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

«‹›1024 »

Horse Clocks: Dec 14, 2004

Horse Clocks posted:

I decided to check out openSUSE Tumbleweed (again) yesterday. And I must say I'm somewhat impressed with LXQt, Snapper, 1-Click installs and whatever else I'm missing.

But I've yet to get (some) videos to play. Youtube works, but .gifv's (Firefox reports "No decoders required for formats: video/mp4") on Imgur, and .wmv's on disk don't play.

I did a 1-click install of: http://opensuse-guide.org/codecs.php, but that didn't seem to work (and I'm not sure why packman is even needed, all those packages seem to be present in the Tumbleweed repo) so I rolled back those changes with snapper.

I'm somewhat stumped, I could have sworn last time I gave suse a crack it 'just worked'.

Managed to solve my problem.

Ended up adding the packman-essentials repo. Running `zypper dup` (This changed the provider of all the vlc,ffmpeg and gstreamer packages to packman), and then making sure `media.ffmpeg.enabled` was True in `about :config`.

# ? Dec 2, 2016 10:37

Adbot: ADBOT LOVES YOU

# ? Apr 19, 2024 03:05

The_Franz: Aug 8, 2003

Is there a way to do fractional desktop scaling on Gnome with Wayland? I have a 27" 4k monitor and with X11 I can use xrandr to render to a virtual 5k desktop and then downscale it by 0.75 for macOS-style scaling which gives me a working space equivalent to 2560x1440. With Gnome running Wayland only 1x or 2x integer scaling is possible which basically means I either get microscopic text or a 1080p-sized workspace on a 27" monitor.

# ? Dec 2, 2016 16:20

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

Any suggestions for a speedtest thingamabob I can host on a server of my own? Clients would include a few mobile devices as well. Doesn't need to be pretty but should be easy for end users.

I see that speedtest.net has Speedtest Mini you can host on your own site...anything else I should look at?

# ? Dec 2, 2016 21:37

RFC2324: Jun 7, 2012; http 418

Thermopyle posted:

Any suggestions for a speedtest thingamabob I can host on a server of my own? Clients would include a few mobile devices as well. Doesn't need to be pretty but should be easy for end users.

I see that speedtest.net has Speedtest Mini you can host on your own site...anything else I should look at?

Ftp server?

# ? Dec 2, 2016 22:22

Methanar: Sep 26, 2013; by the sex ghost

Thermopyle posted:

Any suggestions for a speedtest thingamabob I can host on a server of my own? Clients would include a few mobile devices as well. Doesn't need to be pretty but should be easy for end users.

I see that speedtest.net has Speedtest Mini you can host on your own site...anything else I should look at?

I would say an iperf3 server. Some clients exist for both Android and iOS as well as Windows. I've never used the windows client but there's probably a way you could write a batch script it to make it easy to run. Apparently someone wants 5 bux for the iOS client though :lol:

e; Ehh, actually you probably don't want to have to install anything to do the test. the speed test mini website thing might be your best bet.

Methanar fucked around with this message at 22:28 on Dec 2, 2016

# ? Dec 2, 2016 22:24

Achmed Jones: Oct 16, 2004

Boris Galerkin posted:

In Vim I have the habit of accidentally hitting shift when I want to use a command like 'tabedit' (ie I'll accidentally type Tabedit about half the time). Is there a way I can just force vim to recognize Tabedit = tabedit and vice versa?

in vimrc

code:

if has("user_commands")
  command! Tabedit tabedit
endif

Alternately, use ; instead of :

Achmed Jones fucked around with this message at 14:45 on Dec 3, 2016

# ? Dec 3, 2016 14:30

apropos man: Sep 5, 2016; You get a hundred and forty one thousand years and you're out in eight!

Today I hooked my headless server up to the TV to have a poke around in BIOS. It's running Ubuntu.

I also booted the machine and had a 5 minute play with a wireless keyboard and mouse plugged in. I usually access it with ssh.

I powered it down and put it back where it usually lives, then noticed that root had sent an email via postfix which was headed "*** SECURITY information for <hostname> ***" which detailed a command I ran whilst playing earlier.

I realised that it was a result of me typing the sudo password in wrongly when sat in front of the tty terminal and I quite like the idea of receiving an email every time someone gets sudo password wrong. The problem is, I can't replicate the alert over ssh. Presumably Ubuntu handles tty auth problems with greater severity than someone dicking about in a pty (over ssh).

Any ideas how I can get the alerts for ssh login (and then in the near future grow tired of the emails and end up turning it back off again)?

# ? Dec 3, 2016 19:06

LochNessMonster: Feb 3, 2005; I need about three fitty

evol262 posted:

Hence usability.

If really all you need is containers, Alpine is better. But the size difference is marginal, and I don't know why you'd ever run Alpine as anything other than a base layer for a container.

That's my use case for Alpine. I run stuff in containers that don't need any additional stuff.

# ? Dec 4, 2016 08:30

telcoM: Mar 21, 2009; Fallen Rib

apropos man posted:

Today I hooked my headless server up to the TV to have a poke around in BIOS. It's running Ubuntu.

I powered it down and put it back where it usually lives, then noticed that root had sent an email via postfix which was headed "*** SECURITY information for <hostname> ***" which detailed a command I ran whilst playing earlier.

I realised that it was a result of me typing the sudo password in wrongly when sat in front of the tty terminal and I quite like the idea of receiving an email every time someone gets sudo password wrong. The problem is, I can't replicate the alert over ssh. Presumably Ubuntu handles tty auth problems with greater severity than someone dicking about in a pty (over ssh).

Any ideas how I can get the alerts for ssh login (and then in the near future grow tired of the emails and end up turning it back off again)?

By default, sudo sends an email alert when you type the password wrong 3 times consecutively using sudo. This can be adjusted by using "Defaults" statements in the sudoers file.

On the other hand, sshd logs authentication failures in /var/log/auth.log. There is also the btmp file, viewable using the "lastb" command.
(If your sshd port is Internet-accessible, you really don't want to be spammed with 100 mail notifications each time some automated port-scanning worm on some other system finds it and probes it for common username/password combinations.)

Having the authentication failure notice sent as email in real-time would require that the PAM module responsible for password authentication (usually pam_unix.so) had the functionality to at least run a command when an authentication failure is detected. Currently, there does not seem to be such functionality.

On the other hand, there are lots of tools that can watch the system logs for you and run commands when some condition is satisfied.
And it isn't too difficult to make one yourself with a bit of scripting. Here's one example that only needs a bit of filtering added to only pass sshd authentication failures:
http://www.johnandcailin.com/blog/john/how-setup-real-time-email-notification-critical-syslog-events

# ? Dec 4, 2016 11:10

apropos man: Sep 5, 2016; You get a hundred and forty one thousand years and you're out in eight!

Cheers. Thanks to your post I tried getting the password wrong more than once and if I input 3 incorrect passwords when running sudo I get the email, even from an ssh connection.

The pam module being allowed to run a mail command is probably enabled because I installed postfix and enabled sending emails through a Gmail account.

It's not directly accessible to the Internet via port scanning, just via a second node.

# ? Dec 5, 2016 06:35

salisbury shake: Dec 27, 2011

I have a 2TB drive. It is formatted like so:

code:

me@host:~$ sudo fdisk -l /dev/sdb
Disk /dev/sdb: 1.8 TiB, 2000398933504 bytes, 3907029167 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: <insert_a_uuid_here>

Device          Start        End    Sectors   Size Type
/dev/sdb1        2048  511662079  511660032   244G Linux filesystem
/dev/sdb2  1953515520 3907028991 1953513472 931.5G Linux filesystem

or it can be thought as

code:

2TB
[~240+ GB HFS+ | 680+ GB Free space | 930+ GB btrfs]

I want to include the free space preceding the btrfs partition in the btrfs filesystem. I've come up with a few options:

1) Format the 680+ GB free space as btrfs, add the empty filesystem to the btrfs filesystem like you would another disk using btrfs-progs. From google, I've read that prior to kernel 3.0, there was an problem where the available/reported disk space did not match the actual available disk space from adding the new drive/partition to a pool. I'm stuck on a 3.19 kernel, not sure if this is still an issue.

2) Open up gparted and let it do its magic to expand the last partition to include the free space preceding it. I worry that I'll have to wait out an operation moves 900+ GB of data and wears my disk drive for no reason. If this somehow will just copy the btrfs metadata to the beginning of the partition, cool, but I doubt it will.

3) Find out what happens when I try to run 'sudo btrfs filesystem resize max' though I assume that will only work if the freespace is after the source btrfs filesystem and the partition is expanded to include the free space.

I hesitate because I've lost data messing around with btrfs years ago and because this disk holds data that might not have a backup. Any tips?

# ? Dec 5, 2016 15:22

some kinda jackal: Feb 25, 2003; �
�

I know there's no real definite answer but how do you guys treat FHS when it comes to database content?

I'm guessing /srv is probably the best, right? I'm trying to train myself to keep things organized and trying to follow FHS. I've been throwing my database stuff in /opt/<dbname>/content/ but since I had to rebuild my lab I've been shoving it in /srv/database/<dbname> . No issue, just wondered how other people handle locating their data and whether anyone else is trying to follow FHS.

# ? Dec 6, 2016 14:05

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

Martytoof posted:

I know there's no real definite answer but how do you guys treat FHS when it comes to database content?

I'm guessing /srv is probably the best, right? I'm trying to train myself to keep things organized and trying to follow FHS. I've been throwing my database stuff in /opt/<dbname>/content/ but since I had to rebuild my lab I've been shoving it in /srv/database/<dbname> . No issue, just wondered how other people handle locating their data and whether anyone else is trying to follow FHS.

I was very anal-retentive about this for years, then I realized that literally no one cares and anyone looking for the missing files is just going to check the config file of the running database service

Don't put it under /tmp is the answer I guess

# ? Dec 6, 2016 18:06

xzzy: Mar 5, 2009

I tend to leave stuff under the path the software defaults to because I don't give a poo poo where stuff lives. If that partition isn't big enough for whatever reason, I'll put a mountpoint there.

# ? Dec 6, 2016 18:09

some kinda jackal: Feb 25, 2003; �
�

Yeah, admittedly it's more just me trying to get into some kind of standard than it is me trying to solve a deficiency in leaving things where they belong.

I guess since I've been rebuilding my lab I've tried to be more consistent about where I can find things on each of my servers.

# ? Dec 6, 2016 22:38

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

I want to know if there is a linux text editor that implements a specific feature in the way that notepad++ does on windows.

In notepad++ I can have multiple tabs open and take quick notes that it automatically caches without the need for me to save to a file at all. Closing the program without saving does not give a warning that I will lose all my data, it just closes, and then brings the same data back up when the program is loaded again.

Also this is different than a simple autosave, because I can have unsaved changes that I'm working on in a file, they remain unsaved, (opening the file in another program will not show my changes), instead notepad++ keeps its own copy of the WIP changes.

Its like a neat sort of staging/commit thing built in that I find incredibly convenient. Also its backup interval is on the order of seconds(default 7s) rather than minutes like many other programs.

So is there something for linux that I can use in the same way? I'm looking for something with a GUI interface similar to gedit ( don't say emacs or vim; do not want)

peepsalot fucked around with this message at 01:29 on Dec 7, 2016

# ? Dec 7, 2016 01:25

Methanar: Sep 26, 2013; by the sex ghost

peepsalot posted:

I want to know if there is a linux program that implements a specific feature in the way that notepad++ does on windows.

In notepad++ I can have multiple tabs open and take quick notes that it automatically caches without the need for me to save to a file at all. Closing the program without saving does not give a warning that I will lose all my data, it just closes, and then brings the same data back up when the program is loaded again.

Also this is different than a simple autosave, because I can have unsaved changes that I'm working on in a file, they remain unsaved, (opening the file in another program will not show my changes), instead notepad++ keeps its own copy of the WIP changes.

Its like a neat sort of staging/commit thing built in that I find incredibly convenient. Also its backup interval is on the order of seconds(default 7s) rather than minutes like many other programs.

So is there something for linux that I can use in the same way? I'm looking for something with a GUI interface similar to gedit ( don't say emacs or vim; do not want)

Sublime text does this and I really like it.

# ? Dec 7, 2016 01:29

taqueso: Mar 8, 2004

I believe Atom does too.

# ? Dec 7, 2016 01:30

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

Methanar posted:

Sublime text does this and I really like it.

I've only used Sublime Text on Windows but it's extremely good.

# ? Dec 7, 2016 01:41

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

Is there any reason I shouldn't be using Continuous Release packages for CentOS 7?

# ? Dec 7, 2016 02:53

everythingWasBees: Jan 9, 2013

So my Arch installation is up and running, I fixed all the weird laptop driver issues, got KDE set up, and everything's fine and dandy; I'm actually using it more than Windows on the same machine. However, sometimes I've noticed that when I'm using Windows and restart to go into Arch, it just launches into a black screen and I have to force shutdown. Has this happened with anybody else dual booting?

# ? Dec 8, 2016 00:02

DrSunshine: Mar 23, 2009; Did I just say that out loud~~?!!!

Hey! I just installed Ubuntu and I'm trying to get some of my old Windows software to work, like Adobe Photoshop, using wine. The problem is that every version of Photoshop I try to install seems to crash. I've tried CS6 and an old cd of Photoshop Elements, but no such luck! It always crashes as soon as the installer starts up.

Another problem I've been having is getting the system to recognize my Wacom Intuos Draw CTL490DW Digital Drawing and Graphics Tablet. I'm led to believe that all the drivers should already be installed, and I can move the pen and such, but it doesn't seem to have any pressure sensitivity.

DrSunshine fucked around with this message at 00:13 on Dec 8, 2016

# ? Dec 8, 2016 00:07

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

DrSunshine posted:

Hey! I just installed Ubuntu and I'm trying to get some of my old Windows software to work, like Adobe Photoshop, using wine. The problem is that every version of Photoshop I try to install seems to crash. I've tried CS6 and an old cd of Photoshop Elements, but no such luck! It always crashes as soon as the installer starts up.

Another problem I've been having is getting the system to recognize my Wacom Intuos Draw CTL490DW Digital Drawing and Graphics Tablet. I'm led to believe that all the drivers should already be installed, and I can move the pen and such, but it doesn't seem to have any pressure sensitivity.

I think you're going to have to use GIMP instead of Photoshop, and you get to use xsetwacom to set the pressure sensitivity of your tablet.

# ? Dec 8, 2016 00:42

Mega Comrade: Apr 22, 2004; Listen buddy, we all got problems!

everythingWasBees posted:

So my Arch installation is up and running, I fixed all the weird laptop driver issues, got KDE set up, and everything's fine and dandy; I'm actually using it more than Windows on the same machine. However, sometimes I've noticed that when I'm using Windows and restart to go into Arch, it just launches into a black screen and I have to force shutdown. Has this happened with anybody else dual booting?

Are you dual booting off the same hard drive? Windows has a habit of not 'releasing' a drive unless you do a cold boot, so likely arch is trying to mount the drive and windows won't let it.

I'd recommend you just getting a 2nd drive for each OS, it saves a lot of aggro.

# ? Dec 8, 2016 01:13

Odette: Mar 19, 2011

everythingWasBees posted:

So my Arch installation is up and running, I fixed all the weird laptop driver issues, got KDE set up, and everything's fine and dandy; I'm actually using it more than Windows on the same machine. However, sometimes I've noticed that when I'm using Windows and restart to go into Arch, it just launches into a black screen and I have to force shutdown. Has this happened with anybody else dual booting?

Do you have 'quiet' as a kernel parameter? I recommend removing that and seeing what pops up on boot.

I had a similar issue with dual-booting Windows, and the most likely culprit was a combination of either one or both of:

1) Windows Fast Startup
2) Dirty unmount of NTFS drive(s)

Every time Windows updated itself, it would turn Fast Startup back on. It's the reason why I no longer dual-boot Windows, and just straight up use Linux for everything now.

EDIT:

Mega Comrade posted:

Are you dual booting off the same hard drive? Windows has a habit of not 'releasing' a drive unless you do a cold boot, so likely arch is trying to mount the drive and windows won't let it.

I'd recommend you just getting a 2nd drive for each OS, it saves a lot of aggro.

Even with a drive per OS, this still occurred to me. Windows is notoriously unreliable anyway.

# ? Dec 8, 2016 02:41

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

Windows was notoriously unreliable.

# ? Dec 8, 2016 05:38

Odette: Mar 19, 2011

evol262 posted:

Windows was notoriously unreliable.

That's your opinion and you're entitled to it, and me mine.

Doesn't change the fact that Windows is still a piece of poo poo.

# ? Dec 8, 2016 06:05

apropos man: Sep 5, 2016; You get a hundred and forty one thousand years and you're out in eight!

In addition to disabling Windows fast startup, I'd also disable hibernation. When you use the Ubuntu installer and if you've got Windows with hibernation on another partition Ubuntu will tell you that a certain partition is currently in use/locked by another OS and ask if you want to force unlock it. Easier just to disable Windows hibernation. I don't really like OS hibernation. I've always preferred knowing that it's completely shut down.

# ? Dec 8, 2016 08:10

VikingofRock: Aug 24, 2008

This is probably a stupid question, but for whatever reason I'm having trouble finding the answer: how do I change the default tab width in the terminal? I'm using zsh as my shell and konsole as my terminal emulator--I'm not sure which one controls the tab width. Googling "zsh tab width" turns up a bunch of info about zsh tab completion, and googling "konsole tab width" turns up a bunch of information about using multiple tabs in konsole.

# ? Dec 8, 2016 08:38

gourdcaptain: Nov 16, 2012

A few days ago, my recently constructed desktop running Arch (personal taste, I'm not recommending it to pretty much anyone else I know) had a problem where after a kernel update and then doing an orderly reboot the UEFI system partition corrupted itself, rendering the system unbootable. had to fix it with an install CD and fsck-ing the partition and reinstalling a package (Intel microcode) that provided a corrupted file on that drive. Any idea what might have caused that? I'm using systemd-boot for a bootloader, so my kernel image and initramfs live on the EFI partition.

Having a FAT32 partition necessary for boot on UEFI makes me irrationally nervous (and I recognize I'm mostly being stupid) since it has basically nothing in the means of protection against crashes and other stuff.

# ? Dec 8, 2016 11:01

Kassad: Nov 12, 2005; It's about time.

Mega Comrade posted:

Are you dual booting off the same hard drive? Windows has a habit of not 'releasing' a drive unless you do a cold boot, so likely arch is trying to mount the drive and windows won't let it.

I'd recommend you just getting a 2nd drive for each OS, it saves a lot of aggro.

Do you know if splitting a SSD between Windows 10 and Ubuntu would make the same thing happen? That is, if I don't do a cold boot to go back to Ubuntu after using Windows, might it stop Ubuntu from booting altogether? I'm about to change my computer's motherboard and CPU, so I'm gonna reinstall everything while I'm at it. I'm only going to use Windows to play some games that don't run on Linux, so I won't share any other hard drive between the two OSes or use hibernation. Getting a second SSD just for Windows isn't an option for me.

# ? Dec 8, 2016 12:03

telcoM: Mar 21, 2009; Fallen Rib

gourdcaptain posted:

A few days ago, my recently constructed desktop running Arch (personal taste, I'm not recommending it to pretty much anyone else I know) had a problem where after a kernel update and then doing an orderly reboot the UEFI system partition corrupted itself, rendering the system unbootable. had to fix it with an install CD and fsck-ing the partition and reinstalling a package (Intel microcode) that provided a corrupted file on that drive. Any idea what might have caused that? I'm using systemd-boot for a bootloader, so my kernel image and initramfs live on the EFI partition.

Having a FAT32 partition necessary for boot on UEFI makes me irrationally nervous (and I recognize I'm mostly being stupid) since it has basically nothing in the means of protection against crashes and other stuff.

Intel microcode updates can nowadays be prepended into the initramfs files in order to enable the new microcode as early in the boot process as possible. There's a specific kernel configuration option for that: CONFIG_MICROCODE_INTEL_EARLY. When this functionality is used, the initramfs file will be a concatenation of several .cpio archives, each of which may or may not be compressed individually. For example, with Debian 8 using the early microcode update feature, I could extract the initramfs like this:

code:

mkdir /tmp/foo
cd /tmp/foo
cat /boot/initrd.img-<kernel-version> | ( cpio -idv; zcat | cpio -idv)

The first cpio will extract the uncompressed early microcode part, and the second zcat | cpio will extract the second, compressed part that contains the main initramfs contents.

If your kernel update did not include that configuration option and the microcode update package just assumed it did (or misdetected it because it looked at the wrong kernel), you might have got an initramfs file that was not suitable for booting with the kernel you wound up actually using.

Linux's FAT32 routines are pretty well tested by now, so I don't know anything specific that might have caused that. I would be inclined to suspect that the FAT32 implementation in the UEFI firmware might even be somewhat less robust than the kernel one, and you might have hit some corner case the firmware could not handle.

# ? Dec 8, 2016 12:38

Mega Comrade: Apr 22, 2004; Listen buddy, we all got problems!

Kassad posted:

Do you know if splitting a SSD between Windows 10 and Ubuntu would make the same thing happen? That is, if I don't do a cold boot to go back to Ubuntu after using Windows, might it stop Ubuntu from booting altogether? I'm about to change my computer's motherboard and CPU, so I'm gonna reinstall everything while I'm at it. I'm only going to use Windows to play some games that don't run on Linux, so I won't share any other hard drive between the two OSes or use hibernation. Getting a second SSD just for Windows isn't an option for me.

Yeah same thing will likely happen. There are suggested ways to solve it such as turning off windows quick boot and the like but none if them worked for me personally.

# ? Dec 8, 2016 12:53

Phosphine: May 30, 2011; WHY, JUDY?! WHY?!
🤰🐰🆚🥪🦊

VikingofRock posted:

This is probably a stupid question, but for whatever reason I'm having trouble finding the answer: how do I change the default tab width in the terminal? I'm using zsh as my shell and konsole as my terminal emulator--I'm not sure which one controls the tab width. Googling "zsh tab width" turns up a bunch of info about zsh tab completion, and googling "konsole tab width" turns up a bunch of information about using multiple tabs in konsole.

Check if you have a "tabs" command. On my ubuntu machine i can do 'tabs -N' to set tab stops to every N characters. if it exists, man tabs was fairly readable.

# ? Dec 8, 2016 13:31

telcoM: Mar 21, 2009; Fallen Rib

Mega Comrade posted:

Yeah same thing will likely happen. There are suggested ways to solve it such as turning off windows quick boot and the like but none if them worked for me personally.

You might want to add a "nofail" mount option in /etc/fstab for any NTFS filesystems.
With that option, if Windows has been hibernated or prepared for fast startup or whatever, the system will just skip the NTFS filesystem(s) instead of stopping the boot process if it finds a NTFS filesystem it should mount but cannot.

I've been using a SSD split between Linux and Windows (first 7 and now 10) for years like this with no trouble.

# ? Dec 8, 2016 14:01

Kassad: Nov 12, 2005; It's about time.

Mega Comrade posted:

Yeah same thing will likely happen. There are suggested ways to solve it such as turning off windows quick boot and the like but none if them worked for me personally.

Figures, thanks.

telcoM posted:

You might want to add a "nofail" mount option in /etc/fstab for any NTFS filesystems.
With that option, if Windows has been hibernated or prepared for fast startup or whatever, the system will just skip the NTFS filesystem(s) instead of stopping the boot process if it finds a NTFS filesystem it should mount but cannot.

I've been using a SSD split between Linux and Windows (first 7 and now 10) for years like this with no trouble.

I'm leaning towards setting the Windows partition not to mount at all (won't be anything on it I'd want to access while in Ubuntu, anyway) but it probably wouldn't hurt to add this option as well.

# ? Dec 8, 2016 14:57

gourdcaptain: Nov 16, 2012

telcoM posted:

Intel microcode updates can nowadays be prepended into the initramfs files in order to enable the new microcode as early in the boot process as possible. There's a specific kernel configuration option for that: CONFIG_MICROCODE_INTEL_EARLY. When this functionality is used, the initramfs file will be a concatenation of several .cpio archives, each of which may or may not be compressed individually. For example, with Debian 8 using the early microcode update feature, I could extract the initramfs like this:
code:
mkdir /tmp/foo
cd /tmp/foo
cat /boot/initrd.img-<kernel-version> | ( cpio -idv; zcat | cpio -idv)
The first cpio will extract the uncompressed early microcode part, and the second zcat | cpio will extract the second, compressed part that contains the main initramfs contents.

If your kernel update did not include that configuration option and the microcode update package just assumed it did (or misdetected it because it looked at the wrong kernel), you might have got an initramfs file that was not suitable for booting with the kernel you wound up actually using.

Linux's FAT32 routines are pretty well tested by now, so I don't know anything specific that might have caused that. I would be inclined to suspect that the FAT32 implementation in the UEFI firmware might even be somewhat less robust than the kernel one, and you might have hit some corner case the firmware could not handle.

It was an error that it couldn't find the files and fsck did find massive errors, so who knows. The intel-ucode file is a separate initramfs on Arch you have loaded before the regular one (been doing that for years) and it's working with early updates in the stock kernel after I fixed the partition. Probably never going to get an answer on this one beyond "freak occurrence". (Might also have to do with being on an NVMe drive, either the BIOS or Linux might be a bit flakier with that.)

# ? Dec 8, 2016 18:09

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

Odette posted:

That's your opinion and you're entitled to it, and me mine.

Doesn't change the fact that Windows is still a piece of poo poo.

lol, am I on Slashdot in 2002

# ? Dec 8, 2016 18:13

Harik: Sep 9, 2001; From the hard streets of Moscow
First dog to touch the stars; Plaster Town Cop

Thermopyle posted:

lol, am I on Slashdot in 2002

I didn't see anyone call it "Micro$haft windurrrs" here so I think you're giving /. too much credit.

I'm running across ridiculous IO latency that I can't track down. 2+ second stalls on SSD. Very visible in things like tab completion - it just hangs for a while when it should be basically instant. BTRFS on 4.8.9. It's not consistent, even if I drop the caches and try to reproduce.

I'd say "SSD is failing" but it SMARTs fine and random access benchmarks return the same performance it's always had. Are there any analyzers for blktrace so I can leave a log running and try to run through it after I see a stall?

I'm looking for something causing an IO storm that would explain what I'm seeing.

840 EVO 250GB

code:

ID# ATTRIBUTE_NAME          FLAG     VALUE WORST THRESH TYPE      UPDATED  WHEN_FAILED RAW_VALUE
  5 Reallocated_Sector_Ct   0x0033   100   100   010    Pre-fail  Always       -       0
  9 Power_On_Hours          0x0032   093   093   000    Old_age   Always       -       31448
 12 Power_Cycle_Count       0x0032   099   099   000    Old_age   Always       -       56
177 Wear_Leveling_Count     0x0013   082   082   000    Pre-fail  Always       -       213
179 Used_Rsvd_Blk_Cnt_Tot   0x0013   100   100   010    Pre-fail  Always       -       0
181 Program_Fail_Cnt_Total  0x0032   100   100   010    Old_age   Always       -       0
182 Erase_Fail_Count_Total  0x0032   100   100   010    Old_age   Always       -       0
183 Runtime_Bad_Block       0x0013   100   100   010    Pre-fail  Always       -       0
187 Uncorrectable_Error_Cnt 0x0032   100   100   000    Old_age   Always       -       0
190 Airflow_Temperature_Cel 0x0032   073   054   000    Old_age   Always       -       27
195 ECC_Error_Rate          0x001a   200   200   000    Old_age   Always       -       0
199 CRC_Error_Count         0x003e   099   099   000    Old_age   Always       -       3
235 POR_Recovery_Count      0x0012   099   099   000    Old_age   Always       -       42
241 Total_LBAs_Written      0x0032   099   099   000    Old_age   Always       -       33558280739

The write load seems high (17TB), which is why I'm thinking something's rogue.

# ? Dec 8, 2016 23:43

Adbot: ADBOT LOVES YOU

# ? Apr 19, 2024 03:05

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

Have you thought about sar or iotop?

# ? Dec 8, 2016 23:45

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

«‹›1024 »