[00:02] <matsaman> bulk convert should be about the same as any loop
[00:03] <matsaman> for i in *.mkv; do ffmpeg -i "$i" -c:v libx265 "$i".mp4; done, etc.
[00:03] <matsaman> you might want ffmpeg -i foo bar < /dev/null in certain circumstances
[00:04] <matsaman> and there are GUI frontends aplenty
[00:05] <frad> -x265-params lossless=1, I think I should use this one. THis way I won't lose any quality, but the resulting file will be smaller (x264 to z265), correct?
[00:06] <matsaman> much smaller
[00:08] <frad> so: ffmpeg -i foo.mkv -x265-params lossless=1 bar.mkv?
[00:08] <matsaman> make sure the hardware you plan on playing it back via can handle it before you transcode a large collection, though
[00:08] <frad> testing with 1
[00:08] <matsaman> larger resolution h.265 in particular can be fairly demanding to play back
[00:09] <matsaman> frad: you'll probably still need -c:v libx265
[00:09] <matsaman> if you throw in -t 00:00:30, that'll do it for 30 seconds only, for testing
[00:09] <frad> ffmpeg -i foo.mkv -x265-params lossless=1 -c:v libx265 -t 00:00:30 bar.mkv?
[00:10] <matsaman> looks good; it's possible the -c:v libx265 may have to come before -x265-params
[00:10] <matsaman> it'll let you know if so
[00:22] <frad> matsaman, how dows compressng to x265 affect grainy films?
[00:22] <frad> from the 80's and 90s
[00:23] <frad> does compressing*
[00:23] <matsaman> if you're doing it losslessly, it'll merely be smaller in file size
[00:23] <frad> I mean I wouldn't mind losing some grainy-ness
[00:23] <matsaman> if you're doing it losslessly, probably nothing much
[00:23] <matsaman> whoops
[00:23] <matsaman> if you're doing it lossily*, probably nothing much
[00:24] <matsaman> when you're allowing loss, what you need to watch out for is darks and lights and high contrast
[00:24] <matsaman> animated things are usually higher file sizes to maintain quality, because there are so many large blocks of perceptible color
[00:24] <frad> allowing loss = not running '-x265-params lossless=1'
[00:24] <matsaman> the defaults will be very good for most things, though
[00:24] <matsaman> frad: right
[00:25] <frad> defaults = running your first command, ffmpeg -i foo.mkv -c:v libx265 bar.mp4?
[00:26] <matsaman> yeah
[00:26] <matsaman> that will be technically lossy
[00:26] <matsaman> although perceptibly, as a human, most of the time, you will not notice loss
[00:26] <matsaman> (except in file size, for h.264 -> h.265)
[00:28] <frad> I have a small screen 15 inches, but I want to keep as much quality as possible, for when I buy a 27'' monitor. If this is the plan, would it make sense to just use '-x265-params lossless=1'?
[00:28] <frad> will I then notice loss?
[00:28] <matsaman> I don't think you'll notice loss either way
[00:28] <matsaman> if you're trying to make a collection that will last "forever", it makes sense to go lossless
[00:29] <matsaman> but only if the filesize remains small enough to be financially viable
[00:29] <matsaman> since you're going from h.264 to h.265 that probably isn't an issue
[00:29] <matsaman> because you can reduce size by at least half most of the time
[00:40] <frad> matsaman, the command works. This is a 2 hour movie. How long will it last? if it says 'fps=3', 3 frames per second?
[00:42] <matsaman> frad: there should be a timestamp readout in the output, hours, minutes, seconds
[00:42] <matsaman> if you've got regular consumer grade hardware, it will be a decent while
[00:42] <matsaman> try using -t 00:01:00 and you'll get a time for one minute
[00:44] <frad> strange: I ran ffmpeg -i foo.mkv -c:v libx265 -x265-params lossless=1 -t 00:00:30 bar.mkv but the output is 15 seconds long...
[00:45] <matsaman> that is strange =)
[00:45] <matsaman> maybe :31
[00:46] <frad> encoded 411 frames in 237.54s (1.73 fps), 83098.82 kb/s, Avg QP:4.00
[00:46] <frad> so a second is made of 24 frames, correct?
[00:47] <matsaman> frad: hrmm?
[00:47] <frad> compressing this movie will take me exactly 24 hours at 2 fps...
[00:47] <frad> matsaman, 'encoded 411 frames' 24 frames are a second, correct?
[00:48] <frad> a temporal second is made of 24 frames
[00:48] <matsaman> I would take the seconds of video you got
[00:48] <matsaman> and the seconds of time it took to produce
[00:49] <frad> 4 minutes to create 17 seconds of movie
[00:49] <frad> bejesus
[00:49] <matsaman> it could very well take a long time
[00:50] <matsaman> it'd be interesting to compare not using the lossless option, for time
[00:51] <frad> 29 hours
[00:51] <frad> now I wonder what kind of equipment do pirates use to compress and offer so many movies
[04:00] <robertparkerx> what is default location for mysql log and system log
[04:02] <matsaman> robertparkerx: /var/log/ is the normal place for logs
[04:03] <robertparkerx> ty
[05:31] <today911> hey! any way to list the touchpad module in ubuntu 14.04 ?
[05:31] <today911> (I was using Ubuntu 20.04 which's touchpad is ok, but I find the ubuntu 14.04 driver or module works better?)
[05:32] <today911> (same goes with touchscreen module ?)
[05:45] <cpatrick08> I was on the Cd Image server and noticed their is a Ubuntu-Canary iso. I was wondering what the differences are between it and the regular Ubuntu Daily ISO.
[06:44] <slidercrank> hello. how do I install kernel sources in Ubuntu?
[06:47] <Maik> https://www.cyberciti.biz/faq/installing-full-kernel-source-ubuntu-linux/
[06:47] <Maik> first hit on google slidercrank
[06:48] <slidercrank> I did it
[06:48] <slidercrank> but the problem is it installs linux-source-5.4.0.tar.bz2. Just puts this archive to /usr/src/
[06:48] <slidercrank> without unpacking it
[06:48] <slidercrank> so I thought I'm doing something wrong
[06:49] <slidercrank> in other distros you install kernel sources and they appear in /usr/src/kernels/
[06:50] <slidercrank> and I don't find kernel sources with apt-cache search `uname -r`
[06:50] <slidercrank> it finds only headers and modules and some other files
[06:51] <Baikonur> hey, did something special happen on the 21th, at midnight?
[06:51] <Maik> !discuss | Baikonur
[06:52] <Baikonur> one thing I was running on focal suddenly stopped working at the turn of midnight on the 21th
[06:52] <Baikonur> UTC time
[06:53] <Maik> never heard of such a thing happening
[06:54] <Baikonur> it's a python program that's run once per hour with a systemd timer, it works fine if I run it myself but if systemd time runs, there's a http request that times out
[06:54] <caipiblack2> Hi, I try to debug an issue on an embedded device and to do it i need to "delay" the response of DHCP (dhcpoffer). Did you know a dhcp server that can do it  ? dhcpd seems not.
[06:54] <Baikonur> *if systemd timer runs it
[07:27] <gimpnixon> Do I ask questions about 21.04 in ubuntu-devel or where, i'm sorry
[07:30] <Maik> gimpnixon: in #ubuntu+1
[07:30] <Maik> but don't expect any response
[07:31] <gimpnixon> @maik thank you
[07:31] <Maik> yw
[07:35] <slidercrank> Does anyone know how to properly install kernel sources for the currently running kernel? `uname -r` returns 5.8.0-48-generic. I'm using VERSION="20.04.2 LTS (Focal Fossa)". Yes, I've tried different tutorials. They seem not applicable to Focal Fossa
[07:36] <gimpnixon> I am tryign to figure out the process function of Settings>Accessability>Zoom>Color Effects>Brightness. any suggestions?
[07:57] <overclucker> installing kernel source always felt too complicated in ubuntu
[07:57] <curloxide> Compiling kernels always felt too complicated for me
[07:59] <slidercrank> I managed somehow to install it. Not sure if I did it correctly. The problem is that I can't compile an out-of-tree kernel module. It was so easy in Fedora
[07:59] <overclucker> did you need just the headers, or the entire source tree?
[08:00] <slidercrank> the source tree
[08:01] <overclucker> yeah, the ubuntu has like 20 extra steps to prepare the source
[08:03] <slidercrank> make[1]: Entering directory '/usr/src/linux-hwe-5.8-5.8.0'
[08:03] <slidercrank> make[2]: *** No rule to make target '/home/<username>/Workspace/<drivername>/driver/queue.o', needed by '/home/<username>/Workspace/<drivername>/driver/queue.mod'.  Stop.
[08:03] <overclucker> I want is to unpack a tarball to /usr/src and symlink it to /usr/src/linux
[08:03] <curloxide> Aand there goes my idea of installing anbox :(
[08:03] <slidercrank> and this is where I'm stuck
[08:04] <curloxide> Someone forget to change the string from an example?
[08:04] <curloxide> Or is it just censored?
[08:05] <curloxide> (me forgetting to ping slidercrank)
[08:05] <slidercrank> curloxide, no, I replaced it with <...>
[08:05] <slidercrank> because it's irrelevant
[08:06] <slidercrank> the same driver compiles in Fedora
[08:06] <curloxide> So it's kinda censored
[08:07] <superschnell> In Debian, I used to be able to have the unstable repo and only install packages from there if I explicitly specified it with e.g. /unstable behind the package name.
[08:07] <curloxide> Weird. Normally I'd see "no rule to make target 'clean'" or similar
[08:08] <curloxide> Outside my knowledge
[08:08] <superschnell> Can I do a similar thing in Ubuntu? E.g. run 20.10 and then get something from Hirsute?
[08:10] <superschnell> I also learned about
[08:11] <superschnell> "Ondřej Surý" today, and apparently I can lots of packages from his backports repo
[08:11] <superschnell> can get*
[08:14] <superschnell> Specifically regarding PHP
[08:14] <superschnell> So, any advice about running a 'newer' Ubuntu, optionally with 'newest of the new' for select packages?
[08:21] <slidercrank> I've finally solved the problem with driver compilation
[08:50] <de-facto> in ubuntu server, where is the network config?
[08:51] <rory> de-facto, this documentation page is very good: https://ubuntu.com/server/docs/network-configuration
[08:53] <rory> de-facto, to answer your question: It's in multiple places. Check that page for e.g. "Static IP Address Assignment" and "DNS Client Configuration"
[08:54] <de-facto> hmm ok
[08:54] <de-facto> this looks completely different from what i was used to from debian
[08:55] <rory> probably something to do with systemd
[08:56] <de-facto> i was given a server and want to understand its network setup
[08:57] <de-facto> it seems the interface does not even have an ip address but a whole network ?
[08:57] <de-facto> " inet 192.168.0.10/24 brd 192.168.0.255 scope global ens3"
[08:57] <de-facto> how can i even find my ip address?
[08:59] <rory> hostname -I
[08:59] <rory> mabe
[08:59] <de-facto> ah maybe that is the ip address with hostmask 24
[08:59] <de-facto> weird i would have expected a public address, hmmm
[09:00] <rory> there is a web service you can curl to get your public IP from behind NAT: curl ifconfig.me
[09:02] <superschnell> de-facto: such IP addresses are allowed
[09:02] <de-facto> yeah i am behind a NAT :(((
[09:02] <de-facto> ugh this is ugly
[09:03] <superschnell> de-facto: I'm not sure what you mean
[09:03] <superschnell> What do you think is wrong?
[09:04] <superschnell> 0 doesn't mean "network" here
[09:04] <superschnell> There's literally nothing wrong with any of it, NAT or not
[09:06] <superschnell> https://petri.com/csc_ip_subnet_zero
[09:46] <Bluewolf> Good day all. I have a Gigabyte B365M DS3H Motherboard, with Ubuntu 20.04 - I cannot get the front panel audio ports to work or where to even begin solving it?
[09:50] <toffo> Hi there! My Raspberry Pi4B running on Xubuntu 20.10 64bit doesn't boot from SSD any longer after the 5.8.1020-raspi update :(
[09:51] <VMGuy23> My sugguestion: Try using the Micro SD card slot with a MicroSD to USB adapter and see if it boots.
[09:52] <toffo> VMGuy23: yes, it boots from MicroSD to my backup
[09:52] <toffo> Begin: Waiting for root file system ... Begin: Running /scripts/local-block ... mdadm: No devices listed in conf files were found. [the mdadm errormsg gets repeated multiple times...]
[09:52] <toffo> that's what I get on bootup and then some, it boots straight into BusyBox
[09:52] <VMGuy23> Whats BusyBox
[09:52] <toffo> ~
[09:53] <toffo> "Gave up waiting for root file system device. Common problems: <etc.etc. listed>" ...
[09:54] <VMGuy23> Something went wrong writing the config files
[09:54] <toffo> "ALERT! Label "WRITABLE" does not exist. Dropping to a shell."
[09:55] <VMGuy23> what happens when you unplug the SSD (but keep the card in)
[09:56] <toffo> VMGuy23: unplug the SSD at what point? When it swaps the boot-up process to Ubuntu or .. ?
[09:56] <VMGuy23> Just before you turn it on. Use the card only.
[09:56] <toffo> I'm using James A. Chambers' SSD boot-up script, don't know if that might've botched the boot up
[09:56] <toffo> VMGuy23: works just fine from the MicroSD card, like I said :)
[09:57] <toffo> this was right after I updated the 5.8-1020-raspi today
[09:57] <toffo> that build has been working for half a year with all the latest updates oknp
[09:57] <toffo> (and I've never had any issues like this before tbh)
[09:58] <toffo> should I make a vidcap of the boot process?
[09:58] <toffo> because it DOES boot but something goes haywire at some point of the process
[09:58] <VMGuy23> I can't help anymore, but a screencap would be useful for others.
[09:59] <VMGuy23> Hopefully someone else comes to help soon.
[09:59] <EriC^^> toffo: what's the problem?
[10:00] <toffo> EriC^^: Raspberry Pi 4B won't boot from SSD after today's 5.8.1020-raspi kernel update.
[10:01] <VMGuy23> Here's the error: "Begin: Waiting for root file system ... Begin: Running /scripts/local-block ... mdadm: No devices listed in conf files were found. "
[10:01] <toffo> or well, it does to a degree, but then I get BusyBox'd
[10:03] <EriC^^> does it mention anything at the top before it gives the busybox?
[10:04] <VMGuy23> I said above
[10:04] <VMGuy23> He told me earlier
[10:04] <VMGuy23> right at the end, it was: ""Gave up waiting for root file system device. Common problems: <etc.etc. listed>"
[10:05] <EriC^^> aha, you could try to troubleshoot that from busybox, see what it expects to mount and if it exists or what
[10:05] <VMGuy23> It looks like broken .conf files
[10:05] <EriC^^> toffo: do you have a live usb or some live system you can boot to troubleshoot?
[10:05] <EriC^^> toffo: are you using raid?
[10:06] <VMGuy23> Maybe the james A. Chambers' bootup script got broken by the update, you were using that
[10:07] <EriC^^> try an older kernel from grub > advanced
[10:08] <toffo> https://images2.imgbox.com/c3/b9/7kqniKum_o.png so this is what the "end of the boot" looks like. The cursor is blinking active in the "(initramfs)" part of the BusyBox, but I can't type in anything with any wired or wireless USB keyboard.
[10:09] <VMGuy23> How did you take the screenshot?
[10:09] <toffo> VMGuy23: I have a HDMI video capture USB3 dongle
[10:09] <VMGuy23> toffo: Ok, just wondering
[10:09] <EriC^^> toffo: are you using raid? mdadm is for raid
[10:09] <toffo> VMGuy23: you can get those off of eBay/Amazon for $10-15USD today
[10:10] <toffo> EriC^^: nope. Just a Samsung T5 USB3 SSD
[10:10] <toffo> and note that it boots up until that point from the SSD, there's no MicroSD card anywhere
[10:10] <toffo> I'm not using a "boot jump" via MicroSD or anything. It's straight off the SSD.
[10:10] <EriC^^> toffo: ok, maybe see /etc/mdadm/mdadm.conf in case it has stuff it is trying to use
[10:12] <toffo> EriC^^: will take a look now. I wonder if that has something to do with the fact that I'm using the James A. Chambers' script ( https://jamesachambers.com/raspberry-pi-4-ubuntu-20-04-usb-mass-storage-boot-guide/ )
[10:12] <EriC^^> it looks like it's expecting some /dev/md* devices and they dont exist and it ends up giving up waiting, if you're not using raid i think mdadm /dev/md* shouldn't be relevant
[10:15] <toffo> EriC^^: only two lines that are not #-commented out in /etc/mdadm/mdadm.conf: "HOMEHOST <system>" , "MAILADDR root"
[10:16] <EriC^^> toffo: what's in /etc/fstab ?
[10:17] <EriC^^> toffo: also anything in /etc/mdadm.conf ? or also give "locate mdadm.conf" a try and see what files it gives
[10:19] <toffo> EriC^^: There's a few "mdadm.conf" files, one of them in "/etc/modprobe.d/" that says on an uncommented line: "options md_mod start_ro=1"
[10:20] <toffo> EriC^^: There's also "mdadm.config" and "mdadm.conffiles" in "/var/lib/dpkg/info"
[10:21] <EriC^^> toffo: anything interesting in /etc/fstab?
[10:21] <EriC^^> do you have internet access on the live system?
[10:25] <toffo> EriC^^: https://images2.imgbox.com/3d/ef/ria0w4Mq_o.png fstab
[10:25] <toffo> EriC^^: you mean internet access on the box that is not booting now? yes
[10:26] <toffo> could the problem be in /etc/fstab referring to the disk partitions via labels rather than (part)UUID
[10:27] <toffo> I've always thought that pointing out to partitions by their label instead of (part)UUID is wonky a.f.
[10:27] <EriC^^> yeah it's definitely better practice to put the uuids instead
[10:27] <EriC^^> yup
[10:27] <toffo> EriC^^: I have multiple boxes from the same build type tho ... now this other Ubuntu box, also a RPi4B, running on 64bit Ubuntu 20.10, wants to update the 5.8-1020-raspi kernel ...
[10:28] <toffo> And since this happened to the other box right after the update, I don't want to screw up yet another fully working box
[10:29] <EriC^^> toffo: do you still have older kernels on it? try "ls /boot" and look for vmlinuz files
[10:29] <toffo> I'm a/b comparing the other box and this one, the /dev/sr* ones are for the optical drive I sometimes have attached to the RPi that are not in use
[10:29] <EriC^^> i see
[10:30] <toffo> they shouldn't be a problem tho.. or haven't been. yes, I have the vmlinuz
[10:30] <toffo> the 'vmlinuz' file is from an hour ago which is exactly when everything went to hell ;D
[10:30] <EriC^^> multipe vmlinuz?
[10:30] <EriC^^> or just that one
[10:31] <toffo> one file named 'vmlinuz', dated today, and one file named 'vmlinuz.bak', from March 24th.
[10:31] <toffo> there's also a 'vmlinux' file from March 24th.
[10:32] <EriC^^> aha, try "sudo update-grub && cat /boot/grub/grub.cfg | nc termbin.com 9999" and paste the link it gives you here
[10:32] <toffo> EriC^^: ... so that's what you meant by the internet access :D
[10:32] <toffo> (you're assuming I can ssh into the box)
[10:33] <toffo> I haven't tried that out tbh, but since it booted to busybox and doesn't let me input anything via keyboard, I have a nagging feeling that it didn't load sshd
[10:34] <toffo> seems that just about all of the stuff in the boot directory have a ".bak" file that's dated March 24th.
[10:35] <toffo> EriC^^: I have the SSD mounted to another raspberry pi that's running the same distro and is near-identical in many ways
[10:36] <toffo> EriC^^: but did I misunderstand what you said and should I try ssh'ing into the box or?
[10:37] <toffo> because if I run that grub update, it will obviously do it on my rpi build that's alive?
[10:38] <EriC^^> toffo: sorry im back
[10:39] <EriC^^> toffo: ah, ok if the mounted ssd dir, nevermind the update-grub command, only run 'cat /mountpoint/boot/grub/grub.cfg | nc termbin.com 9999'
[10:41] <toffo> EriC^^: cat: /mnt/sdb2/boot/grub/grub.cfg: No such file or directory
[10:42] <toffo> EriC^^: there's no "grub/grub.cfg" on the FAT32 boot partition (/mnt/sdb1) either
[10:42] <EriC^^> toffo: what's in /mnt/sdb1 ? try "ls -lR /mnt/sdb1 | nc termbin.com 9999"
[10:43] <toffo> EriC^^: https://termbin.com/bbxd7
[10:44] <EriC^^> toffo: are the labels still intact btw? give "sudo parted -ls" a shot and have a look
[10:45] <toffo> EriC^^: should that command show the disk labels ? it doesn't show anything for disk labels ...
[10:46] <EriC^^> yeah it should toffo , under "Name"
[10:46] <toffo> EriC^^: I got gparted to show them, yep, they're "system-boot" and "writable"
[10:46] <toffo> EriC^^: I wonder if I should change them to PARTUUID's and give it another go that way.
[10:47] <toffo> I did record the boot-up process to video tho
[10:47] <EriC^^> yes, change them to uuid, not partuuid
[10:49] <EriC^^> toffo: usually when you boot the rpi, can you get grub or it uses some other bootloader?
[10:51] <toffo> EriC^^: I use the James A. Chambers' SSD boot fix that he made for 20.04 and older Ubuntu versions.
[10:52] <toffo> EriC^^: "I know for Ubuntu I had to take a bunch of steps for it to be able to USB boot correctly on the Pi 4. I even went as far as to write a script called “BootFix.sh” which gets Ubuntu 20.04 to boot (otherwise it gets stuck after the Pi screen similar to how HA is)."
[10:52] <toffo> EriC^^: (from: https://jamesachambers.com/new-raspberry-pi-4-bootloader-usb-network-boot-guide/ )
[10:52] <toffo> oh , well, that's the usb network boot guide
[10:52] <toffo> EriC^^: https://jamesachambers.com/raspberry-pi-4-usb-boot-config-guide-for-ssd-flash-drives/ <= it was probably this one.
[10:55] <toffo> EriC^^: cmdline.txt seems to point, once again, to the label of the drive. I could swear I had made these point to the (part)UUID
[10:56] <toffo> I wonder if kernel updates revert that to default and use the drive label instead. then again, if that's not the problem ... let's see.
[10:58] <EriC^> toffo: maybe you have to run the bootfix script again after installing a new kernel?
[11:00] <de-facto> how do i mount a remove ssh directory?
[11:00] <de-facto> seems there is not sshfs anymore?
[11:01] <EriC^> it seems like /boot/firmware/cmdline.txt has the kernel parameters and such, like how you can use nomodeset in grub, maybe adding a rootdelay=10 or so might help in case that's the issue, if it's the mdadm that's the problem i wonder where it's getting the idea to look for anything raid related
[11:01] <EriC^> !info sshfs
[11:01] <toffo> EriC^: that wasn't a problem before, also, pointing to the UUID's didn't help. it still waits for root file system
[11:02] <EriC^> de-facto: looks like it's still there, it could be that universe repo isnt enabled on your system so it isnt showing up
[11:02] <de-facto> ah
[11:04] <de-facto> hmm on enabling that in software sources the "Cache Refresh" window blocks the dialogue
[11:04] <de-facto> should i xkill it?
[11:05] <de-facto> yeah had to kill it, weird
[11:05] <de-facto> ok now its installed
[11:06] <EriC^> toffo: maybe try to chroot into the install from the running system, and maybe either just purge mdadm since you're not using raid, and try to update-initramfs -c -k all so it recreates it for the kernels
[11:07] <EriC^> and maybe have a look in /etc/initramfs-tools for anything related to raid/md
[11:07] <toffo> EriC^: I'm going through that litany of error messages during booth, the "Running /scripts/local-block ... done" being the last one, it seems that I'm not the only one having that problem  https://forum.mxlinux.org/viewtopic.php?f=23&t=56351
[11:08] <toffo> EriC^: ... should I increase the wait time before timeout? :O
[11:09] <toffo> "
[11:09] <toffo> When the initramfs gets updated due to whatever reason (like new kernel, grub-live, manually running update-initramfs -u -k all) you get an additional boot delay of ~30 sec where it runs some scripts/local-block. This is a known issue: https://bugs.debian.org/cgi-bin/bugrepo ... bug=860533"
[11:10] <toffo> although obviously I can't run the "update-initramfs -u" since I can only boot to the BusyBox ..
[11:10] <EriC^> toffo: yeah, i'd check out /boot/firmware/cmdline.txt and add a bootdelay and see there's no md related stuff, and probably just purge mdadm completely, see there's no config for the initram relating to it and update-initramfs
[11:10] <EriC^> if that works then maybe try removing the rootdelay
[11:10] <toffo> EriC^: is there any way to update-initramfs on a detached box
[11:11] <toffo> a.k.a. can I run it somehow on that other rpi box onto the now-problematic SSD
[11:11] <EriC^> made a typo above, it's not bootdelay it's rootdelay
[11:12] <EriC^> toffo: yeah, chroot into the system, type "for i in /dev /dev/pts /proc /sys; do sudo mount -B $i /mnt/sdb2$i; done"
[11:13] <EriC^> then type "sudo umount /dev/sdb1" and type "sudo chroot /mnt/sdb2" then type "mount /boot/firmware" it should mount it, then you can work as if you're in the running system, use dpkg/apt/update-initramfs etc
[11:13] <toffo> EriC^: cool tip, thanks, I didn't know that was possible :D
[11:13] <EriC^> np, yeah it's very useful :D
[11:19] <toffo> EriC^: alright, I'm now over there in chroot, thanks, will have to probe for a bit
[11:20] <isapgswell> hi
[11:20] <isapgswell> i disabled tpm/vtd and now i am getting 8hrs battery life
[11:20] <toffo> EriC^: so, I know this is a dumb question, but JIC, while in the chroot it's safe to apt purge the adadm packages ?
[11:21] <toffo> EriC^: mdadm, that is.
[11:22] <toffo> EriC^: also noticed when running apt --  E: dpkg was interrupted, you must manually run 'sudo dpkg --configure -a' to correct the problem.
[11:23] <toffo> EriC^: I wonder if something got botched up during the update and it wasn't installed properly
[11:28] <de-facto> trying to do a restore of a server hdd, the server got two hdds one i could read and put a backup tar in out network, the other disk *seems* to be empty (i doubt it is), how can i look into a suspected corrupt disk?
[11:28] <de-facto> the server died of a catastrophic hw failure (psu exploded)
[11:29] <toffo> EriC^: https://pastebin.com/0j5JpJmS
[11:29] <toffo> EriC^: there's that kernel compression/recompression thang.
[11:29] <toffo> EriC^: and it fails.
[11:30] <toffo> EriC^: the part that says
[11:30] <toffo> EriC^: "mv: cannot move '/boot/firmware/vmlinux' to '/boot/firmware/vmlinux.bak': Read-only file system" -- that part (sorry!) baffles me a bit
[11:31] <toffo> EriC^: obviously it's the FAT32 boot partition but ...
[11:31] <de-facto> afaik uefi uses fat32 for its boot images
[11:31] <toffo> de-facto: is raspberry pi 4b
[11:32] <de-facto> oh i see
[11:33] <toffo> I wonder should I reinstall initramfs-tools, oddly enough I get "initramfs-tools is already the newest version (0.137ubuntu12)." but I still get prompted on whether or not to install it
[11:34] <toffo> another ??? is the "update-initramfs: deferring update (trigger activated)
[11:34] <toffo> "
[11:35] <toffo> I get the same error of read-only file system if I try to 'apt-get install initramfs-tools' in chroot.
[11:36] <toffo> "dpkg --configure -a" throws me into the loop of update-initramfs generating /boot/initrd.img-5.8.0-1020-raspi
[11:37] <toffo> next step would be to find out what's causing the file system to be read-only for dpkg
[11:38] <waveform> toffo, I take it mount says both partitions are definitely "rw" instead of "ro"?
[11:38] <toffo> waveform: nope. // /dev/sdb1 on /boot/firmware type vfat (ro,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,errors=remount-ro)
[11:39] <waveform> okay, might be worth trying "mount -o remount,rw /boot/firmware" (if I'm remembering that correctly)
[11:39] <waveform> but that may fail if there's some reason it can't be mounted rw (e.g. dirty filesystem which needs fsck first in some circumstances)
[11:39] <toffo> waveform: thanks! it worked.
[11:39] <waveform> good stuff :)
[11:39] <toffo> waveform: my filesystems are always dirty!
[11:40] <toffo> jk , but fsck found a dirty bit
[11:40] <waveform> heh - well, that would explain things
[11:40] <waveform> and yes, the initramfs packages tend to trigger things like flash-kernel to re-run during postinst which'll require a rw mounted boot partition
[11:40] <toffo> waveform: my rpi crashed right after the update, or it might've been running in the background, it might be the cause
[11:41] <waveform> very likely
[11:41] <toffo> am re-running dpkg --configure -a
[11:41] <toffo> and it's now initramfs updating the 5.8.0-1020-raspi
[11:43] <toffo> oh mi gosh it twerked!
[11:43] <toffo> at least the initramfs / kernel update went oknp
[11:46] <waveform> nice :) probably worth checking if it can boot now (exit chroot, cleanly umount first, etc. etc.), assuming dpkg is all done and exited cleanly
[11:47] <toffo> waveform: I'm running the 'update-initramfs -c -k all' now as EriC^ suggested to make sure that the previous ones are on board as well.
[11:48] <toffo> the initrd.img 's from 5.8.0-1017-raspi to 1020-raspi seem to be updating without any errors or warnings.
[11:49] <toffo> waveform: cant unmount the sdb2 partition btw after exiting chroot :O
[11:49] <toffo> waveform: is busy
[11:49] <waveform> make sure you've got no shells with a current directory in that mount
[11:51] <toffo> what in the -- now my host rpi's PTY is not working
[11:51] <toffo> hmm
[11:52] <waveform> and minor note for the future, "-k all" is usually not *that* useful on the pi given that only the latest kernel (or more precisely the one selected by flash-kernel which defaults to the latest) actually gets copied to the boot partition as there's no fallback boot mechanism in the pi's bootloader (currently)
[11:52] <toffo> doesn't boot.
[11:53] <toffo> it seems to get stuck at "scanning for btrfs filesystems" for a moment.
[11:54] <toffo> the outcome is nonetheless the same ...
[11:56] <toffo> https://images2.imgbox.com/05/2c/X50IgJ69_o.png
[11:57] <toffo> it's very strange, notice that in the previous errormsg before I changed the /etc/fstab to point to the partuuid, it didn't find the drive labeled "writable"
[11:58] <EriC^> toffo: sorry was away, i read the backlog and caught up
[11:59] <EriC^> toffo: it looks like it's saying the partuuid=ab... doesnt exist
[11:59] <toffo> I'm wondering if the J.A.Chambers' SSD boot script for ubuntu is colliding with something
[11:59] <BluesKaj> Hi folks
[11:59] <EriC^^> toffo: what's the contents of '/etc/fstab' ?
[12:00] <Deano59> sigh.
[12:02] <toffo> EriC^^: I changed it to point to the UUID's (not PARTUUID's) of the drive
[12:05] <toffo> EriC^^: cmdline.txt probably.
[12:05] <toffo> EriC^^: I'll try by pointing cmdline.txt to the UUID instead of PARTUUID
[12:08] <toffo> EriC^^: changed cmdline.txt to the UUID and now I end up in the same screen but with tha alert on the UUID not existing instead of the PARTUUID (or as it was originally, before that, the label)
[12:09] <toffo> EriC^^: gah, I mean the alert pointing out that the UUID doesn't exist.
[12:09] <toffo> the "Begin: Running /scripts/local-premount" part seems to run for suspiciously long, too..
[12:13] <toffo> "Error during boot: Gave up waiting for root file system device" => seems to be pointing to a grub error perhaps, https://forum.openmediavault.org/index.php?thread/22879-unable-to-boot-after-installation-dropping-to-a-shell/&postID=187321#post187321
[12:15] <Bluewolf> Hello. I have a Gigabyte B365M DS3H Motherboard, with Ubuntu 20.04 - I cannot get the front panel audio ports to work?
[12:24] <toffo> EriC^^: yep it's gotta be this one that's causing a stir: /boot/firmware/auto_decompress_kernel
[12:25] <toffo> EriC^^: https://cstan.io/?p=12531&lang=en related to that one, (or the James A. Chambers' script -- I originally installed Ubuntu 20.04, then updated to 20.10 a while back -- which btw should support native booting from MSD USB on RPI4b)
[12:29] <toffo> After looking at the instructions found at https://cstan.io/?p=12531&lang=en , it seems I have the file "/etc/apt/apt.conf.d/999_decompress_rpi_kernel" that has a line pointing to run a script ("/boot/firmware/auto_decompress_kernel"), but that script simply isn't there. I'm wondering if I should stick it back in there or not.
[12:30] <BluesKaj> Bluewolf, check your alsamixer with F6, and choose which audio device you want to use, then run, sudo alsactl store
[12:32] <toffo> EriC^^: hah! well, this explains a lot, my config.txt seems to have been rewritten
[12:33] <Maik> good afternoon peeps
[12:33] <Deano59> lol
[12:33] <toffo> EriC^^: the latest update seems to have dne that, instead of having lines i.e.: boot_delay \ kernel=vmlinux \ initramfs initrd.img followkernel , etc. , it's been reverted back to rpi4b's default config.txt
[12:34] <Deano59> Maik: not you as well... sigh. :D
[12:34] <BluesKaj> hehe
[12:35] <Maik> Deano59: stop it for once, thanks in advance
[12:35] <Bluewolf> BluesKaj: Done that. No output or change?
[12:36] <BluesKaj> Bluewolf, you have a nvidia gpu correct?
[12:37] <Bluewolf> BluesKaj: Yes. 1660 Super
[12:39] <BluesKaj> there was abug a while ago that would switch the nvidia audio from the onboard when the gpu choice in the uefi bios was changed from onboard to pci..dunno if that is fixed
[12:41] <toffo> I wonder if I should make a forum post somewhere on this? seems really weird, didn't help if I tried to reinstate that script
[12:41] <Bluewolf> BluesKaj: I've a dual boot with Win10. This issue is the same there. I've even bought a new front port audio and cable with the same result.
[12:42] <Deano59> Bluewolf: is your PC under warranty?
[12:43] <Bluewolf> Deano59: Probably - But my situation doesn't permit a replacement.
[12:43] <pikapika> Is there any way to get the kernel module name from a pulseaudio device?
[12:45] <BluesKaj> Bluewolf, speaking of pulse audio/pavucontrol, have you checked the devices there
[12:45] <BluesKaj> ?
[12:46] <Bluewolf> I think I did something along those lines
[12:46] <Bluewolf> Why are you looking for in particular?
[12:46] <Bluewolf> *what
[12:46] <withered_dragon> odd you guys are speaking audio - does anyone play with the audio while using DWM?
[12:47] <pikapika> Funny
[12:47] <pikapika> Interesting coincidence that all three of us came to talk about audio at the same time
[12:47] <BluesKaj> Bluewolf, output devices/port
[12:48] <withered_dragon> Yeah, I tend to just lurk, but since you guys were speaking audio - I noticed just hte other day when using DWM my line-in isn't registered in Discord (but it is if I launch the normal window manager with Ubuntu)
[12:48] <Bluewolf> BluesKaj: Gimme a command and I'll paste the output?
[12:48] <pikapika> withered_dragon, is it firefox
[12:49] <pikapika> or discord app
[12:49] <withered_dragon> discord app
[12:49] <BluesKaj> look in pavucontrol Bluewolf, it's a gui
[12:49] <pikapika> Yeah pavucontrol, is there any way to find the kernel module name or similar for a particular device listed there?
[12:50] <BluesKaj> pikapika, do you mean a driver?
[12:50] <pikapika> Yes
[12:52] <BluesKaj> pikapika, run, sudo lshw -C sound, then look for "configuration: driver"
[12:53] <pikapika> Thanks that helped
[12:54] <BluesKaj> which driver is it ?
[12:54] <withered_dragon> can a person just apt install pavucontrol? I don't see it in my programs in DWM
[12:54] <BluesKaj> yes
[12:55] <pikapika> I just needed a listing of the drivers associated with audio devices for now
[12:55] <pikapika> And it worked
[12:55] <pikapika> withered_dragon, do a which
[12:55] <pikapika> it should be /usr/bin/pavucontrol
[12:55] <pikapika> if it exists
[12:55] <pikapika> but isn't shown in whatever menu you are using
[12:55] <Bluewolf> BluesKaj: https://imgur.com/a/DqMd7SB - https://paste.ubuntu.com/p/CK7Y2FFmxs/
[12:56] <toffo> not working *sigh*
[12:57] <toffo> it's gotta be that kernel compression script that got screwed up somehow with the 5.8-1020-raspi kernel update.
[12:58] <waveform> toffo, sorry - was off doing a couple of other things - out of interest, what version of ubuntu is this? (apologies if it's in the back-scroll - haven't had much time to read it all between other things)
[12:59] <toffo> waveform: Raspberry Pi 4B, 4GB model, running Ubuntu 20.10 64-bit.
[13:00] <waveform> hmmm, I ask because the later versions of the pi bootloader can implicitly unpack the gzipped arm64 kernels now (no need for a decompression script anymore) - I know we've got that version in hirsute, but it sounds like it's worth backporting if I haven't already
[13:00] <EriC^> toffo: ah, great! it's booting now?
[13:00] <BluesKaj> Bluewolf, in your case bith the nvidia and onboard use the same intel hda driver ,,,the front audio outs are the onboard audio by default so choose that one.
[13:01] <toffo> waveform: the thing is that yep, I've run 20.04 on these boxes before (and been tweaking ubuntus onto rpi4b's since 18.x was unofficially supported) -- these particular installs have been upgraded from 20.04 to 20.10
[13:01] <toffo> waveform: updated first one of the boxes today into 5.8.0-1020-raspi and the whole house of cards collapsed :D
[13:02] <toffo> waveform: had been keeping up to date so the previous versions were like 1018, 1019 even ..
[13:02] <waveform> looks like focal and groovy have basically the same bootloader/firmware release back-ported from hirsute but it was early-ish in the cycle and they haven't got the absolute latest
[13:02] <toffo> waveform: I did a cross-comparison with my non-updated raspi4b box that is still running on 1019-raspi
[13:03] <toffo> waveform: it was either the update itself or me re-running initramfs and so forth via chroot that reset my config.txt at least
[13:03] <Bluewolf> BluesKaj: Regardless of which I choose. Back ports are the only thing that work?
[13:04] <BluesKaj> Bluewolf, have youturned the choices off?
[13:04] <BluesKaj> other
[13:04] <toffo> waveform: I'm wondering if there's a method to get rid of the boot script with the compression/decompression + whatever else if those are causing havoc
[13:05] <withered_dragon> pikapika: I just installed it (it wasn't installed) and then was able to set my input to the appropriate device (line in)
[13:05] <Bluewolf> BluesKaj: Sorry I don't follow. Which do I turn off?
[13:05] <withered_dragon> Thankful I was keeping an eye in here this morning!
[13:05] <toffo> waveform: I mean I obviously know where the boot scripts are but I'm kinda weary on trying out any "patent bending" measures in case I manage to F things up even more
[13:06] <pikapika> withered_dragon, good to hear ^^
[13:06] <waveform> toffo, that's very strange - nothing should be touching config.txt - certainly nothing in initramfs-tools should be doing that. For groovy, ideally you should have a config.txt that has kernel=vmlinuz, initramfs initrd.img followkernel, cmdline=cmdline.txt and nothing about device_tree at all
[13:06] <waveform> (I only mention the last because prior versions had some device_tree stuff in there for u-boot which was ditched on groovy)
[13:07] <BluesKaj> Bluewolf, if you aren't using digital out then turn off everything except the analog stereo
[13:08] <toffo> waveform: so the rpi4b that currently works (with this script, or the James A. Chambers variation of it; https://cstan.io/?p=12531&lang=en ) uses, according to config.txt: "kernel=vmlinux", "initramfs initrd.img followkernel"
[13:08] <BluesKaj> Bluewolf, you may want to relogin afterwards
[13:09] <toffo> waveform: perverse idea, but since these two boxes are essentially the same, well, not even essentially, quite literally so, maybe I should just rewrite the working box's boot partition's contents to the other one
[13:09] <toffo> (obviously will take backups first :D )
[13:11] <waveform> toffo, frankly that's not a bad idea - the only bit I'd caution about is cmdline.txt which tells the kernel where to find the root partition. It sounded earlier like you had some interesting drive setup that demanded using PARTUUID matching instead of LABEL matching (which I think is the default) - might be worth preserving that (or not if it turns out that doesn't work!)
[13:12] <Bluewolf> BluesKaj: Okay. What's the difference between digital out and analog stereo?
[13:12] <Bluewolf> Disable through the terminal or Gui
[13:13] <toffo> waveform: nope, nothing special -- but I have to say that with the default of /etc/fstab pointing to drive labels _only_, that's just imho asking for trouble in the long run. it should always be UUID by default
[13:13] <toffo> waveform: just my two pennies on that subject :D
[13:13] <waveform> toffo, unfortunately given that the only pi images we ship are pre-installed I don't have the option of using UUIDs. It *might* be possible in the desktop which has post-installation process but not on the server images
[13:14] <BluesKaj> Bluewolf, use the gui
[13:14] <toffo> waveform: copied all that junk from my other raspberry pi 4b running the same ubuntu 20.10 and TAH DAH. It twerks.
[13:15] <toffo> I mean I ditched all the stuff in the boot filesystem into a backup directory on another drive and just copied everything from the working rpi4b's boot/firmware over there.
[13:15] <pikapika> toffo, even with uuids you have to be careful
[13:15] <pikapika> I have a funny story regarding that
[13:15] <waveform> toffo, good stuff - incidentally if you want to file a bug about UUIDs instead of LABELs, ping me the number and I'll stick it on my list but I can't promise it'll get much attention until a few other cloud-init issues get squashed (which is the mechanism we'd be most likely to use to re-write fstab)
[13:16] <pikapika> Hint: It involved dd. Guessing the rest of the story is left as an exercize to the reader.
[13:16] <Bluewolf> BluesKaj: If I'm not mistaken, I'm not using digital?
[13:16] <toffo> waveform: although now the next mental anguish comes from the fact that the next update might break it again
[13:16] <BluesKaj> correct
[13:16] <BluesKaj> Bluewolf, ^
[13:17] <toffo> waveform: I'm wondering if I can migrate out of that kernel head compression-decompresion boogie somehow
[13:17] <waveform> pikapika, I can imagine >.<
[13:17] <pikapika> waveform, I was wondering why its randomly booting to an "older state" of my os
[13:17] <pikapika> I decided to check grub config. Oh...
[13:18] <BluesKaj> digital output that is, Bluewolf
[13:18] <toffo> waveform: it works now but it also reverted back to 5.8.0-1019-raspi ... will run apt update for that and see if it breaks down again
[13:19] <waveform> toffo, that's probably a good idea (getting rid of the decomp stuff - that should've been unnecessary in groovy anyway as we ditched u-boot-by-default in that release which also demanded we unpacked the kernel until such time as the native bootloader grew the ability to handle that)
[13:19] <toffo> waveform: considering the uuid's ping you the number of  -- ?
[13:19] <Bluewolf> BluesKaj: It doesn't appear to have been selected in the first place - Doesn't resolve the issue.
[13:20] <toffo> waveform: yep ... the question remains, how :D
[13:20] <toffo> waveform: feels like I'm running an OS that's built on a timebomb now :D
[13:20] <toffo> waveform: waiddaminute, that's just normal computing!
[13:20] <waveform> toffo, oh - if you wanted to file about bug about that (I'm not entirely sure what you'd file it against - erm, probably linux-firmware-raspi2 given that's where the pi bootloader is)
[13:21] <toffo> waveform: odd, btw, apt updates didn't seem to want to update to anything as for the kernel
[13:21] <BluesKaj> Bluewolf, did you turn everything else off ?
[13:21] <toffo> waveform: so it's still 1019-raspi ... actually, perhaps dpkg still thinks I'm already running on 1020
[13:21] <waveform> toffo, hmm, but I take it 1020 is still under /boot/ ?
[13:22] <toffo> yes
[13:22] <toffo> but all sysinfo outputs say that the kernel is 5.8.0-1019-raspi
[13:23] <waveform> right, because presumably what you copied was the older kernel - let me just check what the latest is on groovy
[13:23] <toffo> waveform: what would be the preffered way to reinstall it?
[13:23] <toffo> waveform: oh the 1020-raspi came just today, my other rpi4b box is begging to be updated
[13:25] <Bluewolf> BluesKaj: Yes. https://imgur.com/a/0AYEwYu
[13:25] <Bluewolf> But I'd need to log out and back in?
[13:25] <waveform> toffo, hmmm -1020 is only in -proposed - shouldn't have been updated already. The preferred way to copy "boot stuff" (bootloader, 2nd-stage, kernel, initrd, etc.) to the boot partition is with "sudo flash-kernel" which should select the latest kernel in /boot/ but can be forced to use an older one with "--force <version>"
[13:26] <_lucifer> Hi! I am currently using ssh key authentication to login in to remote Ubuntu server. Before setting up key based auth, I had a password based login. I have forgotten my password. Can I reset my password on my own or do I have to contact an admin?
[13:26] <BluesKaj> Bluewolf,your HDMI2 is still "on"
[13:27] <toffo> waveform: running 'sudo flash-kernel' , it's updating to 1020-raspi. and yes, i'm using the xubuntu-dev etc "cutting edge" ("don't cut yourself!") PPA's
[13:27] <toffo> waveform: flash-kernel went thru without any problems ...
[13:28] <waveform> toffo, could you just "ls -l /boot/firmware/vmlin*" and have a look at the file sizes for vmlinu[xz]?
[13:29] <toffo> ... aaaaaaaaaaand on reboot, I'm stuck on the splash screen :D
[13:31] <toffo> waveform: so yeah, flashing to 1020-raspi defo breaks the whole OS, at least if you have those previous SSD boot workarounds hnaging about
[13:32] <toffo> waveform: oh yes, one moment phleez.
[13:32] <Bluewolf> BluesKaj: https://imgur.com/a/NIiJeGm
[13:32] <waveform> toffo, definitely sounds like that decompression script is breaking something. For now I'd recopy that boot partition, boot it again, re-run flash-kernel, then we can take a look at what it's done to the boot partition
[13:34] <toffo> waveform: okay, so where should I start -- I now have the broken version attached to my working RPi4B build
[13:34] <BluesKaj> Bluewolf, why the equalizer, have you checked it's volume control?
[13:37] <waveform> toffo, I'd start by re-copying the boot partition that you copied before to get it booting. Once that's done, reboot it (hopefully successfully) then we can go from there
[13:37] <toffo> waveform: ## broken rpi4b boot partition [after update to 5.8.0.1020-raspi]: https://termbin.com/xlfh
[13:38] <toffo> I'll re-copy the working 1019-raspi stuff on top of that now.
[13:38] <waveform> okay
[13:39] <waveform> toffo, ahh - there's the first intersting bit: note that vmlinuz (the supposed compressed kernel) and vmlinux (the supposed uncompressed kernel) are the same size - so at least we know firstly that the decompression scripts are definitely not needed
[13:41] <toffo> waveform: ## here's the directory listings for the working rpi4b's /boot/firmware with subdirectories (5.8.0-1019-raspi): https://termbin.com/08u1
[13:42] <toffo> I'll unmount that drive, stick it back into the other rpi4b and boot it up.
[13:44] <toffo> waveform: booted up oknp
[13:44] <Bluewolf> BluesKaj: I installed the equaliser hoping to improve the back sound quality - Cause the front ports weren't working. Didn't do much in the volume control?
[13:45] <BluesKaj> Bluewolf, try turning it off for now
[13:46] <waveform> toffo, okay - firstly I'd ditch that /boot/firmware/auto_decompress_kernel script as it's manifestly not needed; next I'd check if there's a /etc/apt/apt.conf.d/999_decompress_rpi_kernel file which it looks like the original instructions you were following created - ditch that too if it exists
[13:46] <toffo> waveform: I'm going thru dmesg atm and there's this that and the other :D
[13:47] <toffo> waveform: most of them are the usual "you have been warned" tho
[13:47] <Bluewolf> BluesKaj: Pardon me but not sure where too
[13:47] <waveform> toffo, could you pastebin your /boot/firmware/config.txt ?
[13:48] <toffo> waveform: yes :) one sec plz.
[13:48] <toffo> toffo needs a double-cup of espresso or toffo turn into baboon :_D
[13:49] <waveform> I sympathize :)
[13:51] <BluesKaj> Bluewolf, in pavucontrol
[13:56] <toffo> waveform: I've included the contents of 'syscfg.txt' and 'usercfg.txt' as well, just in case. https://pastebin.com/V112yuQt
[13:58] <toffo> waveform: apropos, I'm wondering if the actual KMS driver is soon ready -- the last time I asked wimpy about it, his recommendation was to stick to fkms
[14:00] <waveform> toffo, okay - wow, there's a lot weird stuff in there. Let's start at the top - I'll whip up a fixed config.txt in a mo, without usercfg.txt and syscfg.txt - those latter files aren't needed anymore (they don't exist in the groovy release image) and are basically due to a mistake I made in focal. On KMS: yes, you want to stick to FKMS for now. We'll actually be switching to KMS in hirsute for the desktop but that's *solely* because wayland breaks
[14:00] <waveform> without it - unfortunately with the "full" KMS driver other things break (like the traditional pi camera applications)
[14:00] <toffo> waveform: other than this little "OS nuking mishap" (:D) btw, I have some mad uptimes under near-constant 100% loads on the rpi4b running ubuntu 20.04 or 20.10 for the past few months or so.
[14:00] <Bluewolf> BluesKaj: https://imgur.com/a/Bg4RjFI
[14:00] <Bluewolf> Haven't got the option
[14:00] <toffo> waveform: a.k.a. good work otherwise __b
[14:01] <waveform> toffo, on weird stuff: any particular reason you've got gpu_mem=512 in there? There's usually little point beyond 256 but there's a few specialist cases (usually obscure camera things) that can require more - just wondering if you fall into one of those categories or if I should just stick 256 in the replacement config.txt?
[14:01] <toffo> waveform: I've been more or less migrating stuff since 18.x LTS
[14:02] <toffo> waveform: yes, that's because I am a lunatic. And, I've found that along with a fine tuned zram-swap to be the best way to make Youtube playback in browsers as smooth as possible
[14:03] <toffo> waveform: and it's not just that, I run some experimental video-related stuff on both the camera interface and over GPIO
[14:05] <toffo> waveform: from what I've understood, the video RAM is allocated from the board's main RAM so, well, I thought that I'd just throw that in there and see what happens. Very nice!
[14:05] <toffo> waveform:... and it's not just me, I've seen video RAM allocations like that used especially with emulator gaming and so forth.
[14:07] <waveform> toffo, ah - okay - if you're using camera things you definitely want to stick with FKMS for now (KMS will likely break things for you) - and yes, lots of people stick >256 in there - the *vast* majority of the time it's fairly pointless :)
[14:08] <toffo> waveform: also compositor off in xubuntu helps out a lot, (as an example:) "export CLUTTER_VBLANK=none" >/etc/profile.d/compositoroff.sh
[14:08] <BluesKaj> Bluewolf, now try reloading the kernel module, sudo modprobe snd_hda_intel, then reboot...there won't be any output from that command if the driver loads properly
[14:09] <BluesKaj> then reboot
[14:09] <waveform> toffo, here's a replacement config.txt: https://paste.ubuntu.com/p/7gYwcJ7wN9/ that combines the user/sys/config.txt files into one, removes the cruft (repeated redundant settings under [pi2/3]), and points kernel at the right file (vmlinuz)
[14:09] <toffo> waveform: there's that one xfwm4 tweak to the compositor as well in xubuntu, turn that thing off and the UI is way better.
[14:10] <Bluewolf> BluesKaj: Be back in a moment
[14:10] <Bluewolf> Thanks
[14:10] <BluesKaj> ok
[14:10] <waveform> toffo, I've left the gpu_mem=512 for the moment, but I'd suggest trying 256 and seeing if you notice any signficant difference because if not you may as well have a bit more RAM to play with
[14:10] <BluesKaj> +
[14:11] <toffo> waveform: now that I think of it, the camera module actually isn't in use in some of the stuff I make for customers, and the GPIO's pretty much just used for fan control
[14:11] <toffo> waveform: so, it varies.
[14:12] <waveform> toffo, in that case feel free to play with KMS. As I mentioned, we'll be switching the desktop image to that in hirsute, but not the server image (as I want to keep the camera stuff working wherever I can until the libcamera replacement bits mature)
[14:12] <toffo> waveform: and also, in the 8GB model, I was thinking of RAMdisk possibilities
[14:12] <toffo> waveform: thanks so much for cleaning up my tangled messes btw ! how much do I owe you :D
[14:12] <waveform> toffo, if you have a swapfile around it's also very worthwhile considering zswap as that's built into the kernel and is rather more flexible that zram (but requires that you have a disk-backed swap-file - if you don't then stick with zram)
[14:14] <waveform> toffo, all gratis :) anyway, I must go try and unbreak bluetooth and wifi on an obscure CM4 variant - good luck!
[14:14] <toffo> waveform: thanks! :D
[14:14] <toffo> waveform: have a great one
[14:15] <Bluewolf> BluesKaj: Rebooted. Doesn't appear to have done anything.
[14:15] <toffo> waveform: btw this is the script that i used to tweak up the zram swap, made things lightning fast, esp. on the SSD. https://github.com/foundObjects/zram-swap
[14:16] <BluesKaj> Bluewolf, ok, make sure your alsamixer automute is disabled...think i mentioned it earlier, but i've run out of ideas
[14:16] <waveform> toffo, thanks - I'll take a look - I need to write a blog-post on this and zswap as I'm seriously considering enabling the latter on the desktop images by default given they have a swap-file pre-existing in the image
[14:17] <toffo> waveform: btw, if you ever happen to read this; I've been kinda bouncing back and forth on the optimal "max_framebuffers" setting, especially if the application is for one monitor only -- found some talk on raspberry foundation's forums that it should be 1 if you're only using 1 screen, BUT ...
[14:17] <toffo> waveform: oh ja, I'm using zswap on my desktop ubuntu as well
[14:18] <toffo> things are lightning fast and probably the same thing goes for the SSD's wear and tear too :DDD
[14:18] <BluesKaj> Bluewolf, which media player do you use?
[14:19] <Bluewolf> BluesKaj: Rhythmbox, VLC
[14:19] <Bluewolf> https://imgur.com/a/vVw74Ev
[14:20] <Bluewolf> BluesKaj: I don't know if its something in the Bios. But I couldn't find anything
[14:20] <Bluewolf> Back works fine
[14:20] <Bluewolf> Just not the front
[14:20] <Bluewolf> I dont understand
[14:21] <BluesKaj> neither do I
[14:21] <BluesKaj> makes no sense
[14:24] <BluesKaj> which audio device is VLC using, click on audio tab in VLC
[14:27] <Bluewolf> BluesKaj: Built-In Audio Analog Stereo, same with Rhythmbox
[14:30] <BluesKaj> do you have a headphone volume control in alsamixer?
[14:33] <Bluewolf> BluesKaj: https://imgur.com/a/YOhJkw2
[14:33] <Bluewolf> But changing it makes no difference. The system isn't recognising anything plugged into the front ports.
[14:33] <Bluewolf> I was wondering if its a bios setting
[14:34] <Bluewolf> But I've checked. Should just work
[14:34] <BluesKaj> it could be directing the audio thru hdmi if you have the nvidia gpu chosen in the bios
[14:35] <BluesKaj> bypassing the front audio outs
[14:37] <Bluewolf> I can take another look. Not a fan of playing around in the bios.
[14:37] <Bluewolf> I know this isn't an ubuntu question - but is there an option to change the audio to the front ports?
[14:38] <BluesKaj> not directly
[14:39] <BluesKaj> the front should be active using the onboard audio outs
[14:39] <BluesKaj> unless you have the spdif outs enabled in alsamixer
[14:41] <Bluewolf> Again - This issue extends to windows 10 partition, not just linux
[14:45] <BluesKaj> dunno much about W10, haven't used windows for many yrs
[14:46] <juliandroske> me either, linux sometimes is much powerful than w10
[14:47] <Bluewolf> The point is the issue extends there too. So its likely in the bios I guess...
[14:47] <juliandroske> and sometimes the audio device would power down under w10 because of a driver issue
[14:48] <Bluewolf> Which I tried fixing there and still same thing
[14:48] <Bluewolf> Installed the driver and didn't work
[14:54] <BluesKaj> could be a HW problem, but he left so....
[14:55] <juliandroske> he's back
[14:55] <Bluewolf> woops - network
[14:59] <Bluewolf> Did I miss anything. Disconnected after my last message?
[14:59] <zed-zahir> hello everyone, chkrootkit found this "user zahir deleted or never logged from lastlog!" while my own username is "zahir"
[14:59] <juliandroske> The system's volume control (sound mixer) can be a tester to the speaker
[15:02] <Bluewolf> How does that have an effect?
[15:07] <juliandroske> A ding, you know that, click the seek bar, make a sound....
[15:16] <zed-zahir> is it normal to see in "lastlog" "**Never logged in**" for my main user?
[15:22] <krytarik> If you never logged in at the TTY, then yes.
[15:22] <zed-zahir> ah I see
[15:22] <zed-zahir> yeah I never logged in TTY
[15:23] <krytarik> Yeah, it's rather seldomly needed these days..
[15:34] <imi> hi, there used to be an ignore_nice_load in cpufreq, and now I cannot find anything like that. is there a way of precenting the cpu from speeding up if the cpu load is nice'd?
[15:48] <toffo> waveform , EriC^^ and others -- thanks for the help! I got the RPi4B kernel 5.8.0.1020 working, and got rid of the compressed kernel stuff etc. -- Made a quick script on automating that, probably a lot of bad practice in it but ; https://pastebin.com/VAqBknR8
[15:48] <EriC^^> toffo: good to hear! well done
[15:50] <toffo> EriC^^: thanks! I'm wondering if I should try out normal use with the actual KMS driver or not
[15:50] <Rapture> I am trying to figure out why some of my older instances are hanging on to many linux headers. I am regularly running apt-get clean // apt-get autoremove --purge -y but it doesn't seem to help. A new instance will have only 2 headers which is what I would like the older ones to get. Any suggestions on what I could do? I'm hesitant about manually removing any.
[15:51] <toffo> Rapture: dpkg -l | grep <your search string>
[15:51] <leftyfb> Rapture: by "instance" do you mean older releases of Ubuntu? What versions?
[15:52] <EriC^^> toffo: why not, better and better :D
[15:52] <Rapture> leftyfb - All instances are Ubuntu 16.04
[15:52] <EriC^^> Rapture: have you not rebooted them in a very long time?
[15:53] <Rapture> EriC^^ - Correct, some of them I would prefer to not be rebooted so they have been running for a while.
[15:53] <leftyfb> Rapture: autoremove will only remove the versions older than the kernel version you're running.
[15:54] <leftyfb> Rapture: you know 16.04 will become EOLN next month right?
[15:54] <toffo> EriC^^ and waveform , if you are devs/otherwise associated with canonical or so, take into consideration for the rpi4b in particular (due to the MSD USB boot hassle) on what "safety measures" could be implemented to avoid breakage if users have installed the OS using those bootloader methods
[15:54] <leftyfb> Rapture: sorry, EOL, not EOLN
[15:54] <toffo> ( forgot the pretty please in between there ==^ :D )
[15:54] <toffo> feel free to use my script tho! it fixed things for me.
[15:54] <Rapture> So outside of rebuild I would likely need to first upgrade/update the kernerl to the latest and then try to remove?
[15:55] <Rapture> kernel*
[15:55] <leftyfb> Rapture: you'll need to be running the latest kernel to remove the older ones. As in, reboot into the newer kernel
[15:55] <Rapture> hmm ok, thanks.
[15:56] <EriC^^> Rapture: you can manually remove them if you want to avoid rebooting, just make sure to keep the current running one and the latest
[15:57] <EriC^^> using apt-get purge and "dpkg -l | grep linux-image" to get the names
[15:58] <Rapture> baa yeah through my searches I've seen the manual removal suggestion. Was hoping to avoid that one. Thanks for the suggestions.
[16:04] <R13ose> 20.10: I'm using my bluetooth wireless headphones and they have very low sound.  When I plug them into my computer they work but wirelessly they don't.  How do I fix this issue?
[16:29] <Celphi> anyone here familiar with auditd
[16:30] <lotuspsychje> ask your specific issue into the channel Celphi so volunteers van hel think along with you
[16:30] <Celphi> i need a rule that records failed attempts to create a file by unauthorized users
[16:53] <sarnold> Celphi: here's some rules I've got for that https://paste.ubuntu.com/p/tQZ7zNkWHw/ -- BUT be aware that there's probably new syscalls that need to be added to this list, and it ignores all the 'system service' accounts
[17:31] <Arg-Patagonia-b> hello
[17:43] <Arg-Patagonia-b> Arg-Patagonia-b: hi
[17:43] <Arg-Patagonia-b> archpc:
[17:44] <leftyfb> Arg-Patagonia-b: what can we do for you?
[17:44] <Arg-Patagonia-b> I am learning how to use Finch in text mode
[17:45] <Arg-Patagonia-b> new here
[17:47] <goodTimesImmort> (cdw) Is there a way to tell cdw that I want to burn a folder and all of its content instead of having to choose the files inside said folder one by one?
[17:50] <leftyfb> !test | Arg-Patagonia-b
[17:50] <leftyfb> Arg-Patagonia-b: this is not the place to test and IRC client. Try #test
[17:50] <Arg-Patagonia-b> Thanks!
[17:56] <leftyfb> Arg-Patagonia-b: please do not private message
[17:59] <sarnold> goodTimesImmort: the manpage says "Space key to select files or directories." -- does that work?
[18:00] <R13ose> Any thoughts on my question?
[18:03] <goodTimesImmort> ty @sarnold. It did.
[18:05] <sarnold> goodTimesImmort: yay
[18:05] <sarnold> R13ose: sorry, I've never done bluetooth stuff, it just seems so painful
[18:06] <R13ose> what would be better than bluetooth?
[18:07] <sarnold> I use an android phone for my videoconferencing needs
[18:08] <sarnold> I'm going to try my computer soon, and whether it succeeds or fails will be down to bluetooth., heh
[18:26] <PeGaSuS> today, after a few updates and fixes because the lack of an internal monitor on the laptop: https://imgur.com/a/EWdGd5Z
[18:26] <PeGaSuS> Ubuntu ftw! just to say how much I love this OS
[19:01] <DzzzD> ping
[19:01] <coXZist> pong
[20:24] <Dro> hello, I'm trying "ubuntu-drivers autoinstall" and having this error :  linux-modules-nvidia-460-generic : Depends: nvidia-kernel-common-460 (<= 460.39-1) but 460.67-0ubuntu0~0.18.04.1 is to be installed
[20:24] <Dro> any idea how can I fix this please?
[20:39] <alterjsive> nvme0n1-part2   500.1 GB    btrfs   /
[20:39] <alterjsive> sda-part2       297.8 MB    ext4    /boot
[20:39] <alterjsive> sda-part3       299.7 GB    btrfs   /mnt/disks/sda1
[20:39] <alterjsive> is this a valid partitioning layout?
[20:39] <sarnold> 300M is really small for /boot
[20:40] <alterjsive> I want to boot from sda because hp doesnt allow me to boot from my non hp nvme drive
[20:40]  * cbreak votes for ZFS
[20:40] <cbreak> the ubuntu zfs installer creates a 2GB /boot I think
[20:40] <cbreak> and even that was a bit small I think
[20:41] <learnlearnlearn> Hi all, I'm trying to recompile the kernel being used in my PV guest. Are the sources for kernels used in PV guests maintained in a separate branch or kernel repository? I'm using focal. I've described my question in full here(https://askubuntu.com/questions/1327833/recompiling-pv-kernel-unable-to-get-domain-type-for-domid-4). Any help would be
[20:41] <learnlearnlearn> much appreciated, thanks.
[20:43] <alterjsive> sarnold: ok,  redhat recommends this size IRRC
[20:44] <sarnold> alterjsive: curious, I wonder what they've done differently to let that work well
[20:44] <sarnold> learnlearnlearn: are PV kernels still a thing? I thuoght everyone just kinda switched to using virtio drivers for network, storage, and called it a day..
[20:45] <alterjsive> cbreak: why zfs?
[20:46] <cbreak> because I like it, it never caused me problems (without me provoking it :D), it's reliable, powerful, flexible
[20:47] <alterjsive> ok thanks for the advice guys, I'm gonna pick this back up tomorrow.
[20:48] <cbreak> don't forget the EFI partitions :)
[20:48] <sarnold> ah yes :)
[20:49] <tidux> Is OpenSSH port forwarding broken on 18.04?
[20:50] <tidux> I'm doing "ssh -R <remoteport>:localhost:<localport> user@remote.host" and it throws "bad remote forwarding specification" despite that being the exact syntax required
[20:51] <tidux> I've even tried using "127.0.0.1
[20:51] <tidux> to make sure it's not a DNS error
[20:53] <cbreak> tried with -v?
[20:53] <cbreak> maybe that tells you more
[20:54] <tidux> I tried with -vvvvv, LANG=C, running under /bin/sh to avoid any bashisms, on a tty instead of X to avoid any potential Unicode input problems
[20:55] <tidux> and it just threw the same error
[20:55] <tidux> it looks like that error is due to OpenSSH not recognizing the argument syntax based on the error it's throwing, but that can't possibly be right
[20:55] <tidux> as a toy example I'm testing this with an mpd connection before anything serious
[20:56] <tidux> $ ssh -R 66600:localhost:6600 user@remote.host
[20:56] <tidux> by EVERYTHING I've read that should work
[20:56] <learnlearnlearn> sarnold: got disconnected for a bit. I was following the instructions at https://help.ubuntu.com/community/Xen (under Manually Create a PV Guest VM) so I was assuming that they were still supported
[20:56] <cbreak> tidux: use a valid port
[20:57] <tidux> cbreak: that is a valid port
[20:57] <cbreak> tidux: smaller than 65536
[20:57] <tidux> oh
[20:57] <cbreak> UDP and TCP port numbers are 16 bit
[20:58] <sarnold> learnlearnlearn: aha.. so, xen's never really been an important thing in ubuntu, we standardized on libvirt instead; and even though this page was edited in 2015 I really wouldn't be surprised if it was stale even for 2015..
[21:00] <tidux> ah that works, I'm dumb
[21:00] <tidux> thanks cbreak
[21:00] <cbreak> np :)
[21:01] <learnlearnlearn> sarnold: ah i see, i just noticed the document was last edited in 2015. fwiw the contents on the page are still correct/complete and they work; its just that i'm having trouble re-compiling the kernel inside the focal guest(outside the scope of that document)
[21:02] <sarnold> learnlearnlearn: trouble recompiling? or a successful recompile and then errors elsewhere?
[21:03] <learnlearnlearn> sarnold no problem recompiling. after doing a "dpkg -i *.deb" in the guest with the newly built kernel, it just doesn't boot anymore. the console kind of hangs so im not sure how i can proceed to debug it either
[21:05] <sarnold> learnlearnlearn: aha. dang. that's not a fun place to wind up.
[21:40] <Sven_vB> hi! in bionic, is there an easy way to selectively override enable-wide-area= in /etc/avahi/avahi-daemon.conf but keep having the (reminder of the) file be managed by dpkg? "man 5 avahi-daemon.conf" doesn't look like it supports systemd drop-in files or anything like that.
[21:51] <frad> Mathisen, are you by any chance matsaman?
[21:52] <frad> If I play the same movie encoded in x264 and x265, same quality, but the x265 is 50% the size of the x264, is my computer going to use less cpu to play it?
[21:56] <Mathisen> frad, im not and you can check this by doing > /msg nickserv info matsaman