a good idea but I'd never do it on my daily driver.
|
# ? Jan 26, 2021 19:44 |
|
I love my ZFS NAS even though I spend a lot of time asking for help about it in here. It doesn't take any more work than another FS, it just is better at it and has more features.
|
# ? Jan 26, 2021 20:01 |
xzzy posted:No home system is worth that kind of prep work, just back up the poo poo you can't recreate to a separate drive or a cloud service and if the system blows up wipe and reinstall. Which I've never had happen in 20+ years but it's obviously possible. Intel and IBM absolutely deserve poo poo for not including it in the spec to save a few bucks, when it's part of every single other volatile memory (CPU caches all use ECC, as do the caches inside disks, and workstations, servers and mainframes all use ECC for main memory too).
|
|
# ? Jan 26, 2021 20:08 |
BlankSystemDaemon posted:Memory errors are the cause of much more lost productivity than basically any other type of error that any machine can experience, especially for consumer PCs according to this study by Microsoft from 2010 (I finally found this study again, I thought it was lost), as well as systems in general. I don't think ZFS would help with errors cropping up when stuff getting written to RAM. It only helps in detecting read errors because they differ from the hash written to the storage.
|
|
# ? Jan 26, 2021 20:16 |
|
MrMoo posted:Linux people, may I request some owners of Intel, AMD, or nVidia GPUs on a Linux desktop test this webpage in a browser for at least 20s and see if it can maintain a smooth jank-free animation: It's pretty janky in Firefox 84.0.2, less janky but still janky in Chrome 88.0.4324.96. The machine I'm using right now is Phenom II 965, 16GB RAM, Nvidia 1060 6GB proprietary driver version 460.32.03, on OpenSUSE 15.2 with KDE, fully updated.
|
# ? Jan 26, 2021 20:32 |
|
MrMoo posted:Linux people, may I request some owners of Intel, AMD, or nVidia GPUs on a Linux desktop test this webpage in a browser for at least 20s and see if it can maintain a smooth jank-free animation: The state of Linux is such that you're going to get wildly different responses based on graphics card, browser, display server, and desktop, since browser hardware acceleration is still Getting There(TM). Many, many people cause their CPUs to work overtime to simply watch YouTube still. Nitrousoxide posted:I swapped over to PopOS and this feels much snappier and much more responsive than Garuda. I'll just again suggest Fedora if you'd like a distro that walks the razor's edge of newer versions than buntu/Pop would without all the crashing of Arch. I wouldn't suggest Fedora for grandma, particularly, but developers and gamers can get a better deal out of it.
|
# ? Jan 26, 2021 20:52 |
|
Craptacular! posted:The state of Linux is such that you're going to get wildly different responses based on graphics card, browser, display server, and desktop, since browser hardware acceleration is still Getting There(TM). Many, many people cause their CPUs to work overtime to simply watch YouTube still. That's why the test is 2 triangles sitting inside the GPU, the majority of systems within the last 10 years should be able to manage that. It should be < 1ms per frame with 10+ms to spare for even the worst case. Anything outside of those boundaries will not even support WebGL2. MrMoo fucked around with this message at 21:05 on Jan 26, 2021 |
# ? Jan 26, 2021 20:57 |
Nitrousoxide posted:I don't think ZFS would help with errors cropping up when stuff getting written to RAM. It only helps in detecting read errors because they differ from the hash written to the storage. That's not the point, though, and unfortunately there's been a lot of FUD spread regarding the use of ZFS and ECC. The short of the long is that while ZFS was designed on SPARC hardware where ECC is standard, because that's just how Big Iron works, it doesn't mean ZFS needs ECC to function, and ZFS will still keep your data safe in scenarios where other filesystems would likely end up writing corrupt data to disk, and there's no such thing as a scrub of death. ECC is always good idea for system availability, because of the prevalence of memory errors. BlankSystemDaemon fucked around with this message at 21:06 on Jan 26, 2021 |
|
# ? Jan 26, 2021 21:04 |
|
BlankSystemDaemon posted:ZFS can even help there, although the feature is hidden behind a flag called ZFS_DEBUG_MODIFY, because its datapaths may contain zebras. What this flag does is ensure that as soon as the data is added to memory, it's checksummed, rather than when it's written to disk. was someone talking smack about zfs? because I'm pretty sure everyone likes it, just knows its not in a production ready state on linux
|
# ? Jan 26, 2021 22:07 |
|
RFC2324 posted:because I'm pretty sure everyone likes it, just knows its not in a production ready state on linux Ain't stopping my group! Everyone's butthole clenches a little bit every time it's kernel update time.
|
# ? Jan 26, 2021 22:25 |
RFC2324 posted:was someone talking smack about zfs? EDIT: If you want to check, llnl.gov and lanl.gov appear in commits to the OpenZFS repo, and back when it was started, ZFS on Linux was exclusively used developed by LLNL for use with their own cluster, with Lustre as a middle-layer between the individual ZFS pools and the HPC cluster itself to ensure distributed storage. EDIT2: The project started in 2008 and finished in 2012 according to the announcement. BlankSystemDaemon fucked around with this message at 00:56 on Jan 27, 2021 |
|
# ? Jan 27, 2021 00:34 |
|
|
# ? Jun 8, 2024 05:44 |
|
MrMoo posted:Linux people, may I request some owners of Intel, AMD, or nVidia GPUs on a Linux desktop test this webpage in a browser for at least 20s and see if it can maintain a smooth jank-free animation: Smooth here on Firefox 84.0.2 and Chrome 88. Fedora 33 / Xorg / nvidia (1080 Ti)
|
# ? Jan 27, 2021 09:49 |