The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

«‹›1025 »

taqueso: Mar 8, 2004

ToxicFrog posted:

Some kind of driver wackiness causing windows to use hardware compositing and linux not? I'm not seeing the behaviour you describe on any of my machines.

Sure sounds like it. kujeger, what kind of framerate and CPU use do you get when you run glxgears?

# ? Aug 25, 2016 15:46

Adbot: ADBOT LOVES YOU

# ? Apr 26, 2024 00:16

PBCrunch: Jun 17, 2002; Lawrence Phillips Always #1 to Me

I just installed Ubuntu Gnome on a Dell Latitude D820 (C2D T7200 2GHz, Intel 945 graphics, 4GB RAM, SSD). Gnome ran absolutely terrible on the machine; letters would appear seconds after typing in a terminal window. I installed vanilla Ubuntu and it was infinitely more responsive. It performed like a regular old computer with an SSD.

What is the hang up here? I thought Unity was pretty much the most resource intensive window manager available on Linux. The machine has a 1920x1280 (or so) screen, so I'm guessing that is part of the problem.

# ? Aug 25, 2016 16:46

kujeger: Feb 19, 2004; OH YES HA HA

taqueso posted:

Sure sounds like it. kujeger, what kind of framerate and CPU use do you get when you run glxgears?

I have this on an intel nuc, an intel laptop, and an nvidia desktop using the proprietary drivers. They're all using hardware compositioning. Pretty bog standard Fedora24 with gnome3 (well, the desktop has a little more extra stuff tweaked).
glxgears only gets 60hz since it's v-sync locked

Wild speculation, but it's possible the vsync implementation has something to do with it. I'll give disabling it a shot later.

edit: does not happen in xfce, with or without vsync. Still to try gnome3 without vsync.

kujeger fucked around with this message at 17:27 on Aug 25, 2016

# ? Aug 25, 2016 17:09

evol262: Nov 30, 2010; #!/usr/bin/perl

PBCrunch posted:

Intel 945 graphics. What is the hang up here?

Found the problem.

taqueso posted:

Sure sounds like it. kujeger, what kind of framerate and CPU use do you get when you run glxgears?

glxgears hasn't been a very good benchmark for this stuff for years, unfortunately. glxinfo is still useful, though.

kujeger posted:

I have this on an intel nuc, an intel laptop, and an nvidia desktop using the proprietary drivers. They're all using hardware compositioning. Pretty bog standard Fedora24 with gnome3 (well, the desktop has a little more extra stuff tweaked).
glxgears only gets 60hz since it's v-sync locked

Wild speculation, but it's possible the vsync implementation has something to do with it. I'll give disabling it a shot later.

edit: does not happen in xfce, with or without vsync. Still to try gnome3 without vsync.

I can't explain why, but I feel like this is actually a feature somehow. I'm almost positive I've seen a developer talking about this somewhere...

# ? Aug 25, 2016 18:45

DeaconBlues: Nov 9, 2011

Vulture Culture posted:

Xorg wasn't "all those years ago"

That was my mistaken belief that the X window system and X.Org were the same thing, and both originated in the 1980's. I just checked the X.Org page on Wikipedia.

Every day's a school day with Linux. �\_(ツ)_/�

# ? Aug 25, 2016 21:12

evol262: Nov 30, 2010; #!/usr/bin/perl

DeaconBlues posted:

That was my mistaken belief that the X window system and X.Org were the same thing, and both originated in the 1980's. I just checked the X.Org page on Wikipedia.

Every day's a school day with Linux. �\_(ツ)_/�

How detailed is Wikipedia's entry about the fork slapfight?

# ? Aug 25, 2016 22:23

DeaconBlues: Nov 9, 2011

A quick scan reveals that the X Window System page gives more detail than the X.Org Server page.

Most of the bunfighting seems to have happened around the early 2000's, before things got less argumentative and progress was resumed. That's my quick take on the Wiki, anyway.

# ? Aug 25, 2016 22:51

Suspicious Dish: Sep 24, 2011; 2020 is the year of linux on the desktop, bro; Fun Shoe

the reason your mouse becomes detached from the window is that the mouse cursor movement happens immediately inside a signal handler in the x server, and the window movement has to go through a full gpu pass on modern devices, and triple buffering is enabled by default on intel gpus, adding extra latency (because the guy who makes the drivers is a guy nobody likes)

# ? Aug 26, 2016 05:46

some kinda jackal: Feb 25, 2003; �
�

Want to set up a personal wordpress, debating between installing wordpress through epel on CentOS7 or getting the sources. I know this is likely a personal preference, but I don't want to have to maintain the software too closely.

I know EPEL has no SLA or really any guarantees, but in everyone's experience is it generally safe enough to throw into "production" and let yum-cron worry about keeping wordpress up to date? I guess I really just want to avoid a situation where I have to start worrying about whether my wordpress instance is lagging horribly behind and start checking wordpress releases every day/week/whatever by hand.

Keeping in mind that this is for a personal blog on a server sitting in a DMZ at my house I'm not terribly worried about the implications of wordpress vulnerabilities, I just don't want to wake up in a year and realize that epel is like 20 revisions back or something.

# ? Aug 26, 2016 15:10

fuf: Sep 12, 2004; haha

Wordpress is pretty good at updating itself these days, just put this in wp-config.php:

code:

define('WP_AUTO_UPDATE_CORE', true);

I don't know anything about EPEL so maybe it can do clever stuff but usually updating Wordpress requires admin user credentials (either to login directly or to use tools like wp-cli or mainwp) because the site goes into maintenance mode and makes changes to its DB etc.

# ? Aug 26, 2016 16:48

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

evol262 posted:

How detailed is Wikipedia's entry about the fork slapfight?

all it needs to say is "Keith Packard owned a bunch of scrubs"

# ? Aug 26, 2016 17:36

ExcessBLarg!: Sep 1, 2001

DeaconBlues posted:

That was my mistaken belief that the X window system and X.Org were the same thing, and both originated in the 1980's.

They kind of are. Honestly, if the "X Window System" is any one thing, it's really a protocol that exists in multiple client and server implementations. It happens to be that the reference implementation of the X protocols was developed and released under a free license, so in practice the various implementations of "X" derive at least in part from the reference implementation. These days X.Org is the canonical and most popular implementation of X, and is descended from (in terms of ownership, principle developers, and code) the early reference implementation. Even before the modern X.Org name, the X Consortium-maintained reference implementation was hosted at https://www.x.org since the 90s.

The actual history is a bit convolved. The short version is that X started as a project at MIT, but grew into something that was expected to have significant commercial value, so ownership of the project was spun off into the X Consortium, which was charged with maintaining the protocol and reference implementation while a bunch of Unix workstation vendors would sell hardware-compatible versions for lots of money. The reason for the X Consortium's continued involvement was to make sure that the implementations sold by vendors would remain cross-compatible. During this period, https://www.x.org existed as a website where (I recall) you could download the reference source (e.g., X11R6), but it wasn't really useful to end users.

Meanwhile, XFree86 emerged as an open-source port of the X reference implementation to Linux/Intel hardware. Much as commercial Unix workstations died and were replaced by Linux on Intel (if not Windows and Mac), XFree86 came to be the most popular, and thus, most relevant implementation of X. Things probably would've stayed that way, except that a spat within XFree86 caused core developers to leave, start the X.Org foundation to modernize and continue development of the original reference implementation.

If you read the story there closely, it's clear that a few core developers (e.g., Jim Gettys and Keith Packard) who had been working X since the 80s are the ones who drove the real direction of the project, and so wherever they go (X11 Consortium, XFree86, X.Org, etc.) is where development and popularity follows. The rest of the fluff is around various organizations and vendors that took "ownership" of the project (although the code was always free) in expectation of profit, but in the end, it wasn't really ever all that profitable.

ExcessBLarg! fucked around with this message at 19:19 on Aug 26, 2016

# ? Aug 26, 2016 19:15

jre: Sep 2, 2011; To the cloud ?

fuf posted:

Wordpress is pretty good at updating itself these days, just put this in wp-config.php:
code:
define('WP_AUTO_UPDATE_CORE', true);
I don't know anything about EPEL so maybe it can do clever stuff but usually updating Wordpress requires admin user credentials (either to login directly or to use tools like wp-cli or mainwp) because the site goes into maintenance mode and makes changes to its DB etc.

The advantage of doing it with yum instead of wp auto update is that you don't need to give the web server write permissions to the wordpress directory

# ? Aug 26, 2016 20:47

DeaconBlues: Nov 9, 2011

ExcessBLarg! posted:

...loads of interesting stuff...

Great write-up. I had gone off computers from about 1993, to pursue going out a lot/being hedonistic/working abroad, returning to casual use a decade later. I don't think I even thought about owning one for much of about a decade of my life. Fancy that!

# ? Aug 26, 2016 21:25

YouTuber: Jul 31, 2004; by FactsAreUseless

DeaconBlues posted:

Great write-up. I had gone off computers from about 1993, to pursue going out a lot/being hedonistic/working abroad, returning to casual use a decade later. I don't think I even thought about owning one for much of about a decade of my life. Fancy that!

I used computers the entire time you didn't use them. You didn't miss much.

# ? Aug 26, 2016 23:18

EmmyOk: Aug 11, 2013

Is there a way to make the launcher transparent in Ubuntu 16? I installed Unity Tweak Tool and CompizConfig Settings MAnager and both have an option for launcher transparency but adjusting the value/slider makes no difference. In CCSM I adjusted maximised panel opacity and it went clear so I'm not sure why the launcher transparency isn't changing.

# ? Aug 27, 2016 13:16

anthonypants: May 6, 2007; by Nyc_Tattoo; Dinosaur Gum

I don't suppose there's any way to keep capture groups persistent through multiple sed -e groups? Here's what I've got:

code:

$ cat testmail
To: <recipient@domain.com>
From: <baduser@domain.local>
Subject: this is an email
Message-ID: <this is a message id>
Date: date
$cat testmail | sed -e 's/^From\: <\(.*\)>/From: <gooduser@domain.com>\nReply-To: <noreply@domain.com>/g' -e 's/^Subject:\(.*\)/Subject: \[\1\]\2/g'
sed: -e expression #2, char 37: invalid reference \2 on `s' command's RHS
$ cat testmail | sed -e 's/^From\:.*<\(.*\)>/From: <gooduser@domain.com>\nReply-To: <noreply@domain.com>/g' -e 's/^Subject:\(.*\)/Subject: \[\1\]/g'
To: <recipient@domain.com>
From: <gooduser@domain.com>
Reply-To: <noreply@domain.com>
Subject: [ this is an email]
Message-ID: <this is a message id>
Date: date
$

I don't know if this is a limitation or if there's some other sed command I need to use.

e: I can get that email by itself with sed -n 's/^From:.*<\(.*\)>/\1/p' ...so now I just need a way to get that into the subject line

anthonypants fucked around with this message at 21:16 on Aug 27, 2016

# ? Aug 27, 2016 19:12

ToxicFrog: Apr 26, 2008

anthonypants posted:

I don't suppose there's any way to keep capture groups persistent through multiple sed -e groups? Here's what I've got:

I don't know if this is a limitation or if there's some other sed command I need to use.

e: I can get that email by itself with sed -n 's/^From:.*<\(.*\)>/\1/p' ...so now I just need a way to get that into the subject line

Sed is not a good fit for this. It's very oriented around individual lines and has limited mechanisms for carrying state across lines (basically, appending to either the hold space or the pattern space). I'm pretty sure it can be done in sed but you'd be better served writing a few lines of python/lua/awk/perl.

And while doing so, consider how you will handle emails with multiple subject and/or from lines, because I guarantee you someone is sending emails that look like that.

# ? Aug 27, 2016 21:55

evol262: Nov 30, 2010; #!/usr/bin/perl

Famously,

jwz posted:

Some people, when confronted with a problem, think "I know, I'll use regular expressions." Now they have two problems.

Regular expressions are gonna be bad at this due to the amount of edge cases. You can do it in Python (re.MULTILINE), perl/php/javascript (/x, iirc), but you're much better off using a standard library which does nothing but email.

For python:

code:

import email
e = email.message_from_string(open("email.txt").readlines())
# do some stuff...

# ? Aug 27, 2016 22:47

VikingofRock: Aug 24, 2008

So my virtual console started looking like this recently:

Any idea what might have caused this, and how to fix it? I'm on Arch Linux. The virtual console looks fine during boot, and only starts to look ugly after the graphical stuff boots up. I've recently been switching desktop environments a lot, so maybe that is related? Although it's not really clear to me how that would affect the TTY, since I would have thought that was separate from the graphical environment.

# ? Aug 28, 2016 00:09

evol262: Nov 30, 2010; #!/usr/bin/perl

Can you post lsmod and /proc/cmdline?

Looks like something (kernel update) broke fbdev or kms

# ? Aug 28, 2016 00:24

peepsalot: Apr 24, 2007; ��PEEP THIS...
��BITCH!

VikingofRock posted:

So my virtual console started looking like this recently:

Any idea what might have caused this, and how to fix it? I'm on Arch Linux. The virtual console looks fine during boot, and only starts to look ugly after the graphical stuff boots up. I've recently been switching desktop environments a lot, so maybe that is related? Although it's not really clear to me how that would affect the TTY, since I would have thought that was separate from the graphical environment.

Does the command "reset" help?

# ? Aug 28, 2016 00:31

VikingofRock: Aug 24, 2008

evol262 posted:

Can you post lsmod and /proc/cmdline?

Looks like something (kernel update) broke fbdev or kms

lsmod: http://pastebin.com/YTgBXdRV

/proc/cmdline:
BOOT_IMAGE=/boot/vmlinuz-linux root=/dev/sdb4

peepsalot posted:

Does the command "reset" help?

Unfortunately, no.

# ? Aug 28, 2016 00:42

evol262: Nov 30, 2010; #!/usr/bin/perl

How about nomodeset?

# ? Aug 28, 2016 01:35

VikingofRock: Aug 24, 2008

evol262 posted:

How about nomodeset?

I tried booting with nomodeset set and it didn't seem to change anything. I'm using the proprietary nvidia drivers, in case that matters.

# ? Aug 28, 2016 07:54

gourdcaptain: Nov 16, 2012

VikingofRock posted:

I tried booting with nomodeset set and it didn't seem to change anything. I'm using the proprietary nvidia drivers, in case that matters.

The proprietary NVIDIA drivers don't enable modesetting by default currently, if I'm up to date on that. (My one system with an NVIDIA card currently is an Optimus setup, so I can't check this stuff personally.) Turning on modesetting with "nvidia-drm.modeset=1" might get an interesting result, although I can't say one way or the other.

# ? Aug 28, 2016 18:07

VikingofRock: Aug 24, 2008

gourdcaptain posted:

The proprietary NVIDIA drivers don't enable modesetting by default currently, if I'm up to date on that. (My one system with an NVIDIA card currently is an Optimus setup, so I can't check this stuff personally.) Turning on modesetting with "nvidia-drm.modeset=1" might get an interesting result, although I can't say one way or the other.

This does something! It breaks it part way through the boot screen, as opposed to after the graphical stuff has started up. The boot text was scrolling a little too fast to make out exactly when it broke, but I think this is probably a good hint as to what's going on.

edit: Is there a way to have stuff boot one-item-at-a-time, so that I can see exactly when the squares appear?

edit2: Rebooted a few more times. I believe it's after "Starting Login service...", which would make sense with what I've seen so far.

VikingofRock fucked around with this message at 19:31 on Aug 28, 2016

# ? Aug 28, 2016 19:21

telcoM: Mar 21, 2009; Fallen Rib

VikingofRock posted:

This does something! It breaks it part way through the boot screen, as opposed to after the graphical stuff has started up. The boot text was scrolling a little too fast to make out exactly when it broke, but I think this is probably a good hint as to what's going on.

edit: Is there a way to have stuff boot one-item-at-a-time, so that I can see exactly when the squares appear?

edit2: Rebooted a few more times. I believe it's after "Starting Login service...", which would make sense with what I've seen so far.

The picture looks like something might be corrupting at least parts of the text console display font, and possibly some parts of the text console display memory (the large solid squares). This is probably a bug somewhere.

It seems that without modesetting, the corruption happens when the graphical stuff starts up, and with modesetting, as soon as the nvidia-drm module is activated.

Perhaps reloading the console font after the nvidia driver has taken control of the display might serve as a work-around? If the text console font corruption is just an one-time issue when the graphics driver is initialized, reloading the console font and then clearing the text console screen might be helpful.

I have no experience with Arch, but ArchWiki has useful information:
https://wiki.archlinux.org/index.php/fonts#Console_fonts

Basically, look into /etc/vconsole.conf to see which console font is currently in use. If nothing is specified in there (or everything is commented out), check /usr/share/kbd/consolefonts/ directory to see which fonts are available, and pick one.

After the system has booted up and the corruption has appeared, run

code:

# setfont <fontname>
# clear

to (re)load the console font and clear the screen.

Of course, replace <fontname> with the name of the virtual console font.

# ? Aug 28, 2016 20:51

taqueso: Mar 8, 2004

VikingofRock posted:

edit: Is there a way to have stuff boot one-item-at-a-time, so that I can see exactly when the squares appear?

You can use the linux kernel parameter "boot_delay=xxx" to delay for xxx mS after each line that is printed. However, be aware that it will delay for some lines that aren't printed, so you have to give it some time while it seems dead. More convenient, if you have the HW for it, is to use a serial console and view the output on a different PC.

# ? Aug 28, 2016 20:59

VikingofRock: Aug 24, 2008

telcoM posted:

The picture looks like something might be corrupting at least parts of the text console display font, and possibly some parts of the text console display memory (the large solid squares). This is probably a bug somewhere.

It seems that without modesetting, the corruption happens when the graphical stuff starts up, and with modesetting, as soon as the nvidia-drm module is activated.

Perhaps reloading the console font after the nvidia driver has taken control of the display might serve as a work-around? If the text console font corruption is just an one-time issue when the graphics driver is initialized, reloading the console font and then clearing the text console screen might be helpful.

I have no experience with Arch, but ArchWiki has useful information:
https://wiki.archlinux.org/index.php/fonts#Console_fonts

Basically, look into /etc/vconsole.conf to see which console font is currently in use. If nothing is specified in there (or everything is commented out), check /usr/share/kbd/consolefonts/ directory to see which fonts are available, and pick one.

After the system has booted up and the corruption has appeared, run
code:
# setfont <fontname>
# clear
to (re)load the console font and clear the screen.

Of course, replace <fontname> with the name of the virtual console font.

Typing setfont at the virtual console worked! It cleared it right up. I'm still curious what is causing this in the first place, and whether I have a misconfigured setting somewhere or whether it is a bug upstream, but this is a nice workaround for now. Really I don't spend very much time on my virtual console anyways, but it's always nice to have it working in case something goes wrong with the desktop environment or whatever and I need to use the virtual console to fix it. Thank you so much everyone for your help.

# ? Aug 29, 2016 01:08

DeaconBlues: Nov 9, 2011

Have there been any security issues with running programs which depend on mono, the Windows on Linux library?

I'm running emby on my home server, which installs a bunch of mono stuff. It's the same box I use for external ssh (no passwords, though), so I care about the security of the box.

# ? Aug 31, 2016 15:10

Thermopyle: Jul 1, 2003; ...the stupid are cocksure while the intelligent are full of doubt. �Bertrand Russell

I've got Ubuntu 14.04 desktop running in a VM. I was doing some browsing with Chrome in the VM and suddenly everything froze. Nothing is responsive in Chrome or anywhere else, can't interact with the GUI in any way. I CAN switch terminals via Ctrl-Alt-F8-F12, but all I get is a blinking cursor. I can also switch back to the GUI but it's still frozen. The host system shows VMWare using 2-8 percent of CPU with a few KB/sec of disk activity.

Since I had some unsaved work going on in this VM is there anything I can do to recover this?

edit: If I suspend the machine via VMWare and restore it, I get the notification up by the clock that the network is back up, so the system is doing something. The foreground window (Chrome) has the grayed out look not-responding applications get in Ubuntu.

Thermopyle fucked around with this message at 00:05 on Sep 1, 2016

# ? Aug 31, 2016 23:34

Mao Zedong Thot: Oct 16, 2008

Thermopyle posted:

I've got Ubuntu 14.04 desktop running in a VM. I was doing some browsing with Chrome in the VM and suddenly everything froze. Nothing is responsive in Chrome or anywhere else, can't interact with the GUI in any way. I CAN switch terminals via Ctrl-Alt-F8-F12, but all I get is a blinking cursor. I can also switch back to the GUI but it's still frozen. The host system shows VMWare using 2-8 percent of CPU with a few KB/sec of disk activity.

Since I had some unsaved work going on in this VM is there anything I can do to recover this?

edit: If I suspend the machine via VMWare and restore it, I get the notification up by the clock that the network is back up, so the system is doing something. The foreground window (Chrome) has the grayed out look not-responding applications get in Ubuntu.

Sounds like Chrome is frozen? If your unsaved work is in the browser, then yeah, probably not much you can do. Otherwise `kill -9` chrome and keep on keepin on.

# ? Sep 1, 2016 02:37

Thermopyle: Jul 1, 2003; ...the stupid are cocksure while the intelligent are full of doubt. �Bertrand Russell

so loving future posted:

Sounds like Chrome is frozen? If your unsaved work is in the browser, then yeah, probably not much you can do. Otherwise `kill -9` chrome and keep on keepin on.

Like I said, nothing anywhere is responsive. I can't open any other programs, I can't switch to the editor I have unsaved work in, I can't click anything on the launcher. If I switch to another tty all I get is a blinking cursor.

# ? Sep 1, 2016 03:15

Odette: Mar 19, 2011

Can you SSH into the VM?

# ? Sep 1, 2016 03:50

Mr Shiny Pants: Nov 12, 2012

DeaconBlues posted:

Have there been any security issues with running programs which depend on mono, the Windows on Linux library?

I'm running emby on my home server, which installs a bunch of mono stuff. It's the same box I use for external ssh (no passwords, though), so I care about the security of the box.

Are all the ports available on the Internet or just SSH?

# ? Sep 1, 2016 05:39

telcoM: Mar 21, 2009; Fallen Rib

Thermopyle posted:

I've got Ubuntu 14.04 desktop running in a VM. I was doing some browsing with Chrome in the VM and suddenly everything froze. Nothing is responsive in Chrome or anywhere else, can't interact with the GUI in any way. I CAN switch terminals via Ctrl-Alt-F8-F12, but all I get is a blinking cursor. I can also switch back to the GUI but it's still frozen. The host system shows VMWare using 2-8 percent of CPU with a few KB/sec of disk activity.

Often the virtual consoles F8 .. F12 are left unused by default. Try the Ctrl-Alt-F1 .. F6 range instead: those are typically set up with a login prompt.

Debian has the "console-log" package, which will automatically start a "less" process to view /var/log/syslog on tty9, maybe Ubuntu has it too?

Then you could just press Ctrl-Alt-F9 to try and view the latest system log entries at least.

# ? Sep 1, 2016 08:42

Thermopyle: Jul 1, 2003; ...the stupid are cocksure while the intelligent are full of doubt. �Bertrand Russell

I got fed up with it and just shut it down. I'll just recreate the work.

Thanks for the info though.

# ? Sep 1, 2016 16:04

DeaconBlues: Nov 9, 2011

Mr Shiny Pants posted:

Are all the ports available on the Internet or just SSH?

Just ssh is forwarded out by the router.

# ? Sep 2, 2016 10:10

Adbot: ADBOT LOVES YOU

# ? Apr 26, 2024 00:16

fuf: Sep 12, 2004; haha

I have a server that backs up some files to a Dropbox directory.

I am kinda done with Dropbox (the selective sync is too hard to manage once you have more than a couple of devices) and I'm wondering what I can replace it with.

I looked at Google Drive but their linux client seems kinda unsupported / cobbled together. The Dropbox linux client was kinda unreliable too (I'd have to restart it a lot) so I want someting that actually works well with linux (command line).

I basically use it for two things:
1) Extra reassurance knowing that important configs / backups are synced to the cloud somewhere
2) An easy way to transfer files between servers / my local machine

Any suggestions? Happy to pay.

# ? Sep 2, 2016 13:01

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > The Linux Questions Thread: a bunch of pitfalls, but technically it's possible

«‹›1025 »