Why Linux?

[1nvest]

I recently bought a raspberry pi 4 (4GB version) and am hugely impressed. Very quick for browsing, etc. I do not have a micro HDMI cable, but found it very straightforward to set up headless and access through ssh and VNC. I have not bought a case or heatsink for the board, as I thought I would see if either were necessary first. I have read a lot about reports of overheating/throttling, but also that some of the issues have been fixed with firmware updates. I tried out the sysbench test used on explainingcomputers.com and found that the CPU temperature plateaued at 76C after about 15 minutes, with no sign of throttling. Room temperature was 20C. I have read that placing the board on its side improves ventilation and this seemed to work, with temperature maxing out at 72C. So for anyone wanting to buy one of these, I would advise holding off on the heatsink until you are sure your use case demands it.

Never had one, but tempted to get one just as a low powered always on headless web/file server. Looks like turning ssh on is as easy as creating a empty file called 'ssh' in the root folder and then accessing your router to find its IP address - but is that open ssh i.e. no userid/password initially being required for ssh access?

[ReformedCharacter]

Never had one, but tempted to get one just as a low powered always on headless web/file server. Looks like turning ssh on is as easy as creating a empty file called 'ssh' in the root folder and then accessing your router to find its IP address - but is that open ssh i.e. no userid/password initially being required for ssh access?

https://itsfoss.com/ssh-into-raspberry/

RC

[hiriskpaul]

Never had one, but tempted to get one just as a low powered always on headless web/file server. Looks like turning ssh on is as easy as creating a empty file called 'ssh' in the root folder and then accessing your router to find its IP address - but is that open ssh i.e. no userid/password initially being required for ssh access?

https://itsfoss.com/ssh-into-raspberry/

RC

I put the empty ssh file in place on the sd card, put it in the pi, plugged in an ethernet cable, booted up, did a ssh raspberrypi, which connected straight in. No need to mess around trying to find the IP address. Had to login with default user and password.

After configuring vnc you could disable ssh again if you don't want it, or configure keys each end if you want so you don't have to keep entering the password.

[1nvest]

I put the empty ssh file in place on the sd card, put it in the pi, plugged in an ethernet cable, booted up, did a ssh raspberrypi, which connected straight in. No need to mess around trying to find the IP address. Had to login with default user and password.

After configuring vnc you could disable ssh again if you don't want it, or configure keys each end if you want so you don't have to keep entering the password.

Thanks. I'm a bit of a cli junky, use it still for calcurse (store my calendar in encrypted form on hashbang), htop, mc, irc (weechat), mutt (email) and sshfs mounting (so I just open a file manager on my laptop and the remote filesystem appears as just another folder that I can drag/drop into/from). I use a liveboot (gui) system on my laptop, so once configured no changes are stored across reboots, stays the exact same (so any nasties would get purged by a simple reboot). Data/files stored separately. As part of that I store all bookmarks in a text file and have tilda (that supports clickable links) terminal slide down/hide via a F1 keypress with its tmux sessions and where one links into hashbang where another tmux is also running. Nice being able to tmux detach and later reattach again whilst everything has continued running in the background.

The laptop I'm using has a slow spin up HDD, was gifted it as Windows was unusable, but as I run with everything in ram (Fatdog linux multisession style) that's not a issue. At least not until it comes to opening files that I store on the HDD. Most of my laptop is for google and libreoffice along with music/video playing, where if it were ever lost or stolen is pretty much a non-issue data/security wise.

[hiriskpaul]

I put the empty ssh file in place on the sd card, put it in the pi, plugged in an ethernet cable, booted up, did a ssh raspberrypi, which connected straight in. No need to mess around trying to find the IP address. Had to login with default user and password.

After configuring vnc you could disable ssh again if you don't want it, or configure keys each end if you want so you don't have to keep entering the password.

Thanks. I'm a bit of a cli junky, use it still for calcurse (store my calendar in encrypted form on hashbang), htop, mc, irc (weechat), mutt (email) and sshfs mounting (so I just open a file manager on my laptop and the remote filesystem appears as just another folder that I can drag/drop into/from). I use a liveboot (gui) system on my laptop, so once configured no changes are stored across reboots, stays the exact same (so any nasties would get purged by a simple reboot). Data/files stored separately. As part of that I store all bookmarks in a text file and have tilda (that supports clickable links) terminal slide down/hide via a F1 keypress with its tmux sessions and where one links into hashbang where another tmux is also running. Nice being able to tmux detach and later reattach again whilst everything has continued running in the background.

The laptop I'm using has a slow spin up HDD, was gifted it as Windows was unusable, but as I run with everything in ram (Fatdog linux multisession style) that's not a issue. At least not until it comes to opening files that I store on the HDD. Most of my laptop is for google and libreoffice along with music/video playing, where if it were ever lost or stolen is pretty much a non-issue data/security wise.

In that case you might as well leave ssh in place, but set up keys at each end. You can then control the rpi from scripts that contain ssh commands to the rpi.

I understand that USB drive booting is now official, so you might prefer the reliability of that over sd cards. 128GB SSDs can be bought for under £20 now. You can also run Docker on the rpi, which you might prefer in order to better contain your Web server. Not as efficient, but doubt you would notice unless running loads of stuff on the rpi.

[ReformedCharacter]

Thanks. I'm a bit of a cli junky, use it still for calcurse (store my calendar in encrypted form on hashbang), htop, mc, irc (weechat), mutt (email) and sshfs mounting (so I just open a file manager on my laptop and the remote filesystem appears as just another folder that I can drag/drop into/from). I use a liveboot (gui) system on my laptop, so once configured no changes are stored across reboots, stays the exact same (so any nasties would get purged by a simple reboot). Data/files stored separately. As part of that I store all bookmarks in a text file and have tilda (that supports clickable links) terminal slide down/hide via a F1 keypress with its tmux sessions and where one links into hashbang where another tmux is also running. Nice being able to tmux detach and later reattach again whilst everything has continued running in the background.

The laptop I'm using has a slow spin up HDD, was gifted it as Windows was unusable, but as I run with everything in ram (Fatdog linux multisession style) that's not a issue. At least not until it comes to opening files that I store on the HDD. Most of my laptop is for google and libreoffice along with music/video playing, where if it were ever lost or stolen is pretty much a non-issue data/security wise.

Very interesting! Just looked up tmux, I think I'll have to look into that.

RC

[1nvest]

If you run a local tmux and remote tmux then you pretty much have to redefine the control keys so that they don't conflict. On hashbang I've set it to be the backtick (below the escape key) - which means that if I want to actually type a backtick I have to press it twice in quick succession. Not used that much as a character other than as execute quotes in shell scripts.

mc is great .. you can even create different user menus in different folders. My main one (when I press F2) looks like ...

So F1 to drop down tilda (that terminal screen), F2 in mc to show the menu and I've set P to poweroff, R to reboot ...etc. Three quick succession keystrokes to initiate (highly configurable) actions.

Good to set separate mc themes on different boxes. Otherwise for instance if I'm running mc on hashbang as well as locally in different tmux panes then you can lose track of where you are.

With tmux some like to zoom/unzoom panes, a form of windows where you can see many on a single screen and flip between them and then full screen zoom/unzoom the window/pane, personally I prefer to just flip between full windows rather than using panes. So in my case backtick 1, 2, 3 or whatever window I want to switch to. For instance whilst viewing irc backtick 2 would switch to email

Hashbang has been continually up for 35 days as per that image, its usually over 100 days before they do reboot - which can be a little annoying as all of the things you'd left running such as irc then have to be restarted/logged into again. But otherwise you can (in my backtick as the control key setup) backtick d to disconnect, and then later reattach again from the same or different devices, even your mobile and carry on from where you'd left.

I have tilda set to open up down to the taskbar, so that's still visible. Or F1 to hide tilda and reveal the usual desktop/icons. So all text and gui on the same single desktop, rather than using ctrl-alt-F2 ... or whatever to flip between different desktop sessions.

htop for process/processor monitoring is another nice cli tool

That 128 processor getting on for 1TB of ram ... isn't my laptop before you ask. Probably will have to wait until 2030 or so before that is the case

[Infrasonic]

https://www.youtube.com/watch?v=5aJ9U5t9oD4

Doesn't open-source just make you feel all warm and fuzzy inside? Today Anthony takes a look at System76's Oryx Pro laptop.

[PitBullCH]

For users that like the Windows look, Zorin OS is a very nice Linix distro; and for Mac lovers, Elementary OS is the usual recommendation.

Otherwise Mint or Ubuntu are solid choices.

[servodude]

The problem is it doesn't let me run 'dpkg --configure -a'

does it let you run:
sudo dpkg --configure -a

it will ask for your password

- sd

[servodude]

The problem is it doesn't let me run 'dpkg --configure -a'

does it let you run:
sudo dpkg --configure -a

it will ask for your password

- sd

It did. And it worked. Thanks!

Great!

Think of it as Super User DO and use it when you need elevated access for a command
Welcome to the sharp tools

-sd

[langley59]

I must say as a relatively new user of Linux (Mint 19.3 Cinnamon) there do seem to be an awful lot of updates (practically every day) versus Windows. I wonder why that is?

[Urbandreamer]

I must say as a relatively new user of Linux (Mint 19.3 Cinnamon) there do seem to be an awful lot of updates (practically every day) versus Windows. I wonder why that is?

It's odd, I don't notice that there are updates available, but then I don't check.

I'm old school, if it works, use it until it's not an issue to try a update. As I understand it Mint has something they call timeshift that makes it easy to restore to a previous change, but as I said, I predate such things.

On the subject of available changes every day, what do you expect with millions of developers? There WILL be something new every day. No need to take up the changes, unlike Windows.

Simply use a LTS (Long term Support) version and update every six months or so. After a few years try the new LTS version.

That is unless you want bleeding edge or are gameing, in which case you probably want a cutting edge or tweaked kernel.

BTW, you do know that you can make your own changes to the kernel/OS, compile and test, then suggest the changes don't you? I don't do so, but you might want to. Even if nobody else likes the changes, you can continue to use them. Try that with Windows!

That fact might just explain why there's something new almost every day.

[langley59]

Thanks for the reply Urbandreamer. I've been diligently installing the updates every day. I wonder if that's higher or lower risk than ignoring them?
PS. I'm not sure what you mean by using a long term support version?

[servodude]

I'm not sure what you mean by using a long term support version?

LTS versions are simply ones for which support ( bug/security patches) are guaranteed/promised for an "extended" period; with mint or ubuntu this is presently 5 years. (https://en.wikipedia.org/wiki/Long-term_support)

The only downside is that they might not get any new whizz bang features (or not as quickly as a bleeding edge version)
- sd

[Urbandreamer]

Thanks for the reply Urbandreamer. I've been diligently installing the updates every day. I wonder if that's higher or lower risk than ignoring them?
PS. I'm not sure what you mean by using a long term support version?

Ok, most things Linux are NO risk. Your word processor and spreadsheet applications, your email or video/music player etc.

That's slightly different from what windows was. Windows has changed things since 7, to significantly improve security. They also introduced the idea of changing things.

Linux has ALWAYS been based on the idea that things change and hopefully improve. However unlike Win 10 you don't need to update if you don't want to. If there is a security issue, rare, but it has happened, then you can just update that part of the OS. Word processors and internet browsers were NEVER part of the OS or allowed to change your system.

Long term support is akin to Win 3.1, 95, NT4, Xp, CE, Vista, 7, 8 10. The people behind it promise to provide support to keep it working for a time.
If you load the updates, it won't change how it works or looks (unlike Win 10). If you want the latest version then you download that.

I suspect that users of Mint 18 LTS are still getting updates. It's Long Time Support!. They don't have to change what they have.

FWIW, I just checked, I have 309 packages to upgrade. Has it caused me a problem? No it hasn't. Shall I upgrade, well now I have checked, why not?
Many of them are stuff I rarely if ever use. Stuff installed with Mint assuming that I might like to do so.

Do you do much with python? It's one of the packages that I just updated.

[JohnB]

Mint updates are much less intrusive than Windows. I get no nagging messages, just type 'apt update; apt upgrade' when I think of it. It runs, rarely asking a question, and all continues as before. No reboots, demands to do things now or random powerups in the dead of night.

And I know few hackers are banging on Linux users door trying to get in.

[GeoffF100]

I am using the current Long Term Support version of Lubuntu (20.04). I get updates every few days. Some are updates to the operating system and others are updates to applications. Applications are not usually updated automatically on Windows. I count that as a plus for Linux. You can look and see what is to be updated before clicking the go ahead button if you feel so inclined. I just accept the updates. I have never had a any problems with updates on Lubuntu, Xubuntu or Ubuntu. The updates happen in the background. They do not take over you machine like they do with Windows. In most cases, a reboot is not required to activate the updates. If the Linux Kernel is updated a reboot is required. That can be avoided in some cases by using Live Patch, but I believe that you have to pay Canonical for that service.

[langley59]

Do you do much with python? It's one of the packages that I just updated.

No not at all, I just use Linux as I did Windows.

[stacker512]

If the Linux Kernel is updated a reboot is required. That can be avoided in some cases by using Live Patch, but I believe that you have to pay Canonical for that service.

My experience is that a kernel update doesn't require you to reboot then and there - you continue to use your existing kernel image and modules until the next reboot. If you require any of the kernel updates (a specific bug fix or a feature) to be active, then sure, you would reboot. Otherwise it's going to be enabled on next boot by the simple fact that when you boot up, the kernel image is loaded up and the relevant modules are loaded.

(edit: you do have kexec, which allows you to load and boot a kernel image from an already running kernel image, but I've never tried it and not sure how many of the popular Linux distros use that feature as part of the kernel update mechanism)