golinux | <eyalroz> But, you know, making Devuan from Debian should not have been that difficult either | 00:10 |
---|---|---|
golinux | That is somewhat insulting to our significant efforts of over 6 years. | 00:11 |
golinux | Actually more than somewhat . . . | 00:11 |
ranix | can't just migrate to a different init, it's kind of invasive | 00:14 |
fluffywolf | that init is invasive is why we exist. heh. | 00:15 |
fluffywolf | "init" and "invasive" shouldn't be in the same sentence. | 00:15 |
ranix | lennart poettering sperging about devuan on twitter again | 00:16 |
eyalroz | golinux: I said "should", I didn't say it was | 00:18 |
eyalroz | golinux: If anything, it's an insult against Debian for making it so difficult, while their official decisison was merely making systemd the default, not something you couldn't get rid of. | 00:19 |
tomg | ranix, it's not just his autism, he has narcissistic traits too | 04:08 |
krzych | I upgrade Beowulf end reboot. What happened? https://krzy.ch/p/lvm.jpg I had encrypted lvm... | 08:56 |
tele1234567891 | Hi! Devuan CD probably have bug from Debian. Missing /sbin in PATH in root. Working export PATH=$PATH:sbin , working also su -l | 10:14 |
tele1234567891 | But is there way to detect this bug ? | 10:14 |
tele1234567891 | Ok. So I will load /sbin by default | 10:32 |
tele1234567891 | if [[ ":$PATH:" == *":/sbin:"* ]]; then echo is ; else echo not ; export $PATH=$PATH:/sbin ; fi | 10:32 |
gnarface | if he comes back, tell him it's actually a purposeful change from upstream and we also resent it | 10:57 |
melodie | hi | 13:32 |
melodie | I would like to find out which web browser is the lightest on cpu and ram nowadays, for a very old computer where Devuan 32bits behaves perfectly? I have installed Chromium Browser which is not bad, however a message comes with it, stating it won't be available in a near future. | 13:34 |
melodie | what do you think? | 13:34 |
buZz | melodie: 'surf' by suckless is very lightweight and does many modern websites | 13:34 |
djph | lynx | 13:34 |
djph | :D | 13:34 |
gnarface | i'd recommend moving your swap partition to zram then just using firefox or chrome | 13:35 |
gnarface | chromium* | 13:35 |
gnarface | the problem is obviously lynx is the fastest but you're giving up some features | 13:35 |
nanonymous | just use any text browser | 13:35 |
gnarface | but some of the heavy stuff can be disabled in firefox then you can offset the OOM issue very nicely with zram for swap | 13:36 |
melodie | buZz I'll look into it right away. Thanks. Do you know if Youtube works in Surf? | 13:36 |
gnarface | enough that it might be better than going with a stripped-down browser, but obviously there are other options | 13:36 |
buZz | melodie: hmm not sure, MUCH lightweight-er for youtube is just youtube-dl + mplayer/vlc | 13:36 |
linearain | hello, what's the difference between "allow-hotplug eth0" and "auto eth0" in /etc/network/interfaces? | 13:36 |
gnarface | if you just need youtube links to work you could pipe youtube-dl to mplayer or something like that, for example | 13:36 |
gnarface | you beat me buZz | 13:37 |
buZz | webbrowsers are -not- ideal for video playback | 13:37 |
melodie | buZz is there a gui available in the repos for youtube-dl, such as Clipgrab? | 13:37 |
gnarface | midori might be still worth checking out for not video playback stuff? | 13:37 |
buZz | melodie: no clue | 13:37 |
melodie | the computer is intended for non tech people | 13:37 |
buZz | i dont use guis if i can avoid them on lightweight-needed systems | 13:37 |
nanonymous | enjoy text interface | 13:37 |
buZz | non tech ppl benefit from learning CLI ;) | 13:38 |
buZz | hell, all ppl do | 13:38 |
gnarface | melodie: do really try the zram thing then. the problem with the other browsers is their interfaces are considerably less mainstream too | 13:38 |
melodie | gnarface of course, I have been using zram since before when the name was compcache : it was not even a kernel module then | 13:38 |
gnarface | firefox's problem is probably more memory usage than cpu usage for you would be my guess | 13:39 |
gnarface | my guess is chromium is in a similar situation | 13:39 |
gnarface | unless we're talking about like a pentium2 here | 13:39 |
melodie | gnarface I've had help here a few days ago to make it work in Devuan. This is what I came out with : http://meets.free.fr/Downloads/Zram-For-Devuan/ | 13:40 |
gnarface | but anything later than a core 2 duo is probably mostly ram constrained | 13:40 |
linearain | if i have no graphical user interface, no windowing or desktop, etc, do i still need devuan to get rid of systemd? could i just install system V init from debian repositories and remove systemd? or its not that simple? | 13:40 |
buZz | linearain: not that simple, no | 13:40 |
gnarface | linearain: a bunch of important packages like popular window managers got tied to it and you have to recompile them to stop them from pulling it back in by force | 13:40 |
buZz | many packages (that do not depend on clicky interfaces with guis) depend on systemd | 13:40 |
melodie | linearain Devuan makes it easy, you can get netinstall or whatever server | 13:41 |
melodie | linearain if you want it complicated, why not try LFS? -_- | 13:41 |
buZz | quite sure there's no systemV init in current debian repos anyway | 13:41 |
gnarface | linearain: that was one of their promises before hand but it was like the first one broken too | 13:41 |
linearain | i dont want to reinstall my debian buster os | 13:41 |
buZz | systemV init isnt a monolithic tool like systemderp is | 13:41 |
gnarface | linearain: you might be able to upgrade to devuan directly | 13:42 |
melodie | gnarface if someone knows how to make a package for Devuan, or for Debian, my tarball can be used, then be a "non-systemd-zram" | 13:42 |
gnarface | linearain: it works for most people, you just might have to uninstall and reinstall a few packages | 13:42 |
buZz | melodie: zram already doesnt depend on systemd .. | 13:42 |
melodie | buZz wrong | 13:42 |
buZz | ? | 13:42 |
melodie | the zram-config package does | 13:42 |
melodie | you can check | 13:42 |
buZz | ah | 13:43 |
melodie | ZRAM is the kernel module | 13:43 |
melodie | but | 13:43 |
gnarface | do you need that or can you just load the kernel module manually? | 13:43 |
gnarface | i thought you could put it in the startup scripts pretty easily | 13:43 |
buZz | melodie: zram-tools on devuan | 13:43 |
melodie | it needs to be loaded and configured to fit the needs of a given type of computer | 13:43 |
buZz | gnarface: the latter | 13:43 |
melodie | buZz I know, I tried it at first but it depends on systemd to be loaded at boot | 13:43 |
melodie | no systemd no zram | 13:43 |
buZz | zram-tools doesnt depend on systemd | 13:44 |
melodie | this is why I needed help here to figure it out, and then I provided the tarball on that place I have | 13:44 |
melodie | buZz wrong again | 13:44 |
buZz | root@h81m:/home/buzz# /usr/sbin/zramswap start | 13:44 |
buZz | works directly | 13:44 |
buZz | zero systemd needed | 13:44 |
melodie | you install you start it by hand you reboot : it's not there anymore | 13:44 |
linearain | debian repos have "sysvinit-core - System-V-like init utilities" and sysvinit-utils... | 13:44 |
gnarface | melodie: another minor issue if we are indeed talking about a core2duo recently salvaged from windows; despite how many of them were sold they are all 64-bit | 13:45 |
melodie | gnarface much worse : Asus 4000 A4D : year 2002 | 13:45 |
nanonymous | why try to remove systemdix when you can simply install a distro without it | 13:45 |
melodie | Devuan works perfect there ! | 13:45 |
buZz | gnarface: even core1duo was 64bit, iirc | 13:45 |
melodie | AMD Sempron 3000+ 1.8 Ghz and 2GB ram PC-2700 | 13:46 |
linearain | nanonymous: because i have a server with tons of stuff on it, i would hate to reinstall | 13:46 |
linearain | but systemd is ugly and buggy | 13:46 |
melodie | buZz do your reboot and let me know if you don't need to do " # /usr/sbin/zramswap start " again | 13:46 |
melodie | I've been there! | 13:46 |
buZz | melodie: add it to rc.local | 13:47 |
melodie | in antiX also where systemd is not the default | 13:47 |
melodie | buZz I had tried everything, the set of files I pointed to does the job | 13:47 |
melodie | I invite you to have a look : http://meets.free.fr/Downloads/Zram-For-Devuan/zram/ | 13:47 |
buZz | lol @ typos in it :P | 13:48 |
melodie | also the default file allows refining the amount of RAM used to create the virtual block device : http://meets.free.fr/Downloads/Zram-For-Devuan/zram/etc/default/zram | 13:48 |
buZz | percentages arent factors ;) | 13:48 |
melodie | so what? | 13:49 |
linearain | buZz: which non-gui packages are dependant on systemd? name a few please.... | 13:49 |
buZz | dbus for instance | 13:49 |
buZz | consolekit for another | 13:49 |
buZz | plus many if not all services | 13:49 |
melodie | buZz you are welcome to send an improved version which I will try | 13:50 |
linearain | if i did apt-get --simulate when migrating to devuan, that would show me all the changes that would take place? | 13:50 |
melodie | typos can be fixed, as long as the program still works and allows having zram working out of the box | 13:50 |
linearain | looks like most of my services have scripts in /etc/init.d/ which systemd uses with a wrapper, so that would simplify things i guess | 13:58 |
knofix | ni | 14:31 |
knofix | hi all4 | 14:31 |
knofix | one quick question if anyone is willing and able and has a min of spare time -> smartctl -h outputs SAT command failed, any clue? | 14:32 |
aitor_ | <linearain>: "what's the difference between "allow-hotplug eth0" and "auto eth0" in /etc/network/interfaces?" | 14:32 |
knofix | regular 2.5 hdd via usb 3.0 capable case paired with usb 2.0 on a machine | 14:32 |
aitor_ | the "auto" stanza is deprecated and it's only used for the loopback "auto lo" | 14:33 |
aitor_ | use "allow-hotplug" instead, for the rest of the devices | 14:34 |
knofix | will wait for udev if allow-hotplug | 14:34 |
knofix | or cable connec | 14:34 |
knofix | t | 14:34 |
aitor_ | knofix, during the boot, you mean? | 14:35 |
linearain | i see. So I don't need wicd? | 14:35 |
knofix | not sure :) | 14:35 |
knofix | not using systemd | 14:35 |
linearain | do i need wicd if i only use wired networks and i have no gui? | 14:36 |
buZz | no | 14:36 |
aitor_ | linearain: on the one hand, /etc/network/interfaces is related to ifupdown, not to wicd | 14:37 |
buZz | dhcp etc works fine without | 14:37 |
linearain | great | 14:37 |
aitor_ | knofix: yes, allow-hotplug eth0 leads to a delay in booting | 14:44 |
aitor_ | but you don't need it in a daily use | 14:44 |
aitor_ | but you'll require it if you are using a bridge interface for a virtual machine | 14:46 |
knofix | brb | 14:47 |
fsmithred | since when is "auto" deprecated? | 14:48 |
fsmithred | allow-hotplug tells udev to bring up the network, auto tells ifup to bring it up | 14:48 |
fsmithred | if you use allow-hotplug with the ethernet cable disconnected, you get a boot delay | 14:49 |
linearain | my debian 10.1 interfaces has three lines "source /etc/network/interfaces.d/*, auto lo, iface lo inet loopback", where is the NIC defined? | 14:49 |
fsmithred | linearain, if you use wicd or other network gui, you don't need to define the interface in that file | 14:50 |
aitor_ | fsmithred: once i repackaged ifupdown for beowulf, but i can't remember in detail... maybe you're right | 14:50 |
fsmithred | if you don't use any network manager software, you can define it manually | 14:50 |
linearain | ohh | 14:51 |
fsmithred | ralph gave a nice explanation of the difference between the two ways of bringing up the interfaces in a thread on the forum | 14:51 |
linearain | devuan migration manual says i should add "allow-hotplug eth0" "iface eth0 inet dhcp" to interfaces file. This is in case i don't have a network manager at all? | 14:52 |
fsmithred | also, check in /etc/network/interfaces.d/ for files. | 14:52 |
linearain | its empty | 14:52 |
fsmithred | yeah, that's in case you don't have any network manager | 14:53 |
fsmithred | uh oh | 14:53 |
linearain | so what is the point of a network manager? sorry im dumb | 14:53 |
fsmithred | you're following the migration instructions on devuan.org? | 14:53 |
buZz | fsmithred: isnt that normally empty? | 14:53 |
linearain | yes, i am reading about it, but not doing it yet | 14:53 |
fsmithred | network managers are excellent choice for a laptop that connects to different networks | 14:53 |
fsmithred | good. Those instructions do not work anymore. | 14:54 |
linearain | i see, my case is a server. | 14:54 |
fsmithred | And neither do the ones on the forum. | 14:54 |
fsmithred | Right now I do not know a way to migrate buster to beowulf. | 14:54 |
fsmithred | No shit. | 14:54 |
linearain | i'm interested in migrating buster 10.1 (late 2019) to beowulf 3.0.0 | 14:54 |
fsmithred | server might be ok | 14:55 |
fsmithred | well, you'll get beowulf 3.1 (actually a little newer than that) because that's where the packge versions are now. | 14:55 |
linearain | i have the 3.0.0 .iso file | 14:55 |
linearain | i can use that as a repository, cant i? | 14:56 |
fsmithred | and you're going to use that for the upgrade? | 14:56 |
fsmithred | I guess so. I've never done that. | 14:56 |
linearain | i might try to. I want it reproducible without internet... | 14:56 |
linearain | or I could shrink my partition, and install fresh devuan on a new small partition... i will figure it out eventually | 14:57 |
* iv4nshm4k0v prefers to upgrade by installing the new version to a new set of LVM logical volumes and booting from there | 14:57 | |
fsmithred | check your installed packages against what is in /pool in the iso to make sure everything is there. | 14:58 |
linearain | first i have to prune all the undeeded stuff, since i originally installed debian with gui, and now i use it headless | 14:59 |
fsmithred | oh, that should be fun | 14:59 |
linearain | well it was my first ever linux install of any kind... | 14:59 |
fsmithred | you could probably just purge task-?-desktop for your selected desktop environment | 15:00 |
fsmithred | and that would get rid of all the gui stuff | 15:01 |
fsmithred | there's also a task-desktop package | 15:01 |
linearain | task-desktop task-english task-lxqt-desktop task-print-server task-ssh-server task-web-server | 15:01 |
fsmithred | that might be the primary one. | 15:01 |
fsmithred | keep ssh server! | 15:02 |
fsmithred | lol | 15:02 |
linearain | what does task- mean? | 15:02 |
fsmithred | oy | 15:02 |
fsmithred | oops | 15:02 |
fsmithred | it's a metapackage | 15:02 |
fsmithred | not a real package but depends on several real packages | 15:02 |
fsmithred | so it's an easy way to add or remove groups of packages | 15:02 |
linearain | i need the web server too | 15:03 |
linearain | what the hell is task-english | 15:03 |
fsmithred | git rid of task-desktop and task-lxde-desktop should do it | 15:03 |
fsmithred | dictionaries, I think | 15:03 |
aitor_ | the task work probably comes from the tasksel package | 15:03 |
fsmithred | yeah, and locales | 15:03 |
fsmithred | yes | 15:03 |
fsmithred | tasksel is the source package for all of those | 15:04 |
aitor_ | task has its own meaning: "tarea" in spanish | 15:04 |
linearain | the amount of packages to be purged by task-lxqt-desktop barely fits my screen | 15:05 |
fsmithred | that sounds right | 15:05 |
fsmithred | actually, I'd expect it to scroll off the screen | 15:06 |
linearain | i mean without newlines | 15:06 |
fsmithred | oh! | 15:07 |
linearain | i did apt-get --simulate purge task-lxqt-desktop | 15:07 |
fsmithred | you have backups of this system? | 15:07 |
linearain | well... i have, but not the kind of i could easily rollback to. | 15:08 |
aitor_ | time to eat, see you later | 15:08 |
linearain | i have another computer with the same debian os, so i would first try the process on it | 15:08 |
fsmithred | you should have something ready. Like I said, I just tested buster to beowulf migration by two different methods and failed. | 15:09 |
linearain | what happened? | 15:09 |
fsmithred | you may have better luck with server. On one of the tests, the first thing that happened was I got a black screen. | 15:09 |
fsmithred | This was in a VM, so I couldn't drop to console, but I could boot to single user mode. | 15:10 |
fsmithred | lemme find my notes | 15:10 |
fsmithred | and I'll get the link to the alternate instructions | 15:10 |
linearain | most of those task-lxqt-desktop packages seem to be java related | 15:12 |
fsmithred | for libreoffice | 15:13 |
fsmithred | here's what I wrote about nixer's method (the alternate) | 15:13 |
fsmithred | Install sysvinit-core and screen goes black. End of story. Not end! init 1, login as root, continue the upgade. Try to reboot and get /run/initctl error. Kill it and reboot to console login. Then autoremove the whole system. Reconnect network. Install slim. Install eudev. Reboot to desktop. Wow. Few hundred packages are missing. | 15:13 |
fsmithred | Here's the link to a discussion about migration. Link is to my summary. Scroll up a few posts for nixer's longer description: https://dev1galaxy.org/viewtopic.php?pid=18390#p18390 | 15:15 |
fsmithred | we're getting ready to re-write the instructions on the website, so if you have anything to add or report, please let me know. | 15:16 |
fsmithred | back in a few minutes. need food. | 15:17 |
linearain | sure i will tell how it went once i try to do it | 15:18 |
fsmithred | thanks | 15:31 |
melodie | this midori works perfect in Devuan : https://astian.org/en/midori-browser/download/ | 16:05 |
melodie | Icon | 16:05 |
melodie | 1.1.4_i386.deb | 16:05 |
melodie | this version number is smaller than the one on the repos, however the program is in reality more recent. that's not normal | 16:05 |
melodie | what I can say, this one plays youtube video very well no hickups, in Asus 4000 with proc amd sempron 3000+, so this is the real deal! | 16:10 |
gnarface | midori plays youtube videos for you? i wouldn't have expected video to work at all, interesting | 16:20 |
fsmithred | wow. Confirmed. The astian midori plays youtube videos. | 16:34 |
fsmithred | I don't understand the version. The debian changelog jumped from 0.5 in 2016 to 7.x in the next entry (2018) | 16:34 |
fsmithred | or 6.x | 16:35 |
fsmithred | Settings (three dots), Extensions -> goes to chrome web store | 16:36 |
fsmithred | ps_mem.py says it's using 700MB | 16:38 |
fsmithred | oh, 634MB if I close the chrome store tab that has a slide show running | 16:39 |
fsmithred | I guess they re-versioned when they switched to chrome | 16:41 |
EHeM | Seems stable got broken on pkgmaster. | 16:44 |
gnarface | midori switched to the chrome engine? | 16:44 |
gnarface | weird | 16:44 |
EHeM | The irony is HTML was meant to be LaTeX-like (describe your data, let the /browser/ decide how to render it), yet nearly everything past the first version or two has made it more and more PostScript-like. | 16:47 |
EHeM | Stable being broken on pkgmaster seems like a rather urgent issue. | 16:47 |
buZz | EHeM: yeah, html turned out to be a awefully poor replacement for gopher | 16:48 |
buZz | :( | 16:48 |
EHeM | Stable presently appears to be broken on all mirrors. | 17:01 |
mason | EHeM: Broken in what sense? | 17:03 |
EHeM | Unavailable even though the servers appear to be up, missing Release file. | 17:04 |
mason | EHeM: Check for a local network issue or somesuch. Working here. | 17:05 |
mason | EHeM: Do you refer to it as stable or beowulf? If stable, try beowulf. | 17:06 |
EHeM | mason: A local network issue would make the server appear down (DNS failing or packets being dropped). | 17:06 |
fsmithred | don't even bother to check for network issues | 17:07 |
mason | EHeM: Using the actual name can be perilous when new releases occur, as you'll suddenly start pulling from a different release. | 17:07 |
fsmithred | just repeat in a few minutes (or seconds) | 17:07 |
mason | fsmithred: Oh, was something broken-then-fixed? | 17:07 |
fsmithred | Temporary failure resolving 'deb.devuan.org' | 17:07 |
fsmithred | ^^^ I get that several times a week | 17:07 |
fsmithred | it usually clears up in a minute or less | 17:08 |
EHeM | mason: Stable, such has long traditionally been available since it is quite handy if you use it properly. | 17:08 |
mason | Hrm. I never see that here. I tried against deb.devuan and then switched a box to pkgmaster to compare, and both worked. | 17:08 |
fsmithred | cleared now | 17:08 |
mason | EHeM: It's worth comparing 'beowulf' to rule that out. | 17:08 |
fsmithred | EHeM, are you using "stable" or "beowulf" in sources.list? | 17:08 |
EHeM | fsmithred: "stable", keeps me aware when release updates come out. | 17:09 |
fsmithred | does it differentiate between debian and devuan release updates? | 17:10 |
EHeM | My understanding was stable was maintained on the servers as a symlink. | 17:12 |
mason | fsmithred: stable only refers to devuan | 17:12 |
mason | EHeM: Yeah, but I've seen random anomalies in the past, hence the thought of test-swapping. | 17:13 |
fsmithred | I understand the symlinks on our server, but how does amprolla pull from debian? | 17:13 |
fsmithred | it's keyed to the debian codename? | 17:14 |
mason | fsmithred: It emits merged, and nginx itself diverts as needed to debian, so it's impossible to access an unmodified debian straight through our servers. | 17:14 |
mason | no matter what name you use | 17:14 |
mason | it'll always point to Devan | 17:14 |
mason | Devuan, I can't type | 17:14 |
fsmithred | so whether I use stable or beowulf, amprolla knows to pull from "bullseye"? | 17:15 |
fsmithred | and not from debian "stable"? | 17:15 |
mason | buster | 17:15 |
fsmithred | yeah, I just saw that | 17:15 |
fsmithred | sorry, buster | 17:15 |
fsmithred | and another fucking 'b' name coming after that. | 17:16 |
fsmithred | three in a row | 17:16 |
mason | I'd have to look for complete certainty, but I'm pretty sure it uses the release names, not the aliases, because the aliases change over time. | 17:16 |
mason | the release names are always fixed | 17:16 |
fsmithred | so in theory, it's no worse than it used to be in debian to use the suite names instead of the codenames | 17:17 |
mason | I don't want to think about Amprolla too much now, because it'll send me back to lurcing when we need the wiki more. :P | 17:17 |
mason | right | 17:17 |
fsmithred | accidental dist-upgrad requires that you ignore the fact that over 1000 packages need to be upgraded | 17:17 |
mason | or include an unfortunate -y | 17:18 |
fsmithred | (been there, done that and it included xfree to xorg conversion) | 17:18 |
fsmithred | anyway, I just switched beowulf to stable in sources.list and it still updates without error | 17:18 |
fsmithred | EHeM, are you using deb.devuan.org or a specific mirror? | 17:19 |
mason | he'd noted pkgmaster earlier | 17:25 |
fsmithred | if that were the case, all the mirrors would be screwed up | 17:25 |
mason | That's why I suggested looking for some local network error, as I didn't see either case (deb or pkgmaster) fail here | 17:26 |
mason | The thing I didn't try was stable-vs-beowulf, since I never use the aliases | 17:27 |
fsmithred | I'm trying that now | 17:27 |
fsmithred | with pkgmaster | 17:27 |
fsmithred | works ok here | 17:27 |
Ankokukishi | ive never had an issue with deb.devuan.org connecting | 17:30 |
Ankokukishi | i just tested to see if sudo apt update works and it connected fine, even the backports | 17:31 |
fsmithred | I might get redirected to a mirror that is down more often | 17:32 |
fsmithred | mirror checker is here: https://sledjhamr.org/apt-panopticon/results/Report-web.html | 17:32 |
EHeM | Now double-checking notes, I've tended to stick with us.deb.devuan.org on the theory it might get a closer local mirror (if it simply round-robins, no issue, just if the feature becomes available again...); I'd been doing reverse lookup and seeing pkgmaster. | 17:57 |
EHeM | 152.228.204.144 | 17:57 |
fsmithred | the country codes don't really work | 17:58 |
gnarface | it was never set up to work like that but there aren't any us mirrors anyway | 17:58 |
gnarface | afaik the country-code domains have been officially deprecated | 17:58 |
fsmithred | pkgmaster tends to be slow, too | 17:58 |
EHeM | And I don't consider merely doing round-robin to be a problem. | 17:58 |
fsmithred | try deb.devuan.org | 17:59 |
fsmithred | if that gives you trouble, pick a specific mirror. sledghamr and ungleich tend to be fast | 17:59 |
linearain | I have one partition on a hdd, i want to shrink it and then clone it so i have two identical partitions on one hdd. Could someone give me some tips on how to make it so i can select which one to boot at grub? with different names | 18:00 |
gnarface | grub will pretty much figure it out when you run update-grub | 18:01 |
EHeM | Unless APT provides the hostname to the server, and some servers explicitly reject the country-codes I wouldn't expect the non-country-codes to be different. | 18:02 |
nemo | crud | 18:02 |
nemo | my name was referenced | 18:02 |
nemo | but since I don't bother logging, and I was checked out of freenode for a week on a beach | 18:03 |
fsmithred | deb.devuan.org is the round-robin. I'm pretty sure it won't direct you to pkgmaster | 18:03 |
nemo | I have no idea why | 18:03 |
nemo | oh well. hopefully it wasn't important | 18:03 |
fsmithred | nemo, I think they were talking about cinnamon | 18:03 |
nemo | ah | 18:03 |
fsmithred | false alarm | 18:03 |
nemo | heh | 18:03 |
nemo | that does happen from time to time | 18:04 |
linearain | gnarface: i'm a bit dumb but should i do update-grub from the live distribution that im planning to do the partition cloning with? | 18:06 |
fsmithred | linearain, do that from the existing install (the shrunk one) | 18:06 |
linearain | alright, i'm really not good with linux | 18:07 |
linearain | thanks | 18:07 |
fsmithred | are you going to use rsync to clone? | 18:07 |
fsmithred | or some backup software? | 18:08 |
gnarface | linearain: well you can do it from a live cd too but you'd have to chroot in and mount a bunch of stuff manually | 18:08 |
linearain | i guess i will use resize2fs to shrink it, then parted or gparted to make a partition of same size as the first, shrunk one, and then probably dd to copy the contencts of 1st partition to the new 2nd partition | 18:08 |
gnarface | gparted can do the shrinking too, with some filesystems (optional tools sometimes required) | 18:09 |
fsmithred | you'll probably need to manually edit /etc/fstab in the cloned copy | 18:09 |
gnarface | for that matter it can also do the copying but i like dd too | 18:09 |
gnarface | hmm, yes your UUIDs will all be wrong | 18:10 |
gnarface | minor stuff | 18:10 |
linearain | where do i find the correct UUIDS/ | 18:10 |
linearain | for each partition | 18:10 |
fsmithred | blkid | 18:10 |
linearain | but they will be identical if i just dd... :D | 18:11 |
fsmithred | not if you dd the contents of the partition instead of the partition itself | 18:11 |
gnarface | well you can also just ls -l /dev/disk/by-id/ | 18:11 |
EHeM | `tune2fs` can be used to modify FS UUIDs. | 18:11 |
gnarface | (as any user) | 18:11 |
linearain | ohh right | 18:11 |
gnarface | linearain: uuids regenerated on filesystem resize | 18:12 |
linearain | i keep confusing partitions and filesystems | 18:12 |
nemo | orrr if you know your hardware won't be shuffling around, /dev/sda1 is still an option... | 18:12 |
gnarface | linearain: you'll have to edit the fstab in both partitions | 18:12 |
nemo | but yeah, fixing uuids sounds smarter | 18:12 |
linearain | no big deal | 18:12 |
nemo | I do like the ease of typing ☺ | 18:12 |
gnarface | linearain: keep that live cd happy | 18:12 |
gnarface | linearain: handy* | 18:12 |
onefang | I much prefer labels aver UUIDs. | 18:13 |
EHeM | Yes, installing onto a system by doing `debootstrap` onto fresh media then physically moving the disk is generally not recommended, but if you know what you're doing it can work. | 18:14 |
mason | I'm a big fan of GPT labels over UUIDs. | 18:16 |
mason | Human-readable. | 18:16 |
EHeM | Other issue is many tools tend to prefer FS UUIDs as they are robust against moving the filesystem on storage media; on the flip side that robustness isn't 100% good, if you've got VMs with disks directly mapping to storage medium the filesystem UUIDs are under control of the VM not the hypervisor. | 18:18 |
linearain | gnarface: how would i copy the partition with contents with gparted? | 18:26 |
linearain | oh nevermind | 18:27 |
JPSavage | Unable to install DEB using gdebi | 18:27 |
gnarface | linearain: right-click maybe? i don't have it in front of me | 18:28 |
linearain | yeah it has copy paste | 18:28 |
linearain | i might try gparted live cd | 18:28 |
gnarface | linearain: it's good | 18:28 |
Guest78952 | hi | 20:35 |
aitor_ | hi | 20:36 |
aitor_ | you can also get the FS UUIDs by this other way: | 20:37 |
aitor_ | blkid -s UUID -o value /dev/sdb1 | 20:37 |
aitor_ | gnarface: /dev/disk/by-id is missing in my system with vdev | 20:38 |
aitor_ | and the GPT labels by typing: | 20:39 |
aitor_ | blkid -s LABEL -o value /dev/sdb1 | 20:39 |
aitor_ | and the FSTYPE: | 20:40 |
aitor_ | blkid -o value -s TYPE /dev/sd1 | 20:40 |
aitor_ | */dev/sdb1 | 20:40 |
aitor_ | it requires libblkid1 | 20:42 |
aitor_ | like lsblk, it's part of the miscellaneous utilities of "util-linux" | 20:45 |
linearain | well i did it, bootable gparted is handy. I had to edit /etc/fstab of the cloned partition, and did update-grub, but that wasnt enough. Both entries in grub would boot with the same root filesystem, so i had to manually edit the grub and correct the uuid in one place. uuids are a pain in the ass | 20:46 |
linearain | I mean i had to edit grub.cfg | 20:46 |
linearain | now that i have a backup i can try the migration process! | 20:49 |
aitor_ | go ahead and good luck :) | 20:50 |
linearain | thanks | 20:50 |
aitor_ | i'm finishing my packages of hopman (Didier's project), i recently added a new interface to it | 20:52 |
linearain | aitor_: do you know how to use labels instead of uuids in such cases? linux /boot/vmlinuz-4.19 root=UUID=xx-xx-xx ro quiet | 20:52 |
linearain | this is from grub.cfg | 20:52 |
aitor_ | nope, i never did it | 20:53 |
linearain | probably root=label= | 20:53 |
aitor_ | using search --label | 20:55 |
aitor_ | look at this example: https://wiki.archlinux.org/index.php/GRUB#Using_labels | 20:56 |
aitor_ | one clarification: to get the label give the device name, not the partition name, e.g. /dev/sdb instad of /dev/sdb1 | 21:55 |
aitor_ | blkid -s LABEL -o value /dev/sdb | 21:55 |
aitor_ | i rebooted the system and /dev/disk/by-id an /dev/disk/by-uuid are present with vdev; thay were missing because i removed one action from /etc/vdev/actions during my tests | 23:11 |
aitor_ | but /dev/disk/by-label doesn't exist | 23:11 |
aitor_ | however, blkid works giving the right label | 23:15 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!