[00:58] seriously, eol upgradse are not this difficult [01:03] !cookie | nacc [01:03] nacc: Wow! You're such a great helper, you deserve a cookie! [01:03] Bashing-om: lol [01:03] i've been in a bad mood all day with these users! :) [01:03] oh well, also fixing this nasty packaging bug hasn't helped :) [01:03] nacc: My patience would have expired long past with this one :) [01:04] seriously, it's like they both forgot how to read! :) [01:05] nacc: Addeled brained now and can no longer read for comprehension - reckon ? [01:06] Bashing-om: anything is possible :) [01:10] nacc: Settled down and ya got a "thank you " :) [05:27] good morning to all [05:27] hello [05:27] hey Sveta [05:28] Sveta: all good today? [05:29] i burnt my feet by walking on a melting rock yesterday, but it doesn't hurt much, just sensitive to heat, so no walking for me this weekend [05:29] apart from that everything is fine [05:29] wow [05:29] Sveta: walk too long? [05:30] i walked for one hour, but it was a very hot rock [05:30] ah i see [05:53] Time ! .. g nite guys \o [05:53] nite [05:53] :-) [05:54] hey oerheks [05:54] lotuspsychje: oerheks .. all --- take care till we meet again :) [06:28] bbl shower [06:30] smell ya [06:52] Hi folks 0/ [06:52] hi [06:53] i'm a bit frustrated by how many DE-specific questions get asked in ubuntu channels, i can't possibly know the buttons in them all and often they're not fully documented [06:55] Mmm [06:55] It's kind of like shooting into the aether to ask them [06:56] But like, how do people even know if the flavour of ubuntu they have has a custom build of the DE? [06:57] So, it kind of makes sense to ask IMO. Expecting an answer is another matter [06:58] hey many people download ubuntu and do not even know which flavour they are using [06:59] Good morning [07:01] That's very true. When I first started using linux I managed to install both Ubuntu and Kubuntu on the same drive without formatting in between [07:01] Morning! [07:01] (my laptop was not a happy camper after that) [07:02] Ubuntu and Kubuntu have the same core. If the problem is unrelated to the DE it shouldn't really matter which flavour they are running. If they don't know which version... that is something else. [07:04] In my case the problem was that it corrupted files and led to competiting services etc. Not because they are incompatible but because files from one install were left behind and the second didn't have proper control over them [07:05] (e.g. a bunch of files in /etc/usr/share that weren't registered into apt) [07:06] I was too much of a nub to understand at the time [07:06] that's ok [07:07] It was also like [07:07] not knowing the names of software that you are using is a bit more worrying to me [07:07] makes it a pain to troubleshoot [07:07] 9 years ago, give or take [07:07] Yeah true. [07:08] Sveta: Sure, but if you are completely new to the game it ain't surprising. Suddenly you are bombarded with terms you never heared of. [07:11] after installing a linux distro the users need to be gradually exposed to some key concepts [07:11] we could suggest them to read one small knowledge base article a day if they like [07:11] little load, lots of benefits [07:12] okay I've updated xorg.. now lets see if my damn xorg settings stick this time [07:12] brb [07:14] Sveta: Good idea. Unfortunately people are lazy 😞 [07:14] they're less lazy than you think [07:15] hi all [07:16] Hey immu [07:16] hye lordievader whats up [07:16] Doing allright here, how are you? [07:26] i am going doing maintenace in Windows [07:26] immu: Good luck 😉 [07:28] :) [07:30] still need to keep Windows for various reasons [07:30] Same here. [07:35] On Windows illegal kernel calls where causing blue screens [07:35] by antovirus vendors [07:36] were [07:37] Well, wouldn't be the first time Mircosoft re-added stuff to the system to make antivirus software work.... (suposing these are deprecated calls) [07:38] Okay back, I kind of got it working? [07:38] Good for you. Whatever it was you got working. Kind of. [07:39] good morning, everyone [07:39] ^__^ [07:39] > // ACHTUNG!!! this is a special hack for IBM antivirus software [07:39] immu above quote comes from http://atdt.freeshell.org/k5/story_2004_2_15_71552_7795.html [07:39] Lol. Basically my monitors are default-detected in the wrong positions and my xorg.conf kept getting reset at login for some reason. [07:39] Hey ducasse [07:39] How are you? [07:40] evening, ducasse [07:40] hi lordievader - all good, thanks. and you? [07:40] Still dunno what's messing with xorg, but running xrandr after login now works in 17.10 (it was broken for some reason in 16.04) [07:40] Doing good here 😁 [07:44] thats 2004 stuff @lordievader [07:44] I know [07:54] oops back xD [07:58] https://www.bleepingcomputer.com/news/microsoft/microsoft-says-no-more-windows-security-updates-unless-avs-set-a-registry-key/ [07:58] see lordievader [07:59] immu: 404 ? [07:59] Nope, it's me. Last slash is important (God knows why). [07:59] nope works for me [08:00] https://www.bleepingcomputer.com/news/microsoft/microsoft-says-no-more-windows-security-updates-unless-avs-set-a-registry-key [08:00] check now [08:00] Because some anti-virus vendors are using very questionable techniques they end up cause systems to ‘blue screen of death’ — aka get into reboot loops. === marcoscosmos_ is now known as marcoscosmos [08:14] immu: That last slash makes a difference, apparently. [08:14] ok [08:15] brb === kostkon_ is now known as kostkon [09:47] morning all [10:05] back [10:14] 👋 [10:55] gnome is fucking covoluted wtf [10:57] I bricked ubuntu by adding .desktop and .session files in /usr/share/xsessions and /usr/share/gnome-sessions apparently [10:57] -_- [10:57] realistically it was probably some kind of driver failure I suppose [10:58] I am trying to get away from Xenial's LVM and failing every time because I make some mistake that corrupts my filesystem. [10:58] But I'm not supported if I "import" a version from Yakkety or beyond, and the end result is every time that I can throw away my filesystem. [10:59] That is to say, making "dd" copies of a disk is extremely dangerous. [10:59] yeah I'm doing some unsupported things too but, I can't concieve how they'd be related [10:59] Because Xenial's LVM has no protection against having multiple disks with the same UUID [10:59] dangerous how? [10:59] oh fuck [11:00] ok [11:00] You have to carefully shield the original disk using "global_filter" if you want to examine any of the disks while the other is still connected [11:01] I see [11:01] If you make one mistake, it can start writing to the wrong disk [11:01] To make matters worse I had a cache ;-) that I had forgotten to disconnect [11:01] RIP [11:02] The problem was mostly that LVM depends on /etc/lvm/lvm.conf so in a chroot environment, this file is different [11:02] So if you correctly shield in your live environment [11:02] This correct shielding is gone once you enter a chroot [11:02] Right [11:03] Then, if you somehow forget to deactivate the volumes before the shielding and reactivate them after the shielding [11:03] is it not possible to force the uuid to update/become unique or is it a problem that triggers mid dd? [11:03] I mean if you don't do vgchange -an in between, and then vgchange -ay in the new situation.... [11:04] Well you can [11:04] But if you do that the names and UUIDs of the backup change [11:04] If you then want to copy on to a new disk [11:04] You have to do it again [11:04] I wanted to avoid that [11:04] Having to do it twice [11:05] Because I also had another disk connected that I had not copied (the cache) [11:05] So if I change the UUIDs in the backup, I first have to disconnect it entirely from the cache [11:05] I see [11:05] Because otherwise the names of the cache also change. [11:05] So I have to remove a physical volume from the backup [11:05] Wait why's there a cache anyway? [11:06] set [11:06] Well I should have disconnected it I just forgot [11:06] wait by cache do you mean swap? [11:06] or is this some kind of raid scenario? [11:06] I just have an LVM cache configured with an SSD for the HDDs. [11:06] Sortof. [11:07] Yeah so I made matters worse myself I guess. [11:07] However disconnecting the cache from the backup means doing some LVM manipulation [11:07] that will fail if you don't properly shield first ;-). [11:07] So it's a chicken and egg problem at that point. [11:08] oh cache like hybrid-ssd-hdds [11:08] So I had to disconnect prior to making the backup which I didn't do. [11:08] gotcha [11:08] Yeah [11:08] That [11:08] But [11:08] If I disconnect prior to making the backup [11:08] And remove the extra physical volume [11:08] Then adding it back later is going to be difficult. [11:08] So it's just one big mess anyway. [11:08] why? [11:09] Cause if I really remove the PV I also have to remove the volumes. [11:09] Which means I have to make the entire cache structure again. [11:09] So the only way to do it is to also dd the cache disk [11:09] Which requires an extra disk for that too [11:10] Then I can change them both at the same time and there are no issues. [11:10] shouldn't the volume-making be like. a script that runs at boot anyway? [11:10] boot times wouldn't be so great but [11:10] yknow [11:10] Well no because they are persistent volumes. [11:10] in my experience having a hdd ni the fstab fucks boot anyway [11:10] Ie. a cache has to persist. [11:11] why would a cache persist? [11:12] Because it has to prime which data is often used so you can have fast boot times etc. [11:12] Ah [11:13] E.g. the ureadahead thing [11:13] Took about 5.5 seconds on my raid 0 system (using regular 5400 disks at this point) [11:13] And with the cache connected it went down to 0.7 seconds. [11:14] So that's a direct 5 second boot speedup. [11:14] Anyway [11:15] It's just one big giant mess. [11:15] The way to do it then is to first make the dd, then properly shield the source disks and all the PVs [11:15] Then in the shielded thing disconnect everything you don't want from the backup [11:16] But I didn't want that because I wanted to copy the backup on ;-). [11:16] oh nice [11:16] but [11:16] I'm guessing you're doing this all as like [11:16] a script so [11:16] you don't have to deal with typoes every time [11:17] Well no it is too complicated to write a script ahead of time I guess... [11:17] I mean the first time... [11:17] I don't worry about typos but... [11:17] no yeah I meant like [11:17] for future use [11:17] I still don't know how to do it at this point. [11:18] Ah [11:18] It's not just the cache, this would happen every time you had more than 1 PV. [11:18] Well don't worry too much [11:18] err [11:18] don't feel bad [11:18] Yeah I just had an unbootable system because of a billion dangling inodes [11:18] commercial backup software is often utterly useless so [11:19] I was just scanning what all the lost+found files had come from [11:19] And I still need to find out how to do this. [11:19] I guess the best way is disconnect any cache to be sure. [11:19] I figure the solution most people end up doing is using sync or some such to like, perpetually update key files rather than whole drives [11:20] Then shield the source disks [11:20] Exchange for target disks. [11:20] Copy your own disk back to the target disk. [11:20] Never touch the intermediate. [11:20] Oh boy. [11:21] Never mind, I will just do it slightly more correctly next time I guess. [11:21] This is sorta the 3rd time this has happened to me at least. [11:21] I see [11:22] seems like an ineffectual way to do backups if you lose your original during xD [11:22] Ugh I'm stuck on windows for the night. [11:22] The LVM of Xenial is considered unsupported by the LVM team because it is too old. [11:22] They always tell me: UPGRADE [11:22] Windows and it's stacking window manager and mouse use mandates [11:22] I hate mice [11:22] oh wow okay [11:23] But if I try to upgrade through Ubuntu, Ubuntu people tell me: we can't support that. [11:23] of course not [11:23] So I'm stuck [11:23] is this a commercial grade deal? [11:23] or like [11:23] for private use? [11:23] With an "I will kill your system" version of LVM [11:23] Private use that this point [11:23] Why? [11:24] because depending on the scenario you may not need support for your entire system [11:24] If you happened to do this for a company and you used Xenial LVM you'd be in trouble [11:24] I use unsupported content and still get support for the rest [11:25] e.g. an external IDE [11:25] Well generally I don't require to the detail support [11:25] Just a few pointers. [11:25] Like they wouldn't tell me how to debug initramfs [11:25] Because I was using part of Yakkety [11:25] I'm usually coming here looking for info that's a PITA to google (after trying for a long time) [11:26] When the problem was nothing to do with Yakkety [11:26] like [11:26] how the fuck gnome session even launches these days [11:26] But anyway I have to continue wasting my life away [11:26] I spent like 4 hours trying to figure out how to launch the settings daemons and shit [11:27] only to inexplicably brick it [11:27] I couldn't figure out how to start the KDE Power Savings settings from the command line [11:27] Then I forgot I hadn't disabled it and my system went into standby while doing a download [11:28] Which kinda ruined the entire next day for me [11:28] :\ [11:28] yeah [11:29] Anyway. [11:29] I really now have to return to a clean Xenial and start everything all over again. [11:29] Which I can't do because I'm also on a slow link [11:30] So yes indeed [11:30] Every time I make a backup it destroys my origin AND the backup [11:31] LVM 171 has some protections against more errors [11:31] But the latest version of LVM in Ubuntu is about 168 [11:31] Or maybe now [11:32] Why not just use the latest lvm with xenial? [11:32] You mean compile from source? [11:32] I know it's not difficult to compile it [11:32] No no [11:33] apt-pinning [11:33] Bionic has version 168 which is too old [11:33] But at least new enough to protect again the dual UUID issues [11:33] well you should also be able to pin from debian itself [11:33] not just ubuntu [11:33] including debian-testing [11:33] zomaar: we'll be syncing lvm2 from debian shortly [11:33] But that means using a newer version from Ubuntu [11:34] not with pinning [11:34] Then when there is some issue they'll say I am unsupported [11:34] Okay [11:34] pinning is unsupported but it can't be worse than your current situation [11:35] Bug #1741986 [11:35] bug 1741986 in lvm2 (Ubuntu) "Please merge lvm2 from Debian unstable for lvmlockd and sanlock support" [Wishlist,Confirmed] https://launchpad.net/bugs/1741986 [11:35] Well for some reason I had boot issues with the new LVM from Yakkety [11:35] only lvm and other packages from whatever you pin if any will be unsupported [11:35] Sure but [11:35] Those weren' there yet when coming from kernel 4.4.0 [11:36] I was always able to use the newer LVM in Xenial [11:36] I hadn't figured out yet how to troubleshoot that [11:36] Now I know [11:37] Well no because I didn't even get an initramfs prompt [11:37] I am currently sitting in Yakkety kernel ;-) (4.8.0) with Yakkety LVM [11:38] It has some issues with my USB3 and the Nouveau turns my screen off if the monitor is not connected [11:38] but only the HDMI-monitor [11:38] strange [11:38] I have yet to test the 4.10 kernel now [11:39] And 4.4 I guess [11:39] My download link is also so slow that it takes a few hours to download a new kernel [11:39] So I salvaged them from installer media ;-) [11:40] I'm just dpkg-repacking everything I need [11:41] I grabbed 4.4.0 from the original Xenial DVD [11:41] Newer udev gives issues with my 3G modem [11:42] Through its rules; not the daemon itself [11:42] So I also have to downgrade udev to the version from Xenial [11:43] Which I also grabbed from an installer medium ;-). [11:43] Because it takes too long to constantly update sources ;-). [11:43] Yeah a shitload of trouble isn't it. [11:45] haha yeah [11:45] I am just afraid of compiling LVM myself. [11:46] I first just wanted to try prepackaged versions, because it can't really be worse than that. [11:46] (How does one replace the running LVM if it is depended on by initramfs-tools and such) [11:46] (While not messing up the package system) [11:47] well [11:47] tbh you get used to it [11:47] and its usually just [11:47] make && sudo make install [11:47] Sure but you want to put it in /usr/local [11:48] Because initramfs-tools is going to copy it from /sbin [11:48] But [11:48] So that's more trouble [11:48] the location is usually determined by the make configure right? [11:48] if you're luck you can run ./configure or such and there's an option to redirect it [11:49] Sure but you don't want to overwrite the /sbin files and I would have to check whether any initramfs-hooks get updated etc. [11:49] I guess that I will mostly be safe with the OS patches for spectre........a bad bios upgrade can brick your computer..... [11:50] Just more trouble I guess [11:50] yeah idk anything about initramfs [11:50] In fact some programs always install in /sbin [11:50] zomaar: I missed why you're wanting to maintain your own lvm2, care to summarise? [11:51] No I don't want to, I just want to use a later Ubuntu package [11:51] I was just saying how much trouble it is if I don't [11:51] zomaar: right, but due to what issue? [11:51] LVM Xenial has no protection against duplicate UUIDs and handles that very badly. [11:51] They introduced better handling somwhere in 162 or so [11:52] That's the most important one [11:52] It can basically replace the running PV of an activated VG without you noticing [11:52] runtime [11:53] Because some disk you connect is "preferred" due to its name or something [11:53] This then completely destroys your filesystem in the sense of corrupting all open files and their inodes [11:53] Version 162 will not randomly replace PVs like that [11:54] So e.g. if you were able to connect some USB stick to Xenial with the same UUID as what's running on your system [11:54] You are in for some trouble [11:54] You can basically destroy someone's system once you know the UUID [11:54] Right, but it begs the question of why you'd have multiple PVs with the same UUID /online/ at the same time and not configure lvm.conf to ignore one of the devices [11:54] Well [11:55] That's nice if you can do that correctly [11:55] But chrooted LVM also has a lvm.conf and can mess it up again [11:55] I mean [11:55] This dependency on a file is rather bad [11:56] The first time I didn't know to do that. [11:56] The whole point of a UUID is in it's name: Universally /Unique/ IDentifier [11:56] A user will just use dd and get into a world of trouble [11:57] Then when you discover how to shield you can still mess up [11:57] In this case I messed up because [11:57] so if it is no longer unique ... expect problems especially when you're doing premediated non-standard operations and configurations [11:57] The live system had automatically activated the volume groups of course [11:58] When I shielded the origin they were still activated [11:58] And still used [11:58] Etc. [11:58] dd is actually supported operation by LVM [11:59] clear [11:59] by doing vgimportclone [11:59] but you have to shield first correctly [12:00] And do everything in the right order [12:00] I mean they rewrote LVM to protect against this use case [12:00] Recognising the trouble [12:01] I do data forensics and recovery which involves a lot of cloning, sometimes with LVM, and (so far) I've managed not to have LVM get confused :) [12:01] Also version 168 does not have protection against --split-cache and then re-attach the cache [12:01] It will use the old metadata and corrupt your filesystem [12:01] You must be better at it then I am ;-) [12:01] Right, the way I deal with it is to change the cloned VGs' host value [12:02] Right [12:02] That's also how vgexport/vgimportclone manage it [12:02] That makes sense [12:02] Okay [12:02] Hey I am just a new user in a certain sense [12:02] I do not know how to do everything and have not had formal training or anything [12:03] Zdenek of LVM said that --split-cache was just meant as a debugging feature, not for actual use [12:03] But it was exposed that way and caused mayhem [12:03] In version 171 they now protect against that [12:04] And automatically clean the cache upon reconnect [12:04] So that was one of my "oops, corrupted my filesystem" instances [12:05] Maybe my concentration is not always perfect [12:05] But I always manage to mess it up [12:05] (I have bad short term memory these days) [12:06] I've always had shit short term [12:06] or more specifically [12:06] shit working memory [12:06] Me it started with a hit to the head [12:06] Back in 2014 [12:07] Damn [12:07] I have ADHD and some kind of melatonin/adrenaline imbalance [12:07] I /can/ sleep [12:07] I've gotten better at that once I'm in bed [12:08] but staying awake is just so ridiculously easy [12:08] :). [12:08] Nice [12:08] I am jealous a bit now [12:08] Haha [12:08] I always fall asleep when I don't want to. [12:09] I still don't know if my 4.10 kernel will boot at this point or give me an initrd prompt or something [12:09] Well [12:09] I rarely get 8 hours [12:10] maybe once or twice a month [12:10] Half of the time the kernel hangs on some USB issie [12:10] issue [12:10] so I'm going to die quite a bit younger than you [12:10] :P [12:10] My current boot took a few minutes because it was hanging on usb3 [12:10] Other times, it is hanging on some USB stick [12:11] USB stick = entire kernel stalls [12:11]