[00:00] <gnUser> There is this article that clearly explains that it is supported but no instructions on how to enable it!  https://wayland.freedesktop.org/libinput/doc/latest/touchpad-thumb-detection.html
[00:03] <Maik> gnUser: afaik 20.04 uses Xorg not wayland
[00:06] <gnUser> Maik: true so Xorg does not support thum detection?
[00:09] <Maik> gnUser: i don't know tbh
[00:10] <Ublx> i have a problem: i removed an external hard disk without unmounting before and now 1 of 3 partitions has the error: "error unlocking, failed to activate, file exists". it would be great if someone could help me. thanks.
=== xfnw is now known as vulpine
[00:31] <guiverc> Ublx, did you `fsck` (file-system check) the partition(s)?
[00:45] <Ublx> no, guiverc. what should I type: just fsck or any arguments?
[00:46] <TJ-> Ublx: sounds like you have LUKS encrpyted, and the device-mapper files still exist from when you ripped it out
[00:46] <Ublx> TJ-: yes, it's LUKS encrypted
[00:46] <guiverc> Ublx, I'd boot a live system & `fsck` from there; use whatever tool you prefer, command, gparted, kde partition manager etc (ie. GUI is okay too).   adjust for your box but take note of TJ- first (I didn't see what he does)
[00:47] <TJ-> guiverc: this is a LUKS/device-mapper issue
[00:47] <Ublx> my internal disk is okay, it's an external hard disk
[00:47] <Ublx> what should i do TJ- ?
[00:48] <TJ-> Ublx: OK, so you need to 1) identify the device-mapper name that is stale (the LUKS device name) by looking at "ls /dev/mapper/"  then 2) manually removing the stale name with "sudo dmsetup remove THE_STALE_NAME" - after that LUKS can unlock and recreate the name correctly
[00:48] <Ublx> it says exactly "Error unlocking /dev/sdc3: Failed to activate device: File exists"
[00:49] <TJ-> Ublx: I assume you've an /etc/crypttab with names to be given to the unlocked device, e.g. maybe "sdc3_encrypted"
[00:49] <TJ-> Ublx: in my case the names are LUKS_something_meaningful
[00:50] <Ublx> hm, there is one with luks-e05b2f.... is that the one?
[00:51] <TJ-> Ublx: might be - the e05fb... could be a GUID of the device. compare that with "sudo blkid /dev/sdc3"
[00:51] <Ublx> There's sda5_crypt, too? That's also not my internal drive, as far I as know from df.
[00:51] <TJ-> Ublx: if that matches you've got the correct one
[00:51] <TJ-> Ublx: 'sda' is usually the first storage device discovered which we'd normally expect to be the boot device
[00:52] <Ublx> blkid has no output
[00:52] <TJ-> Ublx: that is strange
[00:53] <Ublx> I have only these: /dev/sda   /dev/sda1  /dev/sda2  /dev/sda5  /dev/sr0
[00:53] <Ublx> And of course, I unplugged the external hard drive completely.
[00:53] <TJ-> you'd expect blkid to report something similar to this for a LUKs partition: " /dev/nvme0n1p2: UUID="db9cd343-9ca9-4115-845f-d27943c14437" TYPE="crypto_LUKS" PARTLABEL="Linux dm-crypt" PARTUUID="15f7095d-5bbb-4a43-8324-9867b91dc7d6"
[00:54] <Ublx> hm...
[00:54] <TJ-> Ublx: check sda5 then, with "sudo blkid /dev/sda5" and see if its GUIDs match the earlier mentioned
[00:55] <TJ-> Ublx: it is possible that when the disk was removed it was known as sdc and now is sda
[00:55] <Ublx> it's just followed by an output and does nothing with sda5, right?
[00:56] <TJ-> Ublx: you'd expect a report to indicate sda5 is LUKS/encrypted
[00:56] <Ublx> /dev/sda5: UUID="43f8104c-c836-44a0-a0eb-c7ffe478abbb" TYPE="crypto_LUKS" PARTUUID="4ebfa39c-05"
[00:56] <Ublx> but my local drive is also encrypted
[00:58] <TJ-> Ublx: that may be it then; check by looking at the tree shown with "lsblk" -- follow sda > sda5 > ... because that will show what is mounted, and where
[00:59] <TJ-> Ublx: basically, you're got to avoid removing the active device(s) by identifying the stale device-mapper device that belonged to the removed disk
[00:59] <Ublx> no, the tree ends at the root and swap as you expected. :(
[01:00] <TJ-> Ublx: for sda5? so that's active then, good. Which by a process of elimination indicates that "luks-e05b2f..." is likely the suspect
[01:00] <Ublx> 'sr0 rom' is the only 'curious' I would say.
[01:00] <TJ-> Ublx: so now check the status of that device with "sudo dmsetup info luks-e05b2f ..." (put in the COMPLETE GUID as shown by "ls /dev/mapper/" )
[01:01] <TJ-> sr0 = SCSI Read-Only device 0
[01:01] <TJ-> a.k.a. CD-ROM :)
[01:01] <remline> Do I see correctly that 'focal-backports' pretty much has only one program in it, i.e., cockpit?
[01:02] <Ublx> TJ-: I have the output. What do you need? ACTIVE, LIVE, ..
[01:07] <TJ-> Ublx: if it is reporting as ACTIVE, LIVE that might not be the stale one
[01:07] <TJ-> Ublx: how many devices does "sudo dmsetup ls" report ?
[01:07] <TJ-> Ublx: if the system is using LVM then you'll get the LVM logical volumes as well as any LUKS devices
[01:08] <Ublx> 4 devices
[01:09] <TJ-> Ublx: so 2 crypto and 2 logical volumes? can you deduce that from the device names?
[01:09] <Ublx> I have root, swap, sda5_crypt (what seems to lead to the first two) and luks-...
[01:09] <TJ-> OK, that sounds correct and understandable
[01:10] <Ublx> puh
[01:10] <TJ-> so the confusion still remaining is the error report mentioned sdc3 but you said there is no sdc* reported (does 'lsblk' not show an sdc ? )
[01:11] <TJ-> Ublx: did the error come from a GUI program, such as file manager?
[01:11] <Ublx> no, should I plug the drive in and then check lsblk?
[01:11] <Ublx> TJ-: yes, from nautilus
[01:12] <Ublx> i do almost anything from cli but not mounting.
[01:13] <TJ-> Ublx: OK, so the file manager uses udisks under the hood to control devices, so my suspicion is it is trying to read the missing device and getting confused
[01:13] <Ublx> ok
[01:13] <Ublx> thank you so much for helping me.
[01:14] <TJ-> Ublx: see if this reports anything: "udisksctl info -b /dev/mapper/luks-e05b2f...." (complete that UUID again!)
[01:14] <TJ-> Ublx: udisksctl is the command-line interface to udisks
[01:16] <Ublx> output looks (too) fine
[01:16] <TJ-> Ublx: does it give a Mountpoint ?
[01:17] <TJ-> Ublx: presumably something like /media/$USER/luks-$UUID
[01:17] <Ublx> /dev/dm-0 ?
[01:17] <TJ-> Ublx: no, that'll be the device. Towards the end of the report
[01:17] <Ublx> no /dev/dm-3
[01:18] <TJ-> Ublx: e.g. for one of my LVs I see " MountPoints:        /media/tj/SourceCode "
[01:18] <Ublx> /dev/mapper/luks-e05b2 ... and /dev/disk/by-id/dm ... nothing of the media/ folder
[01:18] <TJ-> Ublx: if there is an active mountpoint we need to first tell udisks to unmount it
[01:19] <TJ-> Ublx: good
[01:19] <Ublx> can't we just try to unmount it?
[01:19] <TJ-> Ublx: so now we just need to be sure what device that lives on!!
[01:19] <Ublx> what, if it's my local disk?
[01:19] <TJ-> Ublx: it isn't mounted
[01:19] <Ublx> ah
[01:20] <Ublx> I see.
[01:20] <TJ-> Ublx: all this effort is to try and avoid removing a device that is really in use! and so far we've not proved that one way or the other
[01:20] <Ublx> ok
[01:21] <Ublx> tried udisksctl info -b /dev/mapper/sda5_crypt and I also do not get any mount point.
[01:21] <TJ-> Ublx: Going back to the "sudo dmsetup info luks-..." did you notice (or check again) the "Open count:" -- that needs to be 0 for us to be able to remove it
[01:21] <TJ-> Ublx: it is possible sda5_crypt is swap. Check "cat /proc/swaps"
[01:22] <Ublx> open count is 1
[01:22] <TJ-> Ublx: drat! can't remove it unless that is 0. '1' means something has an open handle on that device
[01:22] <Ublx> swapfile, yes
[01:23] <TJ-> Ublx: can you show me "pastebinit <( sudo blkid )"
[01:25] <Ublx> http://paste.ubuntu.com/p/Qp2pZkzqk8/
[01:25] <Ublx> had to install pastebinit
[01:25] <Ublx> but cool, didn't know it's possible within the cli
[01:27] <TJ-> Ublx: OK, definitely no sign of anything there related to this phantom luks-e05bf2...
[01:28] <TJ-> Ublx: looks like you've been eaten by snaps though :)
[01:28] <Ublx> ugh
[01:28] <Ublx> i am nervous
[01:31] <calamari> Can anyone tell me the name of the repo that contains arm64 packages? It seems like the normal repos only have i386 and amd64.
[01:31] <TJ-> Ublx: let's see if you can 'close' the encryption. try "sudo cryptsetup close luks-e05bf2..." (complete UUID again!)
[01:31] <TJ-> Ublx: if the command doesn't report an error check if it has gone with "ls /dev/mapper/"
[01:33] <Ublx> Eh ... I just went to all the pids of ps aux | grep USERNAME and I found a running process on this (no more connected) partition ...
[01:33] <Ublx> ... killed it ... and it works ...
[01:33] <Ublx> so sorry
[01:33] <Ublx> but i didn't thought about it
[01:34] <Ublx> i just wanted to be sure that i close everything if we'll lock out the local disk.
[01:34] <TJ-> Ublx: so you've been able to remove the /dev/mapper/ node for it now?
[01:36] <Ublx> hm... where can i see it
[01:36] <Ublx> ah
[01:36] <Ublx> the 3 partitions are now connected and running ... 2 times luks... but it works.
[01:38] <Ublx> I have to note down all your commands. Learned so much. Thank you!!
[01:38] <calamari> I found it in ports.ubuntu.com. Can anyone tell me how to properly add that repo to sources.list? The normal way doesn't seem to be doing it.
[01:40] <TJ-> calamari: what is the 'normal way' that doesn't work?
[01:40] <calamari> TJ-:  deb [arch=arm64] http://ports.ubuntu.com/ubuntu/ bionic main restricted universe multiverse
[01:40] <oerheks1> sudo dpkg --add-architecture arm64
[01:41] <calamari> TJ-:  The repository 'http://ports.ubuntu.com/ubuntu bionic Release' does not have a Release file.
[01:41] <calamari> oerheks1: already did that
[01:41] <TJ-> calamari: there is no "/ubuntu" directory; delete that.
[01:42] <TJ-> calamari: as in "deb [arch=arm64] http://ports.ubuntu.com/ bionic main restricted universe multiverse "
[01:42] <calamari> TJ-: Thank you!
[01:42] <TJ-> calamari: that URL points to where the dists/ and pool/ directories are
[01:42] <TJ-> calamari: see http://ports.ubuntu.com/
[01:43] <calamari> TJ-: I'm in business, thanks
[01:56] <Ublx> Thanks again for your time, TJ-. I noted all down. Have a good week!
[02:33] <blahboybaz> Does 20.04 use unity or something else?
[02:36] <quadrathoch2> blahboybaz: it does use gnome3, but unity is installable
[03:47] <bitfawkes> Hello, I've an ubuntu server with two nvme disks of 1TB each one, as I don't need Raid I'm trying to set a logical volume of 2TB extended over both disks (I would like to use both nvme as a single partition), there's a solution for that? thanks to everyone spends some time to help me :)
[03:49] <bitfawkes> the release II would like to use on is 18.04 LTS.
[03:51] <guiverc> bitfawkes, LVM comes to mind (can deal with many disks etc); https://wiki.ubuntu.com/Lvm
[03:53] <bitfawkes> I did read that but I'm not able to execute it without support... I'm willing to reward in cryptocurrency for that help...
=== Aleksander is now known as alxd
=== BrianG61UK__ is now known as BrianG61UK_
[06:24] <ducasse> blahboybaz: one filesystem spanning two disks can be a bad idea as the failure of one disk will make you lose all the data
[06:33] <ducasse> blahboybaz: sorry, was not meant for you
=== PowerTower_121 is now known as PowerTower_120
[07:44] <tatertots> +
=== michagogo_ is now known as michagogo
[08:53] <zamba> anyone got any idea about this kernel panic: https://dpaste.org/Nq4V ?
[09:00] <devid> how can I best test the expected battery life, by live boot or by VM ?
[09:15] <dandersson> A mirror returned via mirrors.ubuntu.com/mirrors.txt for me (location: Sweden, mirror: http://se.mirror.guru/ubuntu/) seems to not have renewed its domain name, and is thus dead. Is there somewhere to report this to get it removed, or will it automatically go away?
[09:19] <lotuspsychje> dandersson: there is #ubuntu-mirrors to report repo issues
[09:20] <dandersson> lotuspsychje: Awesome, thanks!
=== ledeni_ is now known as ledeni
[10:03] <alf1975> Is there a easy way to change ubuntu kernel ?
[10:03] <alf1975> some are more effective and speedy
[10:04] <alf1975> Ukuu a util i found is cheat
[10:07] <lotuspsychje> alf1975: whats your end goal exactly?
[10:10] <guiverc> alf1975, LTS releases have two stack options; GA kernel & stack or the HWE kernel & stack
=== ace_me1 is now known as ace_me
[10:29] <aleek> hi!:D I just accidentaly deleted /boot and /dev/ on our ubuntu 20.04 :D MBR/UEFI is fine. How to recover?:D
[10:30] <aleek> I don't have root privileges anymore, because /dev/null is missing :D
[10:30] <aleek> so first step of course is live usb :)
[10:37] <alf1975> and if the Nvidia GPU with the onboard Intel GPU as coprocessor ?
[10:37] <alf1975> more speed
[10:38] <alf1975> yes brothers, processors witch could give more speed ... in games
[10:41] <lotuspsychje> !discuss | alf1975
[10:41] <ubottu> alf1975: Want to talk about Ubuntu, but don't have a support question? /join #ubuntu-discuss for non-support Ubuntu discussion, or try #ubuntu-offtopic for general chat. Thanks!
[10:47] <linuxr> hello all...I noticed that the "atril" pdf viewer creates two "webkit" subprocess for each document I'm viewing...why is that?
=== ace_me^^ is now known as ace_me
=== EriC^ is now known as EriC^^
[11:27] <alf1975> [12:25] <alf1975> Lubuntu 19.04
[11:27] <alf1975> [12:25] <alf1975> Lubuntu 19.04
[11:27] <alf1975> [12:25] <alf1975> This standard release was made on schedule on 18 April 2019.[149]
[11:27] <alf1975> [12:25] <alf1975> This release marked the first Lubuntu version without 32-bit support. Lubuntu developer Simon Quigley wrote in December 2018:[150]
[11:27] <alf1975> [12:25] <alf1975> source wiki
[11:27] <alf1975> [12:25] <alf1975> it worked very well on a Pentium IV with 19 years old
[11:33] <cuttiepie> Hi
=== thafaker5 is now known as thafaker
[11:48] <metbsd> my laptop fail to wake up from suspend
[11:48] <metbsd> screen is black
[11:58] <BluesKaj> 'Morning folks
=== mateen1 is now known as mateen
[13:56] <raub> Where do I need to config bash/terminal/whatever so when I do "cat /etc/default/grub" the double quotes inside the file are printed like normal ASCII characters instead of something fancy that breaks when I try to paste to a text document?
[14:20] <J_Darnley> Is there a dkms package that back-ports a fix for the --source-only option not working on the 18.04 package?
[14:24] <jeremy31> What source are you after?
[14:26] <J_Darnley> My source (specifically a netmap diver) that I have in front of me that I wish to package into a deb using `dkms mkdeb`
[14:32] <karstensrage> need a favor
[14:32] <karstensrage> oopos
[14:32] <karstensrage> sorry
[14:33] <J_Darnley> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=832558
[14:33] <ubottu> Debian bug 832558 in dkms "dkms: mkdeb --source-only fails: can't find package" [Normal,Fixed]
[14:34] <J_Darnley> Typical debian nonsense about deviating from upstream and patching the hell out of their old crap
[14:35] <J_Darnley> Or how about a more generic question:  Where can I search for back ported packages?
[14:45] <J_Darnley> Is this it?  https://packages.ubuntu.com/bionic-backports/allpackages
=== MIF is now known as bTal
=== bTal is now known as MIF
[15:23] <Furai> Is there any world clock shell extension that let's you put next to clock a custom name?
[15:23] <Furai> I'd love to be able to put my coworkers' names next to the timezones.
=== ace_me1 is now known as ace_me
=== jgrim93 is now known as jgrim9
=== Moyst_ is now known as Moyst
=== jgrim95 is now known as jgrim9
=== ravage is now known as Ravage
[16:35] <vmguy23m> Where can I find touchpad drivers (Acer Aspire 3, right-click does not work, only left click)
[16:55] <Maik> vmguy23m: try installing synaptics
[16:55] <Maik> sudo apt install xserver-xorg-input-synaptics
[16:56] <Maik> reboot after that and see if it works
[17:02] <vmguy23m> Thanks
[17:03] <Maik> np, yw
[17:12] <alf1975> Problems with q4wine, Wine ? Windows Vírus attack Ubuntu ?
[17:12] <alf1975> for example a Windows game with virus attacks other games ? in Wine ?
[17:13] <vmguy23m> alf1975: What do you mean?
[17:20] <vincenzoml> Hi there, I am going to mount a new 10tb hard drive on a computation pc running ubuntu
[17:20] <vincenzoml> is it adviceable to use btrfs on it to benefit from snapshots (I already use snapper for smaller drives)? Or should I avoid that?
[17:25] <Maik> vmguy23m: don't min alf1975 too much, he's just posting random stuff in various channels
[17:25] <Maik> i pointed him to discuss and offtopic but he won't listen
[17:26] <Maik> min/mind
[17:26] <vmguy23m> I just don't understand what the question is
[17:26] <Maik> yeah, i don't get him either with most of the things he posts
[17:34] <ducasse> vincenzoml: i don't trust btrfs and use zfs instead
=== ijohnson is now known as ijohnson|lunch
[17:43] <vincenzoml> ducasse: yes I know zfs is way more established; but don't have the expertise to just start using that on a 10tb drive that K people will fill up.
[17:44] <vincenzoml> I just wonder if there are known pitfalls of btrfs specifically on large drives.
[17:45] <cbreak> I haven't heard of anything that makes btrfs problematic with larger disks more than it does with smaller ones
[17:45] <cbreak> but ... since it is problematic, I wouldn't risk it either
[17:46] <cbreak> I use zfs with success though. It's nice :)
[17:46] <vincenzoml> cbreak, I am not an expert; but I have used btrfs for like 3-4 years on three machines with no significant problems, except when I had a failing RAM block
[17:46] <vincenzoml> but cbreak ducasse I am curious to hear:
[17:47] <vincenzoml> since this is a production machine, what are the risks you think should make me think again and use zfs instead
[17:47] <cbreak> do you have redundancy?
[17:47] <vincenzoml> and about zfs is it easy to add a second drive to zfs and make a raid without reformatting the first drive?
[17:48] <vincenzoml> cbreak, I have only used mirroring
[17:48] <vincenzoml> with two drives indeed
[17:48] <cbreak> with zfs you can add more disks to make it a mirror
[17:48] <cbreak> you won't get space by doing that obviously
[17:48] <vincenzoml> cbreak, ok great, that's important (and a reason why in the first place btrfs was ok to me)
[17:48] <cbreak> you can also add more vdevs to get more space (via "striping" (not real striping)), but that won't give you redundancy
[17:49] <cbreak> but you can not create a raidzX vdev from an existing vdev
[17:49] <vincenzoml> so I cannot first extend and then mirror, that's what you mean (I'm not an expert in filesystems and RAID, just need to make that machine large and safe)
[17:50] <cbreak> you can add new vdevs
[17:50] <cbreak> you can change single disk vdevs into n*mirror vdevs or back
[17:50] <cbreak> you can not change anything to a raidzX vdev
[17:51] <cbreak> you can get more space either by replacing harddisks in an existing vdev with bigger ones
[17:51] <cbreak> or by adding new vdevs
[17:51] <cbreak> zfs is more restrictive with this than btrfs
[17:51] <cbreak> but the benefit is that zfs doesn't shredder itself when you try to use those features :)
[17:51] <vincenzoml> sorry but what is raidzX in the first place :) Anyway in that machine I doubt we will ever have more than 4 10tb disks, and likely no more than 2.
[17:52] <vincenzoml> and if 2, they will be mirroring
[17:52] <cbreak> a raidzX (raidz1 / raidz2 / raidz3) is a parity-raid with 1, 2 or 3 disks worth of redundancy
[17:52] <vincenzoml> that's the most likely scenario
[17:52] <cbreak> similar to raid5, raid6
[17:52] <vincenzoml> ah ok
[17:52] <cbreak> raidz1 doesn't make sense with less than 3 disks
[17:53] <vincenzoml> so the most likely scenario is that: 1) tomorrow I mount the drive, format it, and let it live happy for some months
[17:53] <cbreak> zfs isn't 100% self explanatory, and as with any powerful tool, you can destroy your data if you tell it to
[17:53] <vincenzoml> 2) then I buy a second drive, mount it, and it has to become a mirror, raid1-alike
[17:54] <cbreak> what you'd do is probably: create a pool on your drive
[17:54] <cbreak> create datasets on that pool (similar to thin-partitioned file systems)
[17:54] <cbreak> set up snapshot scheduling for those
[17:54] <cbreak> you can later add more redundancy or more striping to that pool
[17:54] <cbreak> without affecting any data on it
[17:55] <vincenzoml> ok, I could experiment with that a bit.
[17:55] <vincenzoml> Returning to btrfs what would be your main concern against using it?
[17:55] <cbreak> you can enable encryption, compression, ... on individual datasets
[17:55] <vincenzoml> I mean, so far, with RAID1 it behaved nicely for me.
[17:55] <vincenzoml> Ah ok
[17:55] <vincenzoml> so I can understand what a dataset is
[17:55] <vincenzoml> do dataset share the free space?
[17:55] <cbreak> you can look at snapshots via ls someDataset/.zfs/snapshots/...
[17:55] <vincenzoml> or are they fixed-size
[17:56] <cbreak> vincenzoml: they share the pool's available space
[17:56] <cbreak> but you can reserve space for datasets, or set max usage quotas
[17:56] <vincenzoml> Ok, So I could start with 1 dataset and create more later
[17:56] <vincenzoml> interesting!
[17:56] <cbreak> you should create datasets in some logical fashion
[17:56] <vincenzoml> so I could for instance, create 1 dataset; when a particular application starts filling up the whole drive, I can confine it to a newly created dataset with a quota, righ?
[17:57] <cbreak> maybe pool/media, pool/vms, pool/media/linux_isos, pool/media/blender_cloud_movies, ...
[17:57] <cbreak> vincenzoml: datasets are independent filesystems
[17:57] <cbreak> so if you want to do that, you have to copy the data over to the new dataset
[17:57] <vincenzoml> I see
[17:58] <cbreak> you can set mountpoints for those datasets individually
[17:58] <cbreak> but normally, they are mounted in the parent
[17:58] <cbreak> so if your pool is in /mnt/pool, then there's some /mnt/pool/media/linux_isos
[17:58] <vincenzoml> But consider the following situation: Just 1 dataset, filling up 98% of the hard drive; then I create a new dataset, and move data from the old one to the new one. Will this work or get stuck?
[17:58] <cbreak> you should really avoid filling zfs to more than 90%
[17:59] <vincenzoml> cbreak, but sh*t happens :)
[17:59] <cbreak> if you want to "move" data between datasets, you have to copy it
[17:59] <vincenzoml> cbreak, ah yes so with very large files that could become a problem
[18:00] <vincenzoml> but a dataset can be created in the pool at any time right?
[18:00] <cbreak> if you have space for it
[18:00] <cbreak> they take up a few kb in meta data
[18:00] <vincenzoml> then a question: with btrfs I have filled up the drive many times due to snapshots preserving too much old data. It worked to just remove the snapshots (even if it's a though decision).
[18:00] <vincenzoml> With ZFS you said, that is not advisable?
[18:01] <cbreak> it's not advisable, but possible
[18:01] <cbreak> (it's not advisable with btrfs either)
[18:01] <vincenzoml> why is it not advisable if I may ask?
[18:01] <vincenzoml> with btrfs they say it's not adivsable but I never knew why
[18:01] <cbreak> because zfs switches to a slower but move efficient data allocation strategy when the pool gets full
[18:01] <cbreak> and when you completely fill a pool, it doesn't have enough space to delete anything
[18:02] <vincenzoml> ah yes
[18:02] <vincenzoml> that was also the reason with btrfs
[18:02] <cbreak> (zfs tries hard to not let it get that far)
[18:02] <vincenzoml> now I remember; but nowadays I believe that's "fixed" in btrfs (probably by reserving space)
[18:02] <vincenzoml> so again, what is in btrfs that you consider more dangerous than zfs?
[18:02] <vincenzoml> I know a lot of people wouldn't use it
[18:03] <vincenzoml> BTW I am happy to have learned a bit about zfs
[18:03] <vincenzoml> I may as well decide to use it tomorrow
[18:05] <cbreak> it's had some issues with redundancy, and self-destructive behavior
[18:06] <cbreak> not sure how current this information is, but last I read that they disabled all raid modes other than mirroring
[18:06] <vincenzoml> and how would I be absolutely sure that zfs does not fill up to the point I can't delete data anymore?
[18:06] <vincenzoml> Only by setting individual quotas on datasets, right?
[18:07] <cbreak> well, or a quota on the root dataset
[18:07] <cbreak> but as I said, you shouldn't even let it get to 90% for performance reasons
[18:07] <cbreak> zfs will try to not get too full for deletion on its own
[18:09] <Perdellian> !compose
[18:10] <Perdellian> what's the current official/recommended way to enable/map the Compose Key in recent versions of non-kde Ubuntu?
[18:11] <Perdellian> I found https://help.ubuntu.com/community/ComposeKey but I also found stuff suggesting that this is out of date
[18:13] <Perdellian> (context: I'm a dyed-in-the-wool KDE user, but I'm going a little bit overboard in a Duolingo comment thread that asked how to type "ß", and I want to be sure I'm giving correct information)
[18:25] <Captain_Proton> is there a setting in SSSD that i can map domain name to pre-2000 domain name
[18:26] <Captain_Proton> the full domian name is Ilikecheeseverymuch.local > per-2000 is ilikecheese
[18:59] <Maik> alf1975: about your questions earlier. If you don't have a Ubuntu support question that needs attention but want to discuss other thing that are Ubuntu related then please join #ubuntu-discuss the next time as we asked a couple of times before now. Thanks in advance.