[00:00] <thyriaen__> sarnold, oh you mean because i wrote .Xsession with a capital X ?
[00:00] <sarnold> yes
[00:00] <thyriaen__> ok let me check if it have it in capitals in my directory
[00:02] <thyriaen__> sarnold, nah, i had it in small letters unfortunately
[00:02] <sarnold> :(
[01:27] <jpmh> I have a linux machine with no GUI and need to set it to connect to a local wifi network.  In the past I edited  /etc/wpa_supplicant/wpa_supplicant.conf to add the device.  That file no longer seems to exist.  What file doI need now.  Ubuntu 20.04
[01:29] <Bashing-om> jpmh: Does https://blog.ubuntu.com/2017/12/01/ubuntu-bionic-netplan help ?
[01:33] <MrSteve> Bash. He there :)
[01:33] <jpmh> Bashing-om: TY - looking at that reference now - ty
[01:33] <MrSteve> I get mailed.
[01:36] <jpmh> Bashing-om: so,I looked on my laptop, which I did configure using the gui forthe old file and any of the new ones referecedin that document and see none of them
[01:39] <jpmh> Bashing-om: I even did: find / -name \*yaml |grep netplan - to see iif it was somewhere else - now I have no idea how my laptop connects either
[01:43] <Bashing-om> jpmh: what shows as the renderer from ' /etc/netplan/01-network-manager-all.yaml ' ?
[01:44] <jpmh> Bashing-om: NetworkManager
[01:45] <Bashing-om> jpmh: Oh boy - with no GUI you need to switch to networkd.
[01:46] <sarnold> you can drive things with nmcli
[01:46] <sarnold> it's more work
[01:46] <sarnold> but it might be less trouble than switching to networkd, depending upon how much of your networkmanager does correctly already
[01:48] <jpmh> Bashing-om: on my laptop, which IS connected, there is a file: 01-network-manager-all.yaml in 01-network-manager-all.yaml but it has NO info on my connection, I was hoping to copy that
[01:48] <sarnold> the netplan yaml to start up nm is basically the same two or three lines for everybody
[01:49] <Bashing-om> jpmh: I would suggest that you let systemd do it. Maybe like: https://www.configserverfirewall.com/ubuntu-linux/ubuntu-network-manager/ <- Disable Network Manager and enable systemd-networkd // But sarnold's advise is also good.
[01:49] <jpmh> sarnold: if that was addressed to me, so, where is the actual info on my connections kept. That file is the same on the laptop and the machine I want to use
[01:50] <sarnold> jpmh: check /etc/NetworkManager/system-connections/
[01:51] <jpmh> Bashing-om: I do see that you two agree, what I don't get, is there must be a file somewhere with the info otherwise I would not be connected through my laptop
[01:51] <Bashing-om> jpmh: Be aware that " 01-network-manager-all.yaml" is for network-manger. A different file is sought for networkd.
[01:52] <jpmh> sarnold: That seems to be it - there is a file in there  for evey network I have conneted to and it seems to contain in each file all that would be needed - I'm about to copy to the on-gui machine
[01:52] <Bashing-om> jpmh: systemd-networkd, and config's under /etc/systemd/network/ .
[01:54] <jpmh> by on-gui I mean non-gui
[01:55] <sarnold> no worries, most of the time typos on irc can be ignored :)
[01:56] <Bashing-om> jpmh: netplan is a configuration generator, for either NetworkManager, or systemd-networkd the idea is to create a config in /etc/netplan that is then applied automatically to whichever service manages network connections
[01:59] <jpmh> guys, thanks for the heklp.I copied the file and all works - TY so much
[02:00] <sarnold> woot
[02:00] <Rakko> Is there an apt* command (or some other command) that acts like grep, but it searches the files inside all packages, not just ones you have installed?
[02:01] <sarnold> Rakko: apt-file can search the filenames
[02:01] <sarnold> Rakko: debian hosts a code search tool that you can use; it only indexes debian stuff, but it's really useful https://codesearch.debian.net/
[02:01] <Rakko> oh, nice. That's probably good enough for me.
[02:01] <Rakko> thanks
[02:12] <samian> hi
[02:40] <mybalzitch> https://ubuntu.com/tutorials/create-a-usb-stick-on-windows#8-write-warnings <== these instructions do not work. you must select "write in DD mode"
[02:41] <mybalzitch> the installer failed to find the iso image after booting if you choose "write in ISO mode", and the installer also fails to run the live image if you let it attempt to netboot. lol.
[02:43] <tomreyn> mybalzitch: please scroll down on said page to where it says "Suggest changes"
[02:43] <tomreyn> that's where your findings should go - here they'll just be lost in the void
[02:46] <Linkandzelda> can I use the ubuntu server ISO to install a bootable system to a usb stick plugged in the same system?
[02:47] <mybalzitch> tomreyn: I'm on IRC. I'm not signing up for some discus account to report the issue when the maintainers for said documentation likely idle here, or someone else has an account.
[02:47] <tomreyn> Linkandzelda: yes
[02:49] <Linkandzelda> tomreyn: great, thanks. gonna put my OS on a usb plugged to my server's motherboard and free up a sata port
[02:50] <tomreyn> mybalzitch: it uses ubuntu sso, so if you already have an account there then you wont need to create another.
[02:50] <mybalzitch> nope.
[02:51] <tomreyn> and the other thing just won't happen, so it'll go to /dev/null
[02:52] <bigpod>  mybalzitch: i would as well recommend you to use that becuause again reality is IRC is basicly /dev/null
[02:52] <tomreyn> Linkandzelda: maybe use an ssd ther,e though, standard flash drives are just not durable enough
[02:53] <mybalzitch> that's great for an official support channel
[02:53] <bigpod> mybalzitch: and those instructions work i used rufus maybe a week ago without having to write in DD mode
[02:54] <mybalzitch> not on my asus board
[02:54] <mybalzitch> the installer boots, but it can't find the image to mount
[02:54] <mybalzitch> reboot and write with dd mode, works fine
[02:54] <mybalzitch> ¯\_(ツ)_/¯
[02:54] <bigpod> wierd
[02:55] <bigpod> on both my HP laptop and asrock MB based desktop i did not have to select Write in DD mode
[02:56] <bigpod> and installer worked just fine, it booted went into Ubiquity and installed everything
[02:56] <mybalzitch> I've never had write in ISO mode work for any flash drive I've tried
[02:57] <bigpod> wierd
[02:57] <mybalzitch> and the installer/initramfs offered to download an ISO. it grabbed network config, downloaded it, said it was gonna do stuff, then just dropped me to the initramfs prompt
[02:57] <bigpod> only time ever i had to do it was with completly different distribution one time and that was it, ubuntu has never required that from me
[02:59] <tomreyn> mybalzitch: this is not the primary (commercial) support channel vanonical provides, if that's what you assumed. this channel is mostly (read: almost exclusively actively) used by volunteers, such as yourself.
[02:59] <tomreyn> *Canonical
[03:00] <bigpod> mybalzitch: i would say that there is something wrong with your system if that is a requirement
[03:00] <mybalzitch> bigpod: I've had it with multiple computers and multiple flash drives, but please go on
[03:01] <bigpod> mybalzitch: now that changes things but still not actualy sure because i have never had to use dd mode so yea
[03:02] <bigpod> tomreyn: as far a si understand its ment for basicly asking questions aka ephemeral stuff
[03:02] <mybalzitch> also, this happens at a point where the board has picked up the usb drive, booted grub from the drive, and loaded part of the installer
[03:03] <mybalzitch> then the installer is unable to find the OS image in a format it likes to continue the installation process
[03:05] <tomreyn> bigpod: yes, *support* channel. i'm just saying don't expect the developers to wait for you to ask your questions and seek an answer for you, or to update docs based on what you type here. (i know you dodn't suggest that's the case.)
[03:07] <mybalzitch> lol
[03:08] <bigpod> mybalzitch: to me that sounds like either HW issue or file validity issue sorry
[03:09] <bigpod> because from my experiance if DD mode is required you dont get into boot of OS at all you will at best get to grub and even that is a tall order sometimes
[03:12] <bigpod> but it is true that DD mode writes the files in a different way then ISO mode but i still dont know how that would affect anything becasue as far as i understand ubuntu ISOs are designed in a way so they can be written in ISO mode
[03:13] <genii> Yes, they are hybrid images
[05:15] <thyriaen> how can i configure certain applications to always run in preset workspaces ?
[06:26] <bumblefuzz> so, I changed my DE to KDE
[06:28] <bumblefuzz> and I can't figure out how to change ctrl+alt+T to terminator
[06:31] <thyriaen> oh my god - i am going insane over these keyboard configurations
[06:32] <thyriaen> alright so i have managed to get xmodmap to load my ~/.xmodmap file but now after some time IT REVERTED back to the default settings - what might have caused this and how can i stop it from happening ?
[06:36] <ILoveAndrewLee> thyriaen: what keyboard configurations?
[06:37] <ILoveAndrewLee> may be related but I have a custom keyboard layout written in "us", and it gets overwritten every time I update my OS
[06:37] <thyriaen> ILoveAndrewLee, i have a few remaps in my ~/.Xmodmap file which i load on startup
[06:39] <ILoveAndrewLee> thyriaen: Ever thought of just `sudo vim /usr/share/X11/xkb/symbols/us`
[06:40] <ILoveAndrewLee> This way, your keyboard remaps also work under root
[06:40] <thyriaen> ILoveAndrewLee, i have thought of it, yes i didnt want to change a system file because, as you are describing once i update it it gets overwritten
[06:41] <ILoveAndrewLee> thyriaen: It happens once every 6 months, or 2 years if you are doing lts. I then just save my US file in a repo, and load it from there
[06:41] <thyriaen> ILoveAndrewLee, maybe i can make a copy of us and put my changes in and call it /usr/share/X11/xkb/symbols/thyriaen ?
[06:41] <thyriaen> ILoveAndrewLee, it only screws it up if you do a dist-upgrade ?
[06:42] <ILoveAndrewLee> thyriaen: Well you should DEFINITELY back up US before editting it, and definitely make a way for you to be able to restore it, because if you f up your US file, your keyboard will stop working
[06:42] <ILoveAndrewLee> thyriaen: Right, only if you do a dist-upgrade
[06:43] <ILoveAndrewLee> thyriaen: I've done enough mucking around on my us file, that I had to devise an effort to `cp` the backup file with my mouse and middle click.
[06:44] <thyriaen> ILoveAndrewLee, i am using the uk layout however this is my output of the symbols directory
[06:45] <thyriaen> https://termbin.com/hib2
[06:45] <thyriaen> what to do you think might be the correct layout ?
[06:45] <ILoveAndrewLee> I'm guessing its us,
[06:46] <thyriaen> but i dont use us
[06:46] <thyriaen> im guessing it might be pc
[06:46] <ILoveAndrewLee> thyriaen: The files borrow from each other. I use Canadian and have used french, but the edits I made to colemak ( found in us ) work none the less
[06:47] <thyriaen> T_T
[06:47] <thyriaen> why are keyboard layouts so confusing
[06:47] <ILoveAndrewLee> thyriaen: don't get me started.
[06:48] <ILoveAndrewLee> thyriaen: Anyways, what sort of remaps are we talking about? For what its worth, I have an 8 layer programmer's layout
[06:49] <ILoveAndrewLee> Take a look here. https://usercontent.irccloud-cdn.com/file/Lt4snw9l/image.png
[06:49] <thyriaen> not that much i am just starting out editing the layout and since everything i touch is just one huge pain to even get it load
[06:49] <thyriaen> i dont want to extend it before i can get it running
[06:49] <ILoveAndrewLee> What I do to achieve this, ( Its technically the wrong way but the right way is confusing and cumbersome ) is that I just overwrite the "Colemak" layout with what I want
[06:50] <ILoveAndrewLee> thyriaen: I did create a spreadsheet that will generate the code for you if you are interested.
[06:50] <ILoveAndrewLee> one sec. It makes things vastly more simple
[06:50] <thyriaen> that would be awesome
[06:51] <thyriaen> i just want to get things done before my uhk gets here
[06:51] <thyriaen> and i get leave all that software stuff behind
[06:51] <ILoveAndrewLee> https://docs.google.com/spreadsheets/d/19paGH1cmvnzNiCFA39nmm9afTIfgCIrt8NUOLpyqM3k/edit?usp=sharing
[06:51] <ILoveAndrewLee> Go there. Make a copy
[06:52] <thyriaen> thanks but that might be overkill :p
[06:53] <makara> hi. I updated my Ubuntu 20.04.2 server a while back and rebooted today. I use sshfs to mount drives, but now when i umount them on my desktop it puts the server to sleep! What's happening?
[06:53] <makara> i have systemd logs
[06:53] <thyriaen> can i make space an modifier such that space+q is F2 ?
[06:54] <ILoveAndrewLee> thyriaen: totally fine. I wrote a tutorial for it here: https://askubuntu.com/questions/510024/what-are-the-steps-needed-to-create-new-keyboard-layout-on-ubuntu
[06:54] <ILoveAndrewLee> thyriaen: Yes you can.
[06:54] <thyriaen> is that tutorial for your spreadsheet ?
[06:55] <ILoveAndrewLee> thyriaen: yes. At the bottom, `Quick and dirty solution`, that details out how to use the spreadsheet
[06:55] <thyriaen> alright
[06:55] <thyriaen> thanks a lot
[06:55] <ILoveAndrewLee> The rest is some basic understanding
[06:55] <thyriaen> i will inhale that and see if that fixes my stuff
[06:55] <ILoveAndrewLee> thyriaen: Do you still want your space to act as a space?
[06:57] <thyriaen> ILoveAndrewLee, no
[06:57] <ILoveAndrewLee> thyriaen: If you look at my layout, I have switched the <Return> key with a <shift> modifier. For a time, I also wanted it to also function as a <Return> key as well, that would return if I ... oh
[06:57] <thyriaen> :p
[06:57] <ILoveAndrewLee> Interesting layout then you got there. What is your idea of making the space a modifier?
[06:58] <thyriaen> oh its not going to be my default layout - currently i am only editing my normal layout to include german letters and the euro sign, switch : and ; around, put capslock as esc and so on
[06:59] <ILoveAndrewLee> capslock as esc? Must be for vim or something.
[06:59] <thyriaen> but once i have figured it all out i wanna make a 2nd starcraft brood war layout where space + q w e are my camera hotkeys
[06:59] <thyriaen> so the space thing is just for sc:bw - not for my usual layout
[07:00] <ducasse> makara, are you mounting them on an ubuntu client?
[07:00] <ILoveAndrewLee> If I can offer a suggestion, I access my alt keys with left and right alt.
[07:00] <makara> ducasse: yes
[07:00] <ducasse> makara, why not use nfs?
[07:00] <thyriaen> ILoveAndrewLee, partly for vim, but partly in preperation of my UHK keyboard which does not have an esc key - so might aswell get used to it right now
[07:01] <ILoveAndrewLee> thyriaen: In order to make the right alt comfortable, I shifted over the right hand by one row. This also had the added benefit of making they keyboard symmetrical and more comfortable.
[07:01] <makara> ducasse: i don't know/can't remember. Is it secure? Is there a tutorial for that?
[07:01] <ILoveAndrewLee> thyriaen: uhk keyboard ... I got to look that up
[07:02] <ILoveAndrewLee> thyriaen: Ah. And Ew. They didn't fix they symmetry!
[07:02] <thyriaen> ILoveAndrewLee, it uses this layout: https://cutt.ly/VnOEobZ
[07:02] <ILoveAndrewLee> Still cool though.
[07:02] <ducasse> makara, it has client-based security, so if the server allows you root access local root is root. very simple to set up, fast, much faster than sshfs
[07:03] <thyriaen> ILoveAndrewLee, it also does all the key remapping on the hw side
[07:03] <ILoveAndrewLee> thyriaen: Now that is cool
[07:04] <ILoveAndrewLee> thyriaen: Are you still going with Qwerty? Or are you going to do Dvorak, Colemak, or Workman?
[07:04] <thyriaen> ILoveAndrewLee, so in preperation to that layout i want to change my stuff to be closer to it to make the transition easier
[07:04] <ducasse> makara, they are on the same network?
[07:06] <thyriaen> i have never looked at dvorak or anything apart from qwerty - i recently moved from qwertz layout to qwerty and started using 10 finger typing, got into vim and everything falls together now which is a good point of breaking bad habbits - before i was kinda a floating 6 finger typist just freestyling it
[07:06] <ILoveAndrewLee> thyriaen: ah I see the layout now. There are a lot of things I like about that layout. For one, the right alt is much more accessible to the right thumb. That should be your modifier key.
[07:07] <thyriaen> ILoveAndrewLee, yea, i don't know what keys to have which modifiers on it
[07:07] <ILoveAndrewLee> thyriaen: I did give a 2 hour presentation on this at my LUG if you are interested, but the short version is that you are extremely smart for investing the time to do this.
[07:07] <thyriaen> ILoveAndrewLee, i definitely want to have multiple layers with different stuff
[07:07] <ILoveAndrewLee> thyriaen: Have you just learned VIM?
[07:08] <thyriaen> im getting started on it - not much more than hjkl, e a A :wq yet
[07:08] <ILoveAndrewLee> thyriaen: Well the 8 layer layout is your best bet. Took me a while to figure out how to do that
[07:09] <makara> ducasse: yes
[07:09] <thyriaen> yesterday i started a haskell tutorial and i want to do it with vim - so learn haskell, new layout, new typing technique and vim in one go
[07:10] <makara> ducasse: i'll definitely have a look, but at the moment I can't have that server going to sleep all the time. It just did it again, without disconnecting sshfs
[07:10] <thyriaen> it think its best to throw out everything you know at the same time so your brain can concentrate on new stuff and don't fall into old patterns
[07:10] <ILoveAndrewLee> thyriaen: okay that is good because the way I learned vim, is having to remap everything out for colemak. Its best to do that when you are fresh
[07:10] <makara> so im trying https://unix.stackexchange.com/a/25151/47987
[07:10] <ILoveAndrewLee> thyriaen: EXACTLY!
[07:10] <thyriaen> ILoveAndrewLee, what are the benefits of not using qwerty ?
[07:11] <ILoveAndrewLee> thyriaen: Comfort, and typing speed are the main ones. For example, if you were to compare typing a novel like war and peace on Colemak vs Qwerty,
[07:11] <ILoveAndrewLee> your fingers would travel like 3 miles less, and 50% less overall compared to qwerty.
[07:12] <ILoveAndrewLee> It prevents carpel tunnel syndrome, which I had bad under qwerty, where I had to literally rest from typing.
[07:13] <thyriaen> ILoveAndrewLee, while i do love stuff like that - my main concern is that it will give me an obstacle in the world to have a different layout - when i am at a different computer, typing on a friends laptop, id assume its quite awkward
[07:13] <ILoveAndrewLee> thyriaen: Those are the main ones. After that, my most important changes to they keyboard was to first make a Programmer's layout.
[07:13] <ILoveAndrewLee> thyriaen: Yeah, that will be pretty much impossible to deal with.
[07:14] <thyriaen> ILoveAndrewLee, i grew up with qwertz ( german layout ) and my main reason to switch to qwerty was to be more compatible with the world, for example when connecting to a server via ssh or just in general being on a pc in the university
[07:14] <ILoveAndrewLee> I have my own laptop, a thinkpad though, so it is almost never an issue. Sometimes it is though, but I don't think its worth the health of your fingers.
[07:15] <ILoveAndrewLee> thyriaen: Well the fact that you will have a hardware based keyboard, solves that issue for you.
[07:15] <ILoveAndrewLee> Frankly, I would have gotten one if I knew that existed.
[07:16] <thyriaen> ILoveAndrewLee, yea maybe - maybe you will get your UHK :p
[07:16] <thyriaen> it also splits for ergonomic stuff
[07:16] <ILoveAndrewLee> thyriaen: That is huge too. Qwerty is horribly unsymmetrical.
[07:17] <thyriaen> ILoveAndrewLee, what kind of thinkpad you got ? I have my X1 Carbon 2017 and thinking about getting a T14s Gen2 AMD
[07:17] <ILoveAndrewLee> Even wads is a travesty, because it is oriented in the worst possible way for comfort for gaming.
[07:18] <ILoveAndrewLee> thyriaen: I got a ton, but I got an x290 silver yoga right now. That carbon is a beautiful machine
[07:18] <thyriaen> ILoveAndrewLee, i am thinking about going esdf when i pick up another shooter :p
[07:18] <ILoveAndrewLee> thyriaen: You just got the best taste in everything, lol.
[07:18] <ducasse> makara, check the logs
[07:18] <thyriaen> ILoveAndrewLee, probably the thing i love the most about myself, my taste
[07:19] <ILoveAndrewLee> thyriaen: interesting. Ive shifted over to the right hand for arrows. Incidentally, i also remapped the arrow keys to alt kl;o
[07:19] <thyriaen> ILoveAndrewLee, right and for arrows and left hand for mouse ? are you left handed ?
[07:19] <ILoveAndrewLee> or jikl. So you hold alt, and you have access to arrows. Great for document editting. And its synonymous with vim.
[07:20] <ILoveAndrewLee> thyriaen: oof I didnt think about the mouse. I play mostly rpgs...
[07:20] <thyriaen> ILoveAndrewLee, i plan to have some kind of uhk modifier map the actual arrow keys to modifier+hjkl
[07:21] <thyriaen> cause some applications dont have hjkl support baked in
[07:21] <ILoveAndrewLee> thyriaen: I did try sdfr. Not baaaaad but still not ideal.
[07:21] <thyriaen> do you know superhuman ? :p
[07:21] <ILoveAndrewLee> superhuman?
[07:21] <thyriaen> ILoveAndrewLee, yea, the email client
[07:21] <ILoveAndrewLee> thyriaen: You mean I can dispense with thunderbird? lol
[07:22]  * BobbyJr ears perked up
[07:22] <BobbyJr> theres finally a Thunderbird rival that actually works?
[07:22] <ILoveAndrewLee> ^
[07:22] <thyriaen> ILoveAndrewLee, email clients are so awful tbh
[07:23] <thyriaen> i absolutely love superhuman
[07:23] <ILoveAndrewLee> yes. Something about thunderbird... I'll check it out.
[07:23] <ILoveAndrewLee> thyriaen: Have you heard of QuteBrowser?
[07:23] <thyriaen> everything is keyboard driven made for vim shortcuts
[07:23] <thyriaen> i have not
[07:23] <ILoveAndrewLee> WOAH IM SOLD!
[07:23] <ILoveAndrewLee> LETS GO! VIM BASED EMAIL
[07:24] <ILoveAndrewLee> thyriaen: Qutebrowser is a vim based browser. I've used it for almost a year now. Its based.
[07:24] <ILoveAndrewLee> Love it.
[07:24] <thyriaen> ILoveAndrewLee, how is the rendering engine ?
[07:24] <ILoveAndrewLee> thyriaen: I'm also right now learning and working on sc-im, which is a spreadsheet terminal app with vim bindings.
[07:24] <thyriaen> i have found everything which is not chromium or firefox based to be unusable in the current web
[07:25] <ILoveAndrewLee> thyriaen: I'd say its excellent. It uses QtWebEngine, which is based on Chrome
[07:25] <thyriaen> you mean i can get rid of freakin libre fkn calc ?
[07:25] <ILoveAndrewLee> hahah, maybe... I like it so far, and the developer is active so I've been chasing down bugs with him.
[07:26] <thyriaen> ILoveAndrewLee, let me know if you need a superhuman invite
[07:27] <ILoveAndrewLee> thyriaen: If you are just getting into coding, I also suggest checking out QtCreator. Qt is the best way to do C++ imo, and the IDE is great for almost all languages. It also has vim built in
[07:27] <BobbyJr> I always end up going back to FF everytime I try a different browser. That said, I like MS Edge. Its a better Chrome than Chrome.
[07:27] <thyriaen> ILoveAndrewLee, i wouldnt say im getting into coding - i use the intellij suite so clion for c++
[07:28] <ILoveAndrewLee> BobbyJr: I find FF unusable now, after I got used to qutebrowser. Its like trying to use a normal text editor when you know vim.
[07:28] <ILoveAndrewLee> thyriaen: Interesting. System Admin then? I always assume someone using vim is doing one of those two things, to be a programmer or to be a sysadmin
[07:30] <thyriaen> ILoveAndrewLee, I am doing my masters in theoretical computer science, mostly writing lots of latex - but i am also self employed general IT Guy, general interest in linux and doing things the open source, and proper way - so i do my invoices in latex for example
[07:31] <ILoveAndrewLee> thyriaen: Oooff latex.
[07:31] <ILoveAndrewLee> I tried writing a parser for latex; it was not fun. haha.
[07:31] <thyriaen> ILoveAndrewLee, i absolutely love latex
[07:31] <ILoveAndrewLee> thyriaen: Is there a vim based latex editor that supports active previews?
[07:32] <thyriaen> i do my presentations in it - was able to completely replace word and powerpoint havent touched that shit in 5 years
[07:32] <thyriaen> ILoveAndrewLee, there is - i did all my previous projects in sublime text but probably going to configure and move over to vim this month
[07:33] <ILoveAndrewLee> I would do the same, and guys like luke smith really proved out the concept of how superior it is, but I have not found a good editor.
[07:33] <thyriaen> sublime text with latexing package is what i used
[07:33] <ILoveAndrewLee> Sublime... mmmmm Hard pass
[07:33] <thyriaen> i wrote my bachelor thesis in it
[07:34] <ILoveAndrewLee> thats epic.
[07:34] <ILoveAndrewLee> What was your thesis on?
[07:34] <thyriaen> ILoveAndrewLee, you don't like sublime ?
[07:35] <ILoveAndrewLee> thyriaen: Call it irrational, but I really don't trust javascript based editors.
[07:35] <makara> ducasse: i have the logs. I don't know what's wrong though
[07:35] <thyriaen> algorithmics and complexity of paths with neighbourhood constraints in graphs
[07:35] <ILoveAndrewLee> Also, I am super happy with QtCreator, that I just use that instead.
[07:35] <thyriaen> ILoveAndrewLee, it isnt javascript based, is it ?
[07:36] <thyriaen> its written in c++ i think you are mixing it up with atom, which is pure aweful
[07:36] <ILoveAndrewLee> thyriaen: Is it? Its cross platform from what I am just reading, with a python api, which Python is the worst.
[07:37]  * ILoveAndrewLee buttresses himself for having an unpopular opinion
[07:37] <ducasse> makara, is there no indication why it went to sleep?
[07:37] <thyriaen> python and javascript is the worst, yes
[07:38] <ILoveAndrewLee> thyriaen: Woah, very based. I am surprised. Most people LOVE python and celebrate it as le best beginner language. I think its the worst beginner language.
[07:38] <thyriaen> yea
[07:38] <thyriaen> it is literally the worst beginner language imaginable
[07:38] <ILoveAndrewLee> thyriaen: Come hand out in #sqt and lets be friends. You are cool
[07:39] <thyriaen> sqt ?
[07:39] <thyriaen> what is that about ?
[07:39] <ILoveAndrewLee> thyriaen: Its a 4channel.org/g/ general, /sqt/ which stands for "Stupid Questions Thread"
[07:40] <thyriaen> but there are no stupid questions
[07:40] <ILoveAndrewLee> In other words, its basically the /g/ irc chat.
[07:40] <ILoveAndrewLee> thyriaen: That sounds like a challenge.
[07:40] <thyriaen> ;p
[07:47] <BobbyJr> Im the kind of heathen that takes python code and rewrites it in PowerShell. Mainly as a learning exercise to see if it can be done. Long story short, it can. Kinda.
[07:54] <ILoveAndrewLee> BobbyJr: Powershell.... how what I just. What?
[07:54] <kyle_> hi, while i am trying to mount a nfs share over rdma, i get mount.nfs: an incorrect mount option was specified
[07:55] <kyle_> i see that the nfs server is accepting the connection, and have modprobed the correct modules, and have testing rdma with a ping over the infiniband network
[08:11] <xu-help43w> hi. can anyone help?
[08:12] <xu-help43w> Wi-Fi connection often falls off. help please set up. Network controller: Ralink corp. RT3290 Wireless 802.11n 1T / 1R PCIe
[08:13] <alan_> hello
[08:13] <alan_> is anybody here ?
[08:14] <xu-help43w> yes
[08:15] <alan_> haha
[08:20] <Trieste> Hi, I'd like to install KVM on my machine, so I followed https://help.ubuntu.com/community/KVM/Installation but according to both the grep command and `kvm-ok`, my CPU (i7-7700K) simply does not support any form of virtualization - could that be right? It's a relatively new and powerful CPU, is it possible I missed a BIOS setting or somesuch?
[08:21] <Trieste> To be precise, `egrep -c '(vmx|svm)' /proc/cpuinfo` outputs 0, and `kvm-ok` outputs "INFO: Your CPU does not support KVM extensions"
[08:24] <ravage> Check your bios
[08:48] <kyle_> is nfs over rdma available in the nfs-common package on ubuntu?
[08:50] <Sauvin> Trying to do a do-release-upgrade, can't because some packages are held and won't update. Help?
[08:50] <xu-help43w> hi. can anyone help?
[08:50] <xu-help43w> Wi-Fi connection often falls off. help please set up. Network controller: Ralink corp. RT3290 Wireless 802.11n 1T / 1R PCIe
[08:53] <Mekaneck> xu-help43w: it doesn't fall off, it disconnects
[08:53] <Mekaneck> also
[08:53] <Mekaneck> |patience
[08:53] <Mekaneck> !patience
[08:54] <xu-help43w> yes)
[08:54] <xu-help43w> yes)
[08:55] <Mekaneck> try searching the web for bugs and/or solutions
[08:55] <holgersson> yep
[08:56] <holgersson> xu-help43w: If I just type your full network card name plus "ubuntu" into a search engine I find plenty threads about ralink devices, e.g. https://www.linux.org/threads/wifi-problem-in-ubuntu.32714/
[08:56] <xu-help43w> already tried several options. nothing helped
[08:57] <holgersson> I ommitted the slash and everything past it though: "Ralink corp. RT3290 Wireless 802.11n 1T ubuntu"
[08:57] <holgersson> Looks to me as there will be some manual work necessary for that card.
[09:00] <xu-help43w> I'm not a great Linux specialist
[09:09] <ledeni[m]> xu-help43w: can you run 'nm-connection-editor' in terminal and check your wi-fi connection  -- general -- auto connection is tick ✔️
[09:13] <xu-help43w> yes installed. everything is working well now
[09:49] <ac5tin> i installed xubuntu-desktop and xrdp on my ubuntu server and then i try to connect to it using remmina i get this error:  [1020140:1028342] [INFO][com.freerdp.core] - ERRINFO_LOGOFF_BY_USER (0x0000000C):The disconnection was initiated by the user logging off their session on the server
[09:52] <xu-help43w> only the signal is weak. although the reuter is near
[10:49] <isapgswell> hi
[10:50] <isapgswell> people
[10:50] <isapgswell> avoid notebook firstboot certified to proprietary OS
[10:52] <gry> isapgswell: hi
[10:52] <isapgswell> gry hi
[10:52] <gry> isapgswell: what exactly is its name?
[10:57] <isapgswell> gry name of?
[10:58] <oerheks> "avoid notebook firstboot certified to proprietary OS" is there a question in this?
[10:59] <isapgswell> oerheks it is a hint
[10:59] <oerheks> explain please?
[10:59] <isapgswell> and do not update the firmware before checking https://fwupd.org/
[10:59] <lotuspsychje> i dont know proprietary Os, is it new?
[11:00] <isapgswell> lotuspsychje paid OS
[11:00] <lotuspsychje> i think your hints are more for #ubuntu-offtopic isapgswell ?
[11:00] <oerheks> so there is no support issue then.
[11:00] <isapgswell> an advice
[11:01] <ogra> also ... full sentences would really be helpful (but i told you that before)
[11:01] <isapgswell> i disabled permanent MEI
[11:01] <isapgswell> without reflasing
[11:02] <gry> isapgswell: brand and make of that evil notebook
[11:02] <lotuspsychje> lol
[11:03] <isapgswell> gry sorry about you
[11:05] <isapgswell> ogra no children here, sorry about you too
[11:05] <mort> so, my understanding is that anyone can upload a snap for anything. For example, I could take something like the rustup program, package it as a snap, and submit it - and if no one else has done that, it will be installed by anyone who types `snap install rustup`. Is this correct?
[11:06] <mort> further, anyone who types `rustup` in their terminal would be adviced by the terminal to install my snap, right?
[11:06] <oerheks> mort, yes, if you claimed that name, they will
[11:06] <ogra> correct
[11:06] <mort> what are the processes in place to prevent me from uploading a rustup with malware for example?
[11:06] <mort> I mean rustup is claimed already, but for a package which isn't claimed yet
[11:06] <ogra> thogh note that "well known package names" are usually reserved by default  ...
[11:07] <ogra> so you would need to ask for the name explicitly
[11:07] <ogra> (typically packages that have a deb equivalent in the archive)
[11:07] <mort> right, so rustup probably isn't reserved then
[11:08] <ogra> the registration page has a "dispute this name" button or so ... that you can use to talk to the store people
[11:08] <oerheks> mort, see the snap developers page, how snaps are reviewed
[11:08] <oerheks> https://snapcraft.io/blog/trust-and-security-in-the-snap-store
[11:09] <rud0lf> .ca
[11:09] <rud0lf> disregard
[11:09] <gry> oerheks: who reviews it? volunteers?
[11:09] <oerheks> snapcraft team, and volunteers
[11:10] <mort> what specifically though would prevent me from making a snap of something like rakudobrew, include some subtle malware which is hard to catch through a quick manual review, uploading it to the snap store, and infiltrating the machines of everyone who follows apt's suggestion when they try to `apt install rakudobrew` and apt tells them "try snap
[11:10] <mort> install rakudobrew"
[11:11] <ogra> the builtin snap confinement would prevent you from doing that
[11:11] <ogra> all you could infiltrate is the snap sandbox
[11:11] <ogra> not the host
[11:11] <mort> not really, this would be a kind of snap which it would make sense to use classic confinement with
[11:11] <ogra> classic gets a full manual review
[11:11] <mort> apt would tell people, "do snap install rakudobrew", snap would tell people "do snap install --classic rakudobrew", people would do that and I would get their cryptocurrency wallets
[11:11] <ogra> so your malware would likely be found
[11:12] <mort> I wonder how thorough those reviews are
[11:12] <ogra> classic snaps undergo a very slow, longwinded approval process
[11:12] <ogra> (of several weeks usually)
[11:12] <mort> do you upload binaries or source code + build script
[11:12] <oerheks> you might want to reask in the snappy channel
[11:12] <mort> to the snap store
[11:12] <ogra> you upload a binary snap
[11:12] <ogra> yeah, lets take this to #snappy
[11:13] <mort> right
[11:46] <xu-help43w> Wi-Fi connection often falls off. help please set up. Network controller: Ralink corp. RT3290 Wireless 802.11n 1T / 1R PCIe
[11:46] <xu-help43w> can anyone help?
[11:48] <gry> xu-help43w: check output of "sudo dmesg" in terminal, please
[11:50] <xu-help43w> ready
[11:51] <gry> xu-help43w: put it into a pastebin.
[11:52] <gry> !pastebin
[11:52] <uluntu1> Hi
[11:53] <uluntu1> I have set up isc-dhcp-server over Ubuntu server 20.04. Isc-dhcp-server starts well. But when I plug a network cable to link the server to the switch, I get the following message: network failed - Connection activation failed. Problem described here: https://askubuntu.com/questions/1345088/ubuntu-server-20-04-set-ip-isc-dhcp-server-network-failed
[11:54] <ledeni[m]> xu-help43w:  can you give me url of ip a | grep wl | nc termbin.com 9999
[11:56] <xu-help43w> https://paste.ubuntu.com/p/xVmrZBPY9C/
[12:07] <ledeni[m]> xu-help43w:  'sudo ifconfig wlo1 | nc term.bin 9999' url
[12:07] <c_89> hi
[12:07] <ledeni[m]>  * xu-help43w:  'sudo ifconfig wlo1 | nc termbin 9999' url
[12:08] <ledeni[m]>  * xu-help43w:  'sudo ifconfig wlo1 | nc termbin.com 9999' url
[12:09] <ledeni[m]> xu-help43w: sorry mistake
[12:10] <ledeni[m]> xu-help43w: can you give url of last post
[12:11] <xu-help43w> https://paste.ubuntu.com/p/xVmrZBPY9C/
[12:13] <ledeni[m]> xu-help43w:i mean 'sudo ifconfig wlo1 | nc termbin.com 9999' you just paste in terminal and you will have termbin url
[12:13] <ledeni[m]> xu-help43w: pate it and enter
[12:14] <ledeni[m]> *paste
[12:15] <xu-help43w> command not found. I have xUbuntu
[12:16] <ledeni[m]> xu-help43w: yes sorry  'sudo iwconfig wlo1 | nc termbin.com 9999'
[12:17] <RupertEverton> hello everyone -- i'm looking for an applet as close to this one as possible https://i.imgur.com/y2Y6TVk.png (this is from a screenshot posted on WhiteSur theme) -- basically a way to display all the info in one applet instead of having one for each
[12:21] <ledeni[m]> xu-help43w:  I'm just interested  about your power management is on or not
[12:25] <ledeni[m]> xu-help43w:  and if yes then just run 'sudo iwconfig wlo1 power off' and see how is going on with your connection
[12:31] <xu-help43w> https://termbin.com/v1vu2
[12:33] <ledeni[m]> xu-help43w:  yes it is on  just run 'sudo iwconfig wlo1 power off'
[12:35] <xu-help43w> https://termbin.com/v1vu2
[12:36] <xu-help43w> everything is the same to mine
[12:39] <ledeni[m]> xu-help43w:  yes because i did not ask for url but we can check again 'sudo iwconfig wlo1 | nc termbin.com 9999'
[12:44] <ExeciN> Hi people. Is there a way to rearrange the drives? On some machines, the OS drive is sda and on some others is sdy
[12:46] <ExeciN> fstab only has uuid references so even if the os drive changes from sdy to sda, the os would still boot fine
[12:46] <oerheks> when does " the os drive changes from sdy to sda" ???
[12:47] <oerheks> never heard of such behaviour
[12:48] <ExeciN> well, I want to induce such behavior
[12:48] <oerheks> there is none.
[12:48] <ExeciN> how can I?
[12:50] <lotuspsychje> ExeciN: whats your endgoal exactly with this?
[12:52] <ExeciN> put the os drive as sda, the "cache" drive as sdb and the rest 24 drives as whatever, repeat for all 5 machines, configure said 5 machines using identical commands
[12:54] <MrMobius> is there an existing PXE server I can use to install ubuntu server?
[12:54] <oerheks> first drive is always sda, so when does it change to sdy ?
[12:55] <ExeciN> oerheks: I'm not exactly sure what you mean with "first drive" but in any case, I want to change sdy to sda
[12:55] <oerheks> MrMobius, no, but you can easily setup one yourself?
[12:56] <MrMobius> oerheks, kind of a chicken and egg problem
[12:58] <oerheks> it is more simpler than that, there is no pxe service on the www AFAIK, you have to do that yourself
[13:05] <ravage> the closest thing to a public server is https://netboot.xyz/ . but you still need to boot that somehow
[13:07] <chalcedony> the community moved over here, right?
[13:08] <Chunkyz-RPi> chalcedony: sure did.
[13:08] <ExeciN> https://wiki.ubuntu.com/IRC/ChannelList
[13:09] <Chunkyz-RPi> ExeciN: has that even been updated?
[13:10] <oerheks> Chunkyz-RPi, yes, 2 lists for freenode and this server
[13:10] <ExeciN> it mentions this very channel on libera
[13:10] <oerheks> easy to read yourself..
[13:10] <Chunkyz> oerheks: don't have a web browser currently installed hence not reading it. :-)
[13:18] <spinningCat> halo
[13:18] <spinningCat> what did i do wrong here? https://bpa.st/M5WQ
[13:19] <oerheks> i see makeup errors, a systemd unit is type sensitive
[13:20] <oerheks> compare with https://www.oreilly.com/library/view/odoo-11-development/9781788477796/04962a35-0ab6-4519-a279-4afc27992049.xhtml
[13:20] <spinningCat> makeup errors?
[13:20] <oerheks> [Install]   WantedBy=default.target = on the same line
[13:21] <spinningCat> oh
[13:21] <oerheks> and there is a tool to check the unit
[13:21] <spinningCat> damn
[13:21] <wez> spinningCat: .o/
[13:22] <spinningCat> what tool
[13:23] <spinningCat> how goez wez  warez
[13:24] <oerheks> oh wait, that is for netplan > sudo netplan try
[13:24] <wez> spinningCat: How goes cycling
[13:24] <wez> I don't warez
[13:24] <spinningCat> morning and evening
[13:24] <wez> I pay for content
[13:24] <spinningCat> like 50km a day
[13:24] <wez> oh>
[13:25] <wez> I have been sick for the last 3 weels
[13:25] <wez> and it has been raining all this week
[13:25] <oerheks> these are validators https://www.liquidweb.com/kb/validate-yaml/
[13:25] <oerheks> !ot
[13:25] <wez> I did do some 1year old bench presses though
[13:25] <wez> 12Kg weight
[13:25] <oerheks> wez, please, use an offtopic channel for chit chat
[13:26] <wez> oerheks: oh sorry, I thoiught i was in web-social
[13:26] <wez> where I usallty speak to spinningCat
[13:26] <Mekaneck> !ot | wez
[13:26] <wez> Mekaneck: Thnx mate, oerheks beat you to it
[13:28] <Mekaneck> just to make clear where you can chit chat
[13:28] <wez> Mekaneck: cheers
[13:55] <chalcedony> My husband is having a problem with a new setup of Ubuntu 20.04. He's still running the live version on a usb stick and the hard drive. It can't see the Internet.
[13:55] <chalcedony> i'm on the internet fine, his other computer is fine, that one says no internet.
[13:55] <chalcedony> i asked him to check the cord - wired lan, and it's plugged in.
[13:56] <chalcedony> it's not a new computer and it's been fine previously.
[13:56] <chalcedony> it can't ping, doesn't have traceroute.
[13:59] <sixwheeledbeast> "lshw -C network" should show ethernet device?
[14:01] <coz_> sixwheeledbeast, lspci   ?
[14:01] <chalcedony> i'll take a look, thanks
[14:02] <sixwheeledbeast> only if it's pci, it maybe builtin
[14:04] <chalcedony> sixwheeledbeast, does it matter if i use sudo for that?
[14:05] <sixwheeledbeast> should show without but best with sudo yes
[14:10] <yuzi> hello
[14:12] <chalcedony> sixwheeledbeast, it shows the wireless interface first, then the gigabit ethernet controller. could it be trying to run wifi?
[14:14] <chalcedony> i took pictures of his screen but i'm not good with uploading pictures. if you tell me what to copy, i can do that.
[14:16] <sixwheeledbeast> thats just a list of detected hardware for the system.
[14:16] <chalcedony> ok
[14:17] <sixwheeledbeast> check the connection information for that, not sure where that is with vanilla ubuntu, possible in a panel top right?
[14:17] <sixwheeledbeast> do you have a local ip?
[14:17] <sixwheeledbeast> "ip addr"
[14:17] <chalcedony> i'll try that
[14:17] <sixwheeledbeast> or ifconfig
[14:18] <chalcedony> sure
[14:19] <sixwheeledbeast> your looking for a connection with letters and numbers like enp3s0 or what have you
[14:22] <chalcedony> it doesn't have ifconfig and can't download it
[14:22] <chalcedony> ip addr seems to give me a v6 address
[14:24] <chalcedony> 1: lo: <LOOPBACK,UP LOWER UP, mtu 1500 qdisc fq_codel state UP group default qlen ..
[14:25] <chalcedony> i think what i am is too tired to do this.
[14:26] <chalcedony> i'll be back later thanks sixwheeledbeast :)
[14:41] <eoli3n> Hi, nautilus doesn't detect a samsung phone, libmtp is installed, don't know what to check
[14:41] <eoli3n> (connected with usb)
[14:48] <oerheks> unlock the phone before connecting perhaps?
[14:49] <oerheks> then in terminal: mtp-detect
[14:51] <oerheks> maybe you have to set enabled USB Debugging in Developers Options on your device too..
[14:52] <eoli3n> oerheks tried the unlock tip with mtp-detect, it doesn't see any "raw device"
[14:52] <eoli3n> i'm rebooting phone and pc to be sure, because i installed mtp-tools (which was missing)
[14:53] <eoli3n> oerheks USB debugging trick, not sure i need this
[14:53] <eoli3n> but lets test
[14:53] <[code]> phone has to be unlocked too
[14:54] <oerheks> [code], indeed
[14:54] <oerheks> that was my 1st hint
[14:55] <[code]> ah i see.
[15:03] <eoli3n> how to be sure i need USB Debugging ?
[15:04] <eoli3n> that's for my mother huhu, i'm debugging her at phone
[15:04] <eoli3n> enable developers options by leading my mother with my voice is funny
[15:04] <eoli3n> but she did not succeed XD
[15:05] <eoli3n> "search for build number and click 7 times on it" she " there is no menu" me "yes this is a hidden one, clic on build number"
[15:14] <eoli3n> ok usb debugging check
[15:14] <eoli3n> when she connects usb, nothing prompt on android, and i run through ssh "watch -n1 mtp-detect" which see nothing
[15:26] <eoli3n> i can't even see usb connected with lsusb, she's searching for another usb cable
[15:28] <TomyWork> I assume you checked other ports too?
[15:28] <eoli3n> i did
[15:29] <eoli3n> she tried 3 usb ports
[15:31] <TomyWork> the quantum superpositioned ports too?
[15:33] <eoli3n> only usb ones :)
[15:35] <heap> hello whe ubuntu package of zfs-dkms is the same as i download zfs from github and compile it manually?
[15:36] <eoli3n> that was the cable
[15:36] <eoli3n> with a new cable everythings ok
[16:05] <bandit-led> how to keep okular and libepub fom filling up .xsession-errors with cruft from opening files?
[16:12] <tomreyn> bandit-led: can you share some of those logs on a pastebin?
[16:14] <tomreyn> heap: zfs.dkms will compile the kernel modules for the currently running kernel, based on the module version released with this ubuntu release. you will likely find both older and newer module versions somewhere on github
[16:15] <tomreyn> bandit-led: make sure you also discuss the ubuntu version and flavour you're using, and relevant customizations (such as a different default desktop environment, X server)
[17:05] <grobkorn> hello channel. need some advice with mounting a NFS drive from synology... installed nfs-common... created a directory in /mnt .. after using mount i cant access to that directory... ubuntu 20.10... any hints?
[17:08] <oerheks> grobkorn, this is the whole tutorial, nas-part, install part, mount part, and more important; file permissions https://acceptdefaults.com/2020/03/16/mounting-a-synology-share-in-ubuntu/
[17:08] <oerheks> after any group changes, logout/login
[17:08] <gt756> grobkorn, it depends on exactly how your mount command looked like, and whats in the logs
[17:09] <gt756> p.s. and on permissions (as oerheks has righly noted)
[17:11] <grobkorn> mounting via CLI --> mount AND fstab... both has the same result
[17:11] <grobkorn> done everything like in that tutorial... mounting itself comes with no errors.. but cant access the directory after mounting
[17:12] <grobkorn> chmod says - nop permission after mounting
[17:15] <gt756> wondering, why would you use chmod after unsuccessful mounting? =#
[17:15] <gt756> what for?
[17:19] <grobkorn> because the posted tutorial says to use chmod if you dont have the permission, right?
[17:20] <grobkorn> have the exact same entries in my fstab on my laptop... everything donel like described in the tutorial... it works perfect on my laaptop,, but not on my pc... both ubuntu 20.10... i have no clue :///
[17:25] <gt756> isn't it that you have to have right permissions on your mount point directory, first. Before mounting?
[17:26] <gt756> And if your mount failed, or you don't have access to mount point after mounting - your chmoding is useless. It neither goes to target, neither to mountpoint.
[17:29] <gt756> But of course that recursive chmodding will work if (and only after) mount was successful, and you can enter mount point dir. Even if it's weird to override all permissions to shared resource each time new user mounts it - it becomes stupid wars of users owning it.
[17:42] <grobkorn> aaah ... thats so stupid... synology changed the SQUASH settings during update to DSM7.... sorry - now everything works... thank u all
[17:47] <raub> You are scaring me: I have a synology box I have not updated recently
[18:34] <bandit-led> how to keep okular and libepub fom filling up .xsession-errors with cruft from opening files? sorry for not adding extra information.   Ubuntu 21.04 hirsute (x86-64) https://paste.ubuntu.com/p/HbkrGRkFRR/
[18:39] <Chunkyz-RPi> anyone ever got this with lightdm? Process: 767 ExecStartPre=/bin/sh -c [ "$(basename $(cat /etc/X11/default-display-manager 2>/dev/null))" = "lightdm" ] (code=exited, status=0/SUCCESS)
[18:42] <bandit-led> host info etc https://paste.ubuntu.com/p/sYjNpM4hrC/
[18:46] <jfcaron> I am on Ubuntu 20.04 LTS but I require packages that are apparently only available up to 18.04 LTS.  Is there a way to install the old packages on the new system?
[18:46] <bandit-led> what are the packages?
[18:47] <jfcaron> qt4-default and all the dependencies
[18:47] <bandit-led> whats the app your trying to install?
[18:47] <jfcaron> (and yes I know qt4 is like a bajillion years old)
[18:48] <jfcaron> it's a custom GUI for talking to a data acquisition hardware
[18:48] <jfcaron> I need to compile the custom GUI and it requires qt4 (and QWT, qtcanvas)
[18:49] <jfcaron> It was written by a lab member many years ago, no one is available to overhaul it to Qt5.
[18:49] <jfcaron> I tried compiling Qt 4.8 from the .tar.gz but I get a compile error I don't understand, I figured maybe it was possible to install the .deb from 18.04
[18:49] <oerheks> focal .. https://launchpad.net/~rock-core/+archive/ubuntu/qt4
[18:50] <jfcaron> oh, awesome oerheks, thanks.
[18:50] <oerheks> use at your own risk.. but you seem to have a legit reason, have fun!
[18:50] <bandit-led> i would also say not to do that
[18:51] <bandit-led> whats the last lts that has qt4?
[18:51] <oerheks> build such in a vm
[18:51] <bandit-led> if there is one even
[18:51] <jfcaron> 18.04 LTS apparently still has qt4
[18:52] <oerheks> https://launchpad.net/ubuntu/+source/qt4-x11
[18:52] <oerheks> jups
[18:52] <jfcaron> but I had trouble with stuff that needed QtCanvas even there
[18:53] <bandit-led> your going to have alot of issues trying to do what you want
[18:53] <bandit-led> there is no updated version of the app your trying to use?
[18:53] <jfcaron> It was written by a grad student like 8 years ago
[18:53] <bandit-led> might be better to install the version that has qt-4 then post the error your getting trying to do what your doing
[18:54] <jfcaron> I tried that before in #qt but everyone just told me to stop using qt4.
[18:54] <jfcaron> It's like asking python 2.x questions in #python, good luck getting any sympathy.
[18:54] <bandit-led> then your going to have to do the vm route to use the app or contact the grad studient
[18:55] <jfcaron> My eyes are set on the "this is my last week at this job" solution right now.
[18:55] <jfcaron> Oh good, the qt4 from that PPA is going to uninstall qt5 on my system.  I'm sure that won't break anything. = )
[18:56] <bandit-led> maybe do a full backup before you start throwing ppa's at it
[18:56] <sixwheeledbeast> that's often the issue with this stuff tho, kick the can down the road...
[18:56] <oerheks> again; build such in a vm
[18:57] <sixwheeledbeast> I'd say vm too
[18:57] <jfcaron> If the bigshot professors with a million dollars in funding can't be bothered to hire software professionals to maintain their stuff...
[18:57] <jfcaron> We have another physicist doing all the networking in the lab too.
[18:57] <sixwheeledbeast> at least the main system is not compromised.
[18:57] <bandit-led> yes but your talking about software from a grad student
[18:58] <jfcaron> Is the software in a VM likely to be able to talk to a card in a PCI-Express slot?
[18:58] <oerheks> pci pass through, sure
[18:59] <bandit-led> depends on the device and what your trying to do but that a different question
[19:00] <jfcaron> God I thought the hard part would be figuring out which version among foo_dev, foo_build, foo_build2 of the code to compile.
[19:00] <jfcaron> Cuz obviously it's been customized over the years and never documented or put into a version system.
[19:02] <jfcaron> How did you get that launchpad URL for the latest Ubuntu that had a package?
[19:02] <jfcaron> If I'm doing a VM anyways I would install a really old one that has QtCanvas
[19:02] <jfcaron> But I don't even know the name of the package.
[19:03] <sixwheeledbeast> Backup the system you have now and turn that into the VM?
[19:04] <jfcaron> No, the same system needs to be able to talk to other equipment with more-modern interfaces (the HV power supply uses Qt5 for its GUI)
[19:04] <jfcaron> Or maybe I'll give up.
[19:05] <tomreyn> don't give up, solve the actual problem
[19:05] <tomreyn> unmaintained / unmaintainable software which the solution you want to use depends on.
[19:05] <jfcaron> lol ok I'll fix the actual problem of postdocs only being given 2 years to work on projects and academics never bothering to document their code or do anything properly because there is no time.
[19:06] <jfcaron> I have until Friday next week. =)
[19:06] <sixwheeledbeast> solve problems you can actually solve.
[19:07] <tomreyn> maybe not this one then. ;)
[19:07] <sixwheeledbeast> I dont see how the "qt5 HV supply" effects my solution but ok.
[19:08] <sixwheeledbeast> affects
[19:08] <sixwheeledbeast> meh
[19:09] <sixwheeledbeast> if you have docs for that it can be installed on the same system not the vm, or another vm etc etc.
[19:09] <Guest64> why cannot I putty into the machine running at runlevel 3 ?
[19:10] <jfcaron> I can't even find what package I'd have to install to get "QtCanvas" working, there's more modern stuff called QtCanvas3D that swamps all the search results.
[19:12] <sixwheeledbeast> sounds like a qtwebkit thing?
[19:13] <jfcaron> Looks like it was a backported Qt3 thing for Qt 4.1
[19:13] <jfcaron> https://dot.kde.org/2005/12/20/qt-41-now-available (Ctrl+F for qtcanvas)
[19:13] <jfcaron> I can find a download for 4.1
[19:13] <jfcaron> https://download.qt.io/archive/qt/4.1/
[19:13] <tomreyn> Guest64: if "the machine" is a system running a supported Ubuntu release (see /topic) and by "putty into" you mean "connect to, through the ssh protocol, with the putty ssh client", then it's probably because the ssh server is not installed / running on the Ubuntu system, or due to unsuccessful authentication, or for lack of permissions.
[19:13] <jfcaron> So maybe I could set up a VM with an Ubuntu version that has Qt 4.1 in the repos.
[19:15] <tomreyn> Guest64: the error message you get threre may help narrowing this down more.
[19:16] <oerheks> jfcaron, xenial perhaps, and use the old-versions trick from the EOL upgrade factoid
[19:16] <oerheks> !eol
[19:18] <jfcaron> !xenial
[19:19] <jfcaron> oerheks What's the EOL upgrade factoid?
[19:19] <oerheks> jfcaron, you would end up with "deb http://old-releases.ubuntu.com/ubuntu/ xenial main restricted universe multiverse
[19:19] <oerheks> https://help.ubuntu.com/community/EOLUpgrades#Update_sources.list
[19:19] <jfcaron> In a 16.04 VM, or on my 20.04 machine?
[19:20] <oerheks> no, in the vm ofcourse
[19:20] <oerheks> there is just 1 server online, for such old versions
[19:27] <jfcaron> Ok I will make a 16.04 VM and try to get it working on that.
[19:27] <jfcaron> I hope I can undo the damage from this PPA.
[19:30] <ash_worksi> I just watched 2 videos where people pronounce gnome "gah-nome"... is that really how people pronounce this?
[19:32] <sarnold> I say something more like "guh-nuhm"
[19:33] <jfcaron> g'nome mate
[19:33] <sarnold> crikey
[19:33] <oerheks> gno-mé
[19:34] <sarnold> gno-me, gno-myou, gno-mus
[19:34] <jfcaron> gee-nom
[19:34] <Mekaneck> offtopic.....
[19:37] <ash_worksi> oerheks sarnold I'm not sure if you're saying you pronounce the g or not
[19:37] <ash_worksi> I'm just stunned anyone pronounces it
[19:38] <krytarik> I guess it depends a little on what your first language is too..
[19:38] <ash_worksi> to me, it's like pronouncing "what", w-hat
[19:39] <Mekaneck> ash_worksi: that chit chat belongs in #ubuntu-offtopic
[19:39] <ash_worksi> sorry Mekaneck; I thought the question was trivial enough to be done by now
[19:40] <Mekaneck> it's a ubuntu support channel, you should know better ;)
[19:41] <ash_worksi> Mekaneck: you're right
[19:48] <krytarik> Mekaneck: I feel like there is a bit much attitude swinging alongside though.
[19:50] <Mekaneck> depends on how you pick it up
[19:50] <Mekaneck> anyhoo, have a nice eveing
[19:50] <Mekaneck> *evening
[19:51] <Linkandzelda> what would cause my ssd write speed to be less than 10MB/s on an idling server with no resources?
[19:53] <Walex> Linkandzelda: O_PONIES :-)
[19:53] <Walex> Linkandzelda: small committed writes.
[19:54] <Walex> Linkandzelda: small committed writes on flash SSD without persistent buffers.
[19:56] <Linkandzelda> Walex: im not sure what that means, could you explain?
[19:57] <Walex> Linkandzelda: web searching helps.
[19:58] <oerheks> why would the writespreed higher on an idle system?
[19:58] <oerheks> seems like to me that it does not do that much.
[19:58] <Walex> Linkandzelda: it is a very common issue. Look for benchmarks for 4K random writes on a consume flash SSD.
[19:59] <oerheks> 10 mb/s measured over a period of time..
[19:59] <Walex> pretty good actually.
[19:59] <oerheks> so, is your server unresponsive ?
[19:59] <Walex> many flash SSDs cannot do as much as 10MB/s on small committed writes.
[20:00] <Linkandzelda> its a fresh install of ubuntu, no services, no data
[20:00] <Walex> it is 20 times faster than on HDDs.
[20:00] <Walex> that 10MB/s is still 20 times faster than on HDDs.
[20:00] <Linkandzelda> im using fio to do a random write test: fio --name=random-write --ioengine=posixaio --rw=randwrite --bs=4k --numjobs=1 --size=4g --iodepth=1 --runtime=60 --time_based --end_fsync=1
[20:00] <Linkandzelda> the same test on mt laptop gives me 400MB/s write
[20:01] <Walex> Linkandzelda: have you heard of something called "Google"? I am told it contains a lot of information.
[20:01] <sarnold> Walex: that's not super-helpful
[20:01] <Linkandzelda> Walex: yea, not helpful at all
[20:02] <Linkandzelda> i come to IRC for targetted help, in case someone knows exactly the same issue, after googling
[20:02] <Walex> sarnold: Linkandzelda: it is pretty basic, storafge 101 information, that can be foung easily just with the keywords I mentioned.
[20:03] <Walex> Linkandzelda: and it is not even related to Ubuntu.
[20:04] <Walex> The explanatiobn takes several paragrpahs, and I think that to be helpful I would have to do the web search myself, and I cannot understand why the person requesting the information has not done that already.
[20:05] <Walex> Note that I also already gave the obviously correct answer in detail "small committed writes on flash SSD without persistent buffers". has someone else done more? They are welcome to try.
[20:15] <Linkandzelda> Walex: still nowhere with it, but i'll stop asking if its not related to ubuntu directly
[20:18]  * Walex has looked at https://www.google.com/search?num=50&as_q=small+committed+writes+on+flash+SSD+without+persistent+buffers and the results are pretty explicit
[20:22] <Linkandzelda> Walex: you are basically telling me "thats how it is"?
[20:32] <Linkandzelda> Walex: im just not making any sense of what you are pointing me at. I am only trying to assertain if there's a problem with this system or the hardware. what pointed me to believe there was is that im getting faster sequential writes to an ssd connected via usb 2 to a raspberry pi than an internal sata ssd. 101 or not, that does not make sense
[20:33] <sarnold> how's smart output look on the drive?
[20:34] <Linkandzelda> sarnold: basic health check passed
[20:35] <sarnold> Linkandzelda: and the error counts?
[20:37] <Linkandzelda> sarnold: looks like a very old drive, power on hours 4446 and all marked as "Old-age" and one "Pre-fail"
[20:38] <sarnold> that isn't even half a year of uptime..
[20:41] <Linkandzelda> sarnold: this is the report https://hastebin.com/wubabiqoqa.yaml
[20:41] <sarnold> 199 UDMA_CRC_Error_Count    0x0032   100   100   050    Old_age   Always       -       26573
[20:42] <sarnold> I wonder if that's "bad cable" kind of thing
[20:43] <Linkandzelda> sarnold: thats totally possible, gonna swap the cable on the drive now and see if that has any difference
[20:43] <sarnold> two ssds I've got with 38559 hours power-on time report zero udmo crc error counts
[20:57] <webchat79> I'm on an Ubuntu focal.
[20:57] <webchat79> I just found out I have 2 docker containers running on my machine for mysql and ghost.
[20:57] <webchat79> I vaguely remember having played with them a month ago but no more.
[20:57] <webchat79> Where should I look to find how they startup everytime I restart?
[20:58] <webchat79> doing a `docker ps --no-trunc` give me the following:
[20:58] <webchat79> ```
[20:58] <webchat79> ```
[20:59] <webchat79> ```
[20:59] <webchat79> CONTAINER ID                                                       IMAGE            COMMAND                                        CREATED       STATUS        PORTS                                       NAMES
[20:59] <webchat79> 238ae6052d38c59c2f72de51672a0fad9b986ea534bb0b4291bc04efd77e1b98   ghost:3-alpine   "docker-entrypoint.sh node current/index.js"   4 weeks ago   Up 10 hours   0.0.0.0:8080->2368/tcp, :::8080->2368/tcp   newtest_ghost_1
[20:59] <webchat79> 91f6c45310654bf9c3c9d46de52b0d1608aa07b759e0cad40af2cd3cccf9b519   mysql:5.7        "docker-entrypoint.sh mysqld"                  4 weeks ago   Up 10 hours   3306/tcp, 33060/tcp                         newtest_db_1
[20:59] <webchat79> ```
[20:59] <tomreyn> !paste | webchat79
[21:00] <webchat79> got it, sorry, I'm more used to Discord
[21:01] <tomreyn> no problem then. ;-) i'm afraid i don't know your very answer, i'm not too much into docker.
[21:01] <tomreyn> docker itself will be a systemd unit
[21:01] <webchat79> https://paste.ubuntu.com/p/Jk8CRgVmBn/
[21:01] <Slart> from a quick google-search it seems like docker handlers much of this itself.. there's also ways to do it from systemd etc...  I would start checking for docker commands
[21:02] <webchat79> This is the output of `docker ps --no-trunc`
[21:02] <tomreyn> https://docs.docker.com/config/containers/start-containers-automatically/
[21:02] <webchat79> I looked into this file from that output `docker-entrypoint.sh`
[21:03] <webchat79> and found a bunch of them in the the `overlay2` subfolders that are managed by docker
[21:04] <tomreyn> see the link i posted above, i think it explains how to choose which ones gets started automatically
[21:04] <webchat79> O God
[21:05] <webchat79> So, this means that 4 weeks ago, I ran some containers with `--restart unless-stopped` flag
[21:05] <webchat79> and everytime I restart my machine they also startup again bc I haven't stopped them specifically
[21:05] <webchat79> I just forgot about them
[21:06] <webchat79> Gonna test that hypothesis
[21:06] <webchat79> Restart my pc and see them containers running again
[21:06] <webchat79> then stop them by command and restart again
[21:06] <webchat79> if they don't start it's correct
[21:06] <webchat79> Thanks a lot
[21:07] <tomreyn> you're welcome
[21:07] <webchat79> btw, is this the right/best channel for asking these normal linux/ubuntu question
[21:07] <webchat79> ?
[21:08] <tomreyn> if you run an ubuntu release that's supported, yes
[21:08] <webchat79> thanks, that's a relief, ran this question on a lot of discord servers with no response
[21:08] <tomreyn> for general linux questions, you'd head over to the #linux channel, and to #docker for docker (i assume, have not checked this exists)
[21:09] <webchat79> thanks again
[21:29] <svm_invictvs> I've got a docker container running ubuntu, how do I know which version it is?
[21:29] <ZeZu> lsb_release -a
[21:30] <svm_invictvs> No command lsb
[21:30] <ZeZu> either you forgot the '_' between lsb and release
[21:31] <tomreyn> cat /etc/os-release    also works
[21:31] <ZeZu> or it's not running ubuntu and is extremely cracked out for not seeing the _ and giving full cmd in result
[21:31] <svm_invictvs> Yeah no lsb_release
[21:31] <svm_invictvs> Yeah my IRC client doesn't render underscrores for some reason
[21:31] <svm_invictvs> Linux e8c494837310 5.8.0-55-generic #62~20.04.1-Ubuntu SMP Wed Jun 2 08:55:04 UTC 2021 x86_64 GNU/Linux
[21:31] <ZeZu> 20.04
[21:31] <svm_invictvs> That's my kernel, though right?
[21:32] <ZeZu> yeah
[21:32] <svm_invictvs> Which shares with the host machine.
[21:33] <ZeZu> I forget the docker containers can be very minimal and might not have lsb package,  i never use them : perhaps it's normal
[21:34] <svm_invictvs> Yeah
[21:34] <svm_invictvs> Yeah i"M not using the official ubuntu docker image
[21:34] <tomreyn> it's not ubuntu if it doesn't comply with ubuntu defaults, though,probably some derivative which just claims to be ubuntu.
[21:34] <svm_invictvs> okay
[21:35] <svm_invictvs> tomreyn, The host machine is Ubuntu20.04 right? So uname is sharing hte kernel of the host machine
[21:35] <svm_invictvs> uname doesn't tell me what I want.
[21:36] <svm_invictvs> And my cat just bit my face lol
[21:37] <tomreyn> svm_invictvs: i can't watch over your shoulder (nor bite your face, you're safe there!), so couldn't tell what the host is running.
[21:37] <Walex> Linkandzelda: "a problem with this system or the hardware" there is neither a problem with the system or hardware, as I wrote you have a "flash SSD without persistent buffers". Some flash SSDs have them, and some don't. Perhaps another useful keyword for searching is "supercapacitor" and "SSD"
[21:45] <svm_invictvs> .
[21:45] <svm_invictvs> tomreyn, I just change the base container to explicitly use ubuntu 20.04
[21:46] <svm_invictvs> tomreyn, I was originally using the openjdk 11 containers. It's easy enough to just switch to ubuntu and then manually install the jdk.
[21:46] <svm_invictvs> I guess I could make alpine editions of my containers which is what all the cool kids do, but that's well beyond the scope of this discussion
[21:55] <Linkandzelda> Walex: its a cheap ADATA 120gb so i believe that it doesnt have good components
[21:56] <Walex> Linkandzelda: it is not good or bad components, it is whether it has a persistent write buffer (and good firmware).
[21:57] <Walex> Linkandzelda: usually only "enterprise" SSDs have a persistent write buffer, backed by supercapacitors.
[21:57] <Linkandzelda> Walex: and what about NVMe SSDs?
[21:57] <Walex> Linkandzelda: the same, it is a limitation of the flash chips.
[21:58] <Linkandzelda> Walex: and HDDs?
[21:58] <Walex> Linkandzelda: much worse, because of mechanical positioning.
[21:59] <Walex> Linkandzelda: what do you think is the typical HDD write rate on comitted 4KiB random writes?
[22:00] <Walex> very few people seem to have done the relevant multiplication and give the right number :-)
[22:00] <Walex> Linkandzelda: please guess a number for the HDD case
[22:01] <Walex> Linkandzelda: for HDDs there is a big difference in many case between enabling or disabling the HDD write buffer with 'hdparm', and it is similar for flash, only more/less bad.
[22:03] <Linkandzelda> Walex: im not sure i could guess, without doing the tests
[22:03] <Walex> Linkandzelda: put another way, what is the random IOPS rate of an HDD, usually?
[22:04] <Walex> for HDDs read and write IOPS rates are the same...
[22:04] <Linkandzelda> Walex: ive never looked, the only thing i can tell you is the typical everyday use read/write and what i expect or have seen
[22:04] <Linkandzelda> i just know that HDDs are slow, and SSDs are fast haha
[22:05] <Walex> Linkandzelda: then a detailed explanation would indeed be a bit long... Anyhow some details
[22:05] <Linkandzelda> Walex: a 4kib text on pi with usb2 hdd: WRITE: bw=74B/s (74B/s), 74B/s-74B/s (74B/s-74B/s)
[22:06] <Linkandzelda> test*
[22:06] <sarnold> 74 bytes per second?
[22:06] <Linkandzelda> apparently
[22:06] <Walex> Linkandzelda: a typical HDD can do around 100 random IOPS, and with a 4KiB write size, that is 0.5MB/s.
[22:06] <Walex> sarnold: Linkandzelda: that probably is blocks.
[22:06] <sarnold> even random 4k writes that feels shockingly low.. you usually get around 100 random ops, so 400kbps is about what I'd expect...
[22:06] <sarnold> Walex: oh that makes a lot more sense
[22:06] <Linkandzelda> Walex: this is a usb
[22:06] <Linkandzelda> oh wait, yes, its B = blocks
[22:07] <sarnold> even ancient usb ought to be able to do 400kbps :)
[22:07] <Linkandzelda> block size is 4k
[22:07] <Walex> Linkandzelda: and the RPi USB2 is famously crap.
[22:07] <Linkandzelda> so... 74 * 4k per s?
[22:07] <Walex> Linkandzelda: yes.
[22:07] <Linkandzelda> Walex: i know, but its better than it's usb 3...
[22:08] <Linkandzelda> what i dont understand is the same test on the server in question using a usb hdd gives this result: WRITE: bw=6327KiB/s (6479kB/s), 6327KiB/s-6327KiB/s (6479kB/s-6479kB/s)
[22:08] <Linkandzelda> is that kbytes or kBlocks?
[22:08] <Walex> Linkandzelda: anyhow what really slows down all sort of storage is "committed", that is implicit or implicit
[22:09] <Walex> Linkandzelda: anyhow what really slows down all sort of storage is "committed", that is implicit or implicit 'fsync'.
[22:09] <Walex> Linkandzelda: that#s kibibytes. It is too high.
[22:09] <Walex> Linkandzelda: probably it does not do proper committing.
[22:10] <Linkandzelda> Walex: its fio, i thought that was an amazing tool to check this kind of thing
[22:10] <Walex> Linkandzelda: USB2 mass storage protocol is not good for data integrity.
[22:10] <Walex> Linkandzelda: 'fio' is excellent, but you have to know what the various knobs mean.
[22:11] <Linkandzelda> Walex: fio --name=random-write --ioengine=posixaio --rw=randwrite --bs=4k --size=4g --numjobs=1 --iodepth=1 --runtime=60 --time_based --end_fsync=1
[22:12] <Linkandzelda> Walex: end_fsync means it syncs at the end of the test
[22:12] <Walex> Linkandzelda: those are plausible but perhaps they don't measure what you think. For example do you know how SSDs are likely to respond to increased 'iodepth' and increased 'bs'?
[22:13] <Walex> Linkandzelda: 'end_fsync' is "optimistic" as it makes the results too dependent on how much memory you got.
[22:13] <Linkandzelda> Walex: i know that this is very complex, and different setups and hardware cause completely different results than what a typical situation would be like
[22:13] <Linkandzelda> this is the SSD in question with the same test: WRITE: bw=9361KiB/s (9586kB/s), 9361KiB/s-9361KiB/s (9586kB/s-9586kB/s)
[22:14] <Walex> Linkandzelda: it is 20 times faster than a typical HDD.
[22:14] <Linkandzelda> Walex: what would you expect from the same test on an NVMe?
[22:15] <Walex> Linkandzelda: same, if it does have committed writes and same memory size.
[22:15] <Walex> Linkandzelda: a 4G test file may be larger than RAM or a lot smaller.
[22:15] <Linkandzelda> Walex: server ram is 2GB, pi 4GB
[22:16] <Linkandzelda> NVMe is a laptop with 16GB ram
[22:16] <Linkandzelda> Walex: the NVMe results WRITE: bw=433MiB/s (454MB/s), 433MiB/s-433MiB/s (454MB/s-454MB/s)
[22:17] <Linkandzelda> am i reading the results wrong or this lightening speed?
[22:18] <Walex> you have already mentioned these numbers before, and I have already mentioned the 20 times difference BTW
[22:18] <Walex> Linkandzelda: that is not necessarily the actual write rate... RAM caching.
[22:19] <Walex> Linkandzelda: while you run that test consider also running 'iotop'
[22:25] <Walex> Linkandzelda: what's the actual write rate reported by 'iotop'?
[22:29] <Linkandzelda> Walex: sorry, a thunderstorm happened and the network went down
[22:30] <svm_invictvs> When configuring sshd, is tehre a way to specify the environment for the user session?
[22:30] <Walex> "just for fun" I have done a 'fio' run on my amazing 2TB NVME flash and I am getting random write rates of 50-70KiB/s...
[22:31] <Linkandzelda> Walex: iotop reports values around 350-600MB/s for both total and actual
[22:31] <Linkandzelda> thats for the nvme
[22:32] <Walex> Linkandzelda: that's pretty good, but then I suspect that it is not doing committed writes.
[22:32] <Linkandzelda> maybe it just has amazing cache or buffers or something, handled by the firmware itself
[22:32] <Linkandzelda> something the OS doesnt know about
[22:32] <Linkandzelda> and the others dont, which give real values
[22:33] <Walex> Linkandzelda: the problem is that manufacturers deliberately omit the persistent buffer from consumer drives to force enterprise users to buy enterprise drives.
[22:33] <Linkandzelda> either way i understand now a bit better what to expect from these tests, but hope you also know why i was original alarmed at the different result and why i felt my SSD had a problem :)
[22:33] <Walex> Linkandzelda: sure, but that's how things are.
[22:34] <Walex> Linkandzelda: and there are some papers and benchmarks that show that :-)
[22:35] <Walex> Linkandzelda: BTW drive buffering on/off can also be set in BIOS. Many laptops have the default to "on".
[22:35] <Linkandzelda> Walex: come to think of it i disabled write cache in the bios of this server
[22:36] <Walex> Linkandzelda: then it should be slower, unless the firmware ignores that.
[22:36] <Linkandzelda> should i re-enable it?
[22:37] <Walex> Linkandzelda: only to test if it makes any difference.
[22:37] <Linkandzelda> i only disabled it because unfortunately the front 4 bays are broken and i was trying to get them working
[22:40] <Walex> Linkandzelda: I just remembered that this page discusses small committed write rates on SSD and has some simple test results: https://www.sebastien-han.fr/blog/2014/10/10/ceph-how-to-test-if-your-ssd-is-suitable-as-a-journal-device/
[22:41] <Walex> Linkandzelda: as you can see the "1 JOB" write rates go from "hundreds" to "a few" MB/s.
[22:42] <sarnold> oh that's cool page
[22:43] <Walex> sarnold: Linkandzelda: this explains supercapacitors: https://www.anandtech.com/show/11765/liteon-ssds-at-fms-new-controllers-and-tsv-nand-packaging
[22:43] <Walex> sarnold: that page on Ceph journaling also applies to other types of journaling.
[22:44] <sarnold> Walex: yeah I figured it'd be useful for zfs slog devices, too
[22:44]  * Walex has just remember where he had put some relevant saved links
[22:44] <sarnold> well, useful for filesystems in genheral, but specifically useful for slogs etc
[22:44] <Walex> something about latency issues with flash SSDs: https://www.anandtech.com/show/13131/the-silicon-motion-sm2262en-ssd-controller-preview/3
[22:45] <Walex> something about long term steady state testing: https://www.tweaktown.com/articles/6104/ssd-consistency-testing-18-drives-tested-for-real-world-performance/index.html
[22:46] <sarnold> no wonder people loved those optane 900 drives..
[22:47] <Walex> sarnold: well, the write blocks are much smaller and the consistency is much higher, and the rw cycles a lot higher too.
[22:47] <Linkandzelda> Walex: i turned off the write cache because there's a lot of power cuts here. i just turned it on, and gonna do the tests again
[22:47] <Walex> sarnold: flash SSD are essentially RAIDs of 8MiB BIOS flash chips.
[22:48] <Walex> sarnold: every time you do a committed write to a flash SSD you are reflashing a BIOS chip :-)
[22:49] <sarnold> Walex: lol
[22:49] <Walex> sarnold: I know it sounds ridiculous, but it is (almost) the literal thing.
[22:50] <sarnold> thank goodness it's faster than the bios updates I've done :)
[22:51] <Walex> sarnold: yes, but that is largely because BIOS chips are updated using a terrible buis (a serial link IIRC)
[22:51] <Linkandzelda> Walex: with it off the ssd's results WRITE: bw=38.1MiB/s (39.9MB/s), 38.1MiB/s-38.1MiB/s (39.9MB/s-39.9MB/s)
[22:51] <Linkandzelda> on*
[22:52] <Walex> Linkandzelda: that explains a lot :-)
[22:53] <Linkandzelda> Walex: how real is the data loss from enabling the cache? i assume it means that if something is written and the OS thinks its done, its not commited til the cache is flushed
[22:54] <Walex> Linkandzelda: it is not a cache, it is a bffer for writes (even if lots of people confuse read caches with write buffers)... :-)
[22:55] <Walex> Linkandzelda: if buffering is disabled, and the drive lies to the OS, terrible things may happen, but often they don't
[22:55] <Linkandzelda> even the bios manafactuers lol
[22:56] <Walex> Linkandzelda: actually sometimes the buffer and cache overlap: so that just-written stuff can also be read back from the write buffer
[22:56] <Linkandzelda> Walex: if the write buffer is enabled its essentially lying isnt it?
[22:56] <Linkandzelda> how long does it take for such things to be commited out of that cache?
[22:59] <src> I use go on ubuntu 21.04 and it's version 1.16.2 which is out of date, so I was looking for an easy way to track stable releases, and found a snap https://snapcraft.io/go however I am unsure now what happens if I install the snap? will the versions co-exist, will it somehow destroy my go build system
[23:06] <Walex> Linkandzelda: the storage unit buffer may have a timeout or only write out when it gets full, depends on the firmware.
[23:10] <Walex> ah I had specified an odd paramter. With typical parameters, my NVME and SATA 2TB both do 30+30MB/s with mixed rw of 32KiB blocks, committed every 8 blocks.
[23:13] <sarnold> src: it probably all boils down to the PATH that's in use in each process
[23:14] <sarnold> src: if the snap go is in the PATH first in most but not all processes, you might run into problems, yeah; it might be worth uninstalling the deb packaged go if you're going to use the snap packaged go, just so you'll have fewer potential problems to troubleshoot
[23:14] <Walex> src: another small issue is that SNAPs can only access files under '/home/', so if you want to build anywhere else you need a "bind" mount
[23:15] <Walex> my current quick-and-rough 'fio' test: https://paste.ubuntu.com/p/7gZ27GPxbb
[23:16] <Walex> note that the test file size is much larger at 100G than the RAM size.
[23:19] <Walex> pure 32KiB random reads get me roughly 450MB/s on the NVME and 250MB/s on the SATA drive (on the SATA drive 'fio' reports 550MB/s, probably caching).
[23:27] <src> sarnold: Walex: mhm kay thanks, good things to consider
[23:28] <src> given ubuntu's six month release cycle and me not needing the newest version of Go that desperately atm, I'll postpone the issue for now
[23:29] <src> seems easier to at some point install the official go package and use their infrastrcture fully not any os packages
[23:30] <sarnold> it could be; when I was dabbling in rust, it was certainly much easier to use the rustup tooling to get rust and cargo
[23:31] <src> I can imagine, and once you screw up using language specific ways to install stuff with packages in distro repositories etc
[23:31] <src> it all goes to hell from there
[23:31] <sarnold> yeah, pip is a good way to make a real hash of things
[23:38] <Walex> sarnold: actually it is possible to use 'pip' in a safe manner, but it is very tedious.
[23:38] <Walex> sarnold: the key is to always, always, always use '--user'
[23:41] <Walex> the dream: https://xkcd.com/353/ the common reality: https://xkcd.com/1987/
[23:42] <sarnold> Walex: oh, I thought the trick was "always use virtenv"
[23:42] <Walex> sarnold: the key is to always, always, always use '--user' and always always use 'python3.N', and invoke 'pip' as something like 'python3.8 -m pip3 ...' and never user 'virtenv'
[23:43] <sarnold> Walex: this second xkcd rings far more accurate
[23:43] <sarnold> Walex: though it's missing the "stuff that only worked for 14 months a dozen years ago"..
[23:44] <Walex> sarnold: I know the 1987 XKCD strip number by heart as I have to use it sometimes several times a day :-)
[23:44] <sarnold> lol
[23:45] <Walex> sarnold: sometimes with: https://boyter.org/static/books/Cf7eHZ1W4AEeZJA.jpg
[23:46] <sarnold> perfect :)
[23:46] <Walex> https://boyter.org/static/books/CfSQdwUW8AErog1.jpg
[23:46] <Walex> https://boyter.org/static/books/pRDPMx2.jpg