| al1r4d | \o/ https://lists.dyne.org/lurker/message/20150214.203545.7ba2d477.en.html | 01:39 |
|---|---|---|
| golinux | al1r4d: Yes! It is 10 years since the first pre alpha isos were released!! | 01:47 |
| Xenguy | <3 | 01:48 |
| rustyaxe | lawd, the curse of systemd has really been around so long? | 01:53 |
| rustyaxe | Thanks for the freedom from that crap <3 | 01:54 |
| TorC | Running into an Excalibur netinstall issue. | 02:08 |
| TorC | "Unmerged /usr is not compatible with excalibur" | 02:08 |
| TorC | This is attempting a clean install into blank / and mounting an extant /home, though I'm at the installing base system step. | 02:09 |
| fsmithred | TorC maybe you need to choose expert install to get a question about merged usr, for which you should answer yes. | 02:12 |
| TorC | Ah, hadn't thought of trying that. Will do, one moment. | 02:13 |
| fsmithred | it's been a few months since I did a netinstall, so I might not be remembering right | 02:13 |
| TorC | Well, I've dealt with bugs of various kinds before. Figured Excalibur should be established enough to be reasonably stable for production use by now, and I've ended up on relatively new hardware after some failures, so figured I'd go straight there from Beowulf. | 02:16 |
| fsmithred | the testing suite will be better after debian freezes trixie. Did expert install help? | 02:35 |
| TorC | Still working on it. HW detection steps have been running slow for reasons unknown. It was happening on Daedelus live, too, so I'm ignoring it. | 02:36 |
| TorC | There was a question in expert install, looks like it's installing fine now. Thanks for the pointer, fsmithred. | 02:54 |
| TorC | Is it possible the netinst image is somehow set up for IPv6 only on the download? | 03:00 |
| TorC | Having trouble syncing either deb.devuan.org or us.deb.devuan.org. | 03:01 |
| gnarface | does it persist if you retry? | 03:01 |
| gnarface | could just be a problem with a particular mirror | 03:02 |
| TorC | On second check, it looks like it's the update repo, as it goes through for the initial set OK first. Done it three times on deb, twice on us.deb, and now once on beard.ly. | 03:03 |
| TorC | Well, continuing for now, and we'll see what happens, I guess. I think it's trying all four options it offered, so we'll see how it goes. | 03:04 |
| gnarface | i don't know for sure that it's not a problem with the testing isos, but we have also seen some weird issue a couple times where a certain (Australian?) ISPs' DNS servers weren't handling ipv6 records right, or something like that | 03:06 |
| gnarface | we usually recommend running your own local nameserver in such cases, but switching to google's public dns also works fine | 03:08 |
| gnarface | (if you don't care about privacy) | 03:08 |
| TorC | Well, it's a case where I have two ISPs. One does IPv6 fine, but it's poor by comparison. The good one is completely without IPv6, so I have it turned off on both. | 03:11 |
| gnarface | hmm, that might be what's doing it, if you are getting ipv6 responses over ipv4 | 03:13 |
| gnarface | i don't think any of the mirrors are ipv6 only, but some of them are both | 03:13 |
| TorC | well, seems to have downloaded and installed, so I guess we'll call it a mystery error for now. | 03:23 |
| rrq | excalibur (=testing) doesn't have updates and security set up yet | 03:36 |
| TorC | Ah, right. That makes sense. Thanks for the explanation. | 03:39 |
| TorC | Sigh. Fatal error initializing (AMD) GPU at boot. System already ran Daedelus live, so it should be possible to work. And... I can't see the USB network I used just fine on the same, since the RTL8125 2.5G drivers weren't available. | 04:13 |
| TorC | It claimed network autoconnect worked fine, and didn't throw errors ad being asked to install KD, so in theory it had drivers for at least the USB. | 04:17 |
| gnarface | TorC: missing firmware maybe | 05:27 |
| gnarface | firmware-amd-graphics for the amdgpu and i'm guessing firmware-realtek for the RTL | 05:28 |
| gnarface | both in non-free-firmware | 05:28 |
| gnarface | the live iso would have had them already | 05:28 |
| TorC | Could be. I did enable non-free firmware, although now I've got bigger problems than are appropriate for this channel before I can get back to what is going on. | 05:28 |
| TorC | Started getting a USB overcurrent error on boot after removing a suspected misbehaving internal USB card reader, and now I can't even get BIOS display or a boot. | 05:29 |
| gnarface | hmm, that rings a bell, might actually be a kernel issue i think... | 05:30 |
| TorC | Probably not, since it was BIOS throwing the error. This whole upgrade has been by far the worst computer upgrade I've ever had. | 05:30 |
| gnarface | oh, maybe not then | 05:30 |
| TorC | Quite a few issues before I ever got so far as even having the live Daedelus running. | 05:31 |
| TorC | Ah well. That's troubleshooting. One step at a time. | 05:31 |
| TorC | Well, in case it's useful to anyone: USB overcurrent error not visibly an issue cleared from Asus MB by clearing CMOS. | 05:36 |
| TorC | After that it took a few boot attempts to get even the internal GPU to start behaving again (awaiting new graphics card, which is another story.) | 05:37 |
| TorC | Well, I guess we'll blame the GPU issue on gremlins or that USB card reader. Once I reset the BIOS settings it booted right up slick as can be. | 05:43 |
| Guest67 | fsmithred on devuan 5 live here. i use refractasnaphot and get "/tmp/extracted /mnt/refracta/work" and then "Decompession error.." (maybe compRession is ment?) | 13:26 |
| Guest67 | and cant find the iso. | 13:26 |
| Guest67 | i run devuan live on a ~800mb partition on that usb, the rest of the stick (~13gb) is a ext4 partition used as a mountable storage. i edited /etc/refractasnapshot.conf to point the dirs that are by default pointing to /home, to /mnt/heremy-usb-storage | 13:28 |
| Guest67 | oh also the symbolic links in devuan live for the initrd and i think also the kernel are all broken. | 13:29 |
| fsmithred | Guest67, if you have amd or intel microcode installed, you probably need the newer version of refractasnapshot | 14:06 |
| fsmithred | I'll get a link | 14:06 |
| fsmithred | https://pkgmaster.devuan.org/devuan/pool/main/r/refractasnapshot-base/refractasnapshot-base_10.4.1_all.deb | 14:07 |
| fsmithred | it will install in daedalus or chimaera. | 14:08 |
| Guest67 | i dont know if i have, but i know that i used your refracta in the past to build isos on the same computer. | 14:14 |
| fsmithred | what problem are you having with the sym links? The boot menu uses /vmlinuz and /initrd.img. Which iso are you using? | 14:15 |
| Guest67 | does the refracta on the devuan 5 live still support 32-bit btw? | 14:15 |
| fsmithred | yes | 14:15 |
| fsmithred | you're using the i386 minimal-live iso? | 14:15 |
| fsmithred | dpkg -l | grep microcode | 14:16 |
| Guest67 | initrd.img is a symlink to /boot/initrd.img wich is not present. instead theres /boot/initrd.img-6.1.0-9-686 and that file is a symlink to /boot/initrd.img :) | 14:17 |
| fsmithred | which iso? | 14:17 |
| Guest67 | iso is: | 14:19 |
| Guest67 | devuan_daedalus_5.0.0_i386_minimal-live.iso | 14:19 |
| Guest67 | from here: https://files.devuan.org/devuan_daedalus/minimal-live/ | 14:19 |
| fsmithred | I think that gets corrected if you install the system. And the live iso doesn't use that to boot | 14:19 |
| fsmithred | yeah, I just booted that one in qemu | 14:19 |
| fsmithred | I see what you're talking about | 14:19 |
| Guest67 | the live init is in live or something (starting the live usb now) | 14:20 |
| fsmithred | did you copy the iso to hard drive to run it live from there, or did you install it with refractainstaller? | 14:20 |
| Guest67 | no i dd it to a partition of an usb | 14:20 |
| Guest67 | live-usb | 14:20 |
| fsmithred | ok | 14:20 |
| fsmithred | I'm not sure how that happened, but I don't think it will affect anything. | 14:21 |
| Guest67 | the first computer doesnt have microcode ill check the second now, need a sec | 14:21 |
| fsmithred | ok | 14:21 |
| Guest67 | the second doesnt have microcode too. | 14:23 |
| Guest67 | the initrd is in: /lib/live/mount/medium/live/initrd.img and as it seems also in /run/live/medium/live/initrd.img | 14:26 |
| Guest67 | i searched for: find / -iname "initrd.img*" | 14:27 |
| Guest67 | all others initrdfoobar are broken symbolic links | 14:27 |
| Guest67 | btw: will devuan drop 32-bit support? | 14:42 |
| fsmithred | yes, we use debian kernels | 14:44 |
| Guest67 | and debian dropped 32-bit support? | 14:45 |
| fsmithred | I made a 32-bit live with the last 686-pae kernel in excalibur (testing), which is gone now. | 14:45 |
| Guest67 | then the day has sadly come to move away from debain i guess... | 14:46 |
| Guest67 | dropping support for old hardware is just idiotic and against the idea of freedom. | 14:46 |
| fsmithred | https://get.refracta.org/files/testing/ | 14:46 |
| fsmithred | I don't actually recommend using the goodbye iso unless you have another source for kernels that get patched. | 14:47 |
| freem | "dropping support for old hardware is just idiotic and against the idea of freedom" dropping support for old hardware is not against the idea of freedom: you are still free to do it yourself (or pay someone to do it). It is not idiotic neither, as it allows to remove code that increase maintenance costs for the whole application. | 14:47 |
| freem | I'd be interesting to see where the info that debian is dropping entire 32bit support though | 14:49 |
| Guest67 | freem yeah all the time and money debian devs had to spent... -.- how to save time by stopping supporting shitty software like systemdeath or hundreds of thousands variations of frontends for the smae backend? | 14:49 |
| fsmithred | AFAIK they are still building i386 packages | 14:49 |
| freem | https://www.debian.org/releases/testing/release-notes/whats-new.en.html debian 13, the next stable, still officially supports i386 (which really is i686 but hey) | 14:49 |
| Guest67 | freem i dont know where to find that i heard that the first time here. | 14:50 |
| freem | yes, people who contribute to debian pay. They pay their personnal time. | 14:50 |
| freem | systemd's impact on systems not using a monolithic DE is overrated, imo. | 14:51 |
| Guest67 | dropping 32-bit is also an ecological desaster and purely capitalistc: "oh your software doesnt work anymore? no problem just buy a new computer and throw aways your old reliable one!1!" | 14:51 |
| freem | I've been running debian without systemd on all my systems since many years, works well; | 14:51 |
| freem | Guest67: except they do NOT. And dropping 32bit would have surprised me for a technical reason: 32 bit systems are still produced. | 14:52 |
| Guest67 | thats typical google, aplle, microsoft marketing logic. | 14:52 |
| Guest67 | freem yeah and they are more then enough for the most tasks. | 14:52 |
| freem | yes and no. | 14:53 |
| fsmithred | this is moving toward an off-topic conversation | 14:53 |
| Guest67 | 64-bit can address more tahn 4gb of ram sure, but programs are also bigger than their 32-bit version | 14:53 |
| Guest67 | fsmithredyes sorry | 14:53 |
| freem | that mostly defines the addressable size of per-process image | 14:53 |
| Guest67 | fsmithred back to refracta | 14:53 |
| Guest67 | why i get this error? (Decompession error...) | 14:54 |
| freem | in any case, the point which is on-topic is this: debian 13 is NOT dropping i386 arch | 14:54 |
| fsmithred | Guest67, run 'refractasnapshot -d' to get a more verbose log. | 14:54 |
| Guest67 | fsmithred ok | 14:54 |
| fsmithred | Guest67, that's the error you get if microcode is installed | 14:55 |
| fsmithred | but you said it's not | 14:55 |
| Guest67 | yes, dpkg -l | grep microcode gives no output | 14:57 |
| Guest67 | oh i found the log | 14:58 |
| Guest67 | var/log/refractasnapshot.log has exactly one line -> ls: cannot access '/mnt/refracta/snapshot/*.iso': No such file or directory. | 14:59 |
| Guest67 | mnt/refracta/ is what i set in /etc/refractasnapshot.conf | 14:59 |
| Guest67 | for /snapshot and /work | 14:59 |
| Guest67 | and /mnt/work is full of files | 15:00 |
| fsmithred | right. That's why I say to run it with -d|--debug | 15:00 |
| Guest67 | and that directory exist: /mnt/refracta/snapshot | 15:00 |
| Guest67 | fsmithred ok ill do it now | 15:00 |
| fsmithred | better yet, run it with 10.4.1 and it will probably work better with the zstd compression | 15:01 |
| fsmithred | you just warped my brain. I have to think about whether you can use /mnt or not | 15:01 |
| fsmithred | uh, yeah it should be ok | 15:02 |
| Guest67 | fsmithred you mean while its exclude by default? | 15:02 |
| fsmithred | just not ok for me to use it here | 15:02 |
| fsmithred | it's ok. work dir needs to be excluded or you get in a loop | 15:03 |
| fsmithred | https://pkgmaster.devuan.org/devuan/pool/main/r/refractasnapshot-base/refractasnapshot-base_10.4.1_all.deb | 15:04 |
| Guest67 | fsmithred i get the same error | 15:13 |
| Guest67 | an no further output | 15:14 |
| Guest67 | so i mean same output at the end like without ---debug | 15:15 |
| Guest67 | fsmithred oh that debug stuff is in that log then! :o | 15:22 |
| Guest67 | ^^' | 15:22 |
| Guest67 | I send you the log wait a minute please | 15:23 |
| Guest67 | fsmithred https://paste.debian.net/plainh/77a0268c/ | 15:27 |
| Guest67 | I searched for the terms "fail", "cannot" and "error" in that log. Maybe it fails just because of that: | 15:34 |
| Guest67 | ls: cannot access '/mnt/refracta/snapshot/*.iso': No such file or directory ? | 15:34 |
| Guest67 | That directory exists and it is empty before running refractasnapshot, but also after running it... | 15:37 |
| fsmithred | back. checking... | 15:52 |
| fsmithred | 14G2.9M13G1% /mnt/dev/sdb2' | 15:55 |
| fsmithred | don't mount your second partition where you want work_dir and snapshot dir | 15:56 |
| fsmithred | what version of refractasnapshot are you using? | 15:58 |
| fsmithred | the error message about no such file means that you don't already have a snapshot.iso in the snapshot directory. It's counting the isos to tell you how many you have and how much space you have. | 16:01 |
| fsmithred | compression error is the last thing | 16:01 |
| fsmithred | refractasnapshot -v | 16:01 |
| fsmithred | what does it say? | 16:01 |
| fsmithred | how much RAM do you have? | 16:03 |
| fsmithred | forget what I said about mounting sdb2. You need that for the space. | 16:05 |
| fsmithred | file -L /run/live/medium/live/initrd.img | 16:07 |
| fsmithred | wget https://pkgmaster.devuan.org/devuan/pool/main/r/refractasnapshot-base/refractasnapshot-base_10.4.1_all.deb | 16:13 |
| fsmithred | and then as root, | 16:14 |
| fsmithred | dpkg -i refractasnapshot-base*.deb | 16:14 |
| Guest67 | "don't mount your second partition where you want work_dir and snapshot dir" what do you mean by that? - refracta version is the version that devuan 5 minimal live ships with: 10.3.0 - I have 1gb of ram - file --dereference /run/live/medium/live/initrd.img: Zstandard compress data (v0.8+), Dictionary ID: None | 16:20 |
| Guest67 | fsmithred I'll try 10.4.1 now. | 16:23 |
| fsmithred | right. The part I said about mounting was wrong. You are doing it right. (except for using the version that's not working with zstd) | 16:26 |
| Guest67 | oh ok! | 16:27 |
| Guest67 | then the minimal live iso should contain version 10.4.1 of refracta. | 16:28 |
| fsmithred | should but does not | 16:28 |
| Guest67 | as it seems :) | 16:30 |
| fsmithred | I'm checking the amd64 | 16:31 |
| fsmithred | same | 16:36 |
| Guest67 | Success. Now it is creating the filesystem :) | 16:42 |
| Guest67 | that will take a while on that old pc | 16:42 |
| amarsh04 | It took me 3 weeks to migrate/crossgrade an installation that ran Debian i386 initially and then Devuan to x86_64 - might have been quicker if the Debian wiki had referenced the crossgrader package. | 16:47 |
| Guest67 | fsmithred All finished! :) | 17:08 |
| Guest67 | fsmithred I'll test the iso. | 17:09 |
| fsmithred | cool | 17:10 |
| Guest67 | fsmithred awesome it works! :D | 17:18 |
| Guest67 | fsmithred Thank you very much for your help! :) | 17:18 |
| Guest67 | And lets hope that 32-bit will not be dropped. | 17:18 |
| fsmithred | yw | 17:19 |
| fsmithred | there's no 686 kernel in excalibur. Not that I can find now. I checked when we started talking. | 17:20 |
| Guest67 | I hope they change that again... | 17:20 |
| fsmithred | maybe some third-party will continue to maintain the kernel. | 17:25 |
| fsmithred | not continue - step up | 17:25 |
| Guest67 | are there lists of what packages the devuan isos contain? | 17:34 |
| fsmithred | I don't think so | 17:36 |
| fsmithred | I thought there used to be package lists with the isos, but I don't see any | 17:37 |
| Guest67 | fsmithred yeah cant find them too | 17:37 |
| Guest67 | np i make one with dpkg then :) | 17:37 |
| fsmithred | git has the package list used to build the isos, but that doesn't show all the dependencies that get installed | 17:37 |
| fsmithred | dpkg -l | 17:37 |
| rustyaxe | why does qemu-system-x86 drag in ppc, sparc, riscv, etc?! :( | 18:02 |
| n4dir | rustyaxe: looks like a recommends | 18:07 |
| n4dir | --no-install-recommends; might be worth a try | 18:07 |
| rustyaxe | Ahhhh this machine doesnt have the usual no recommends in apt config.. wonder how that got missed | 18:08 |
| rustyaxe | thanks | 18:09 |
| rustyaxe | someday hopefully that becomes a default | 18:09 |
| fsmithred | qemu-system-gui pulls them in | 18:10 |
| Guest67 | cya | 18:28 |
| fsmithred | bye | 18:28 |
| fsmithred | weird. if I install qemu-system-gui it wants to pull in qemu-systsem-s390x but not the -x86 package. | 18:29 |
| hagbard | probably an or-dependency | 18:36 |
| fsmithred | yeah, it is. a whole string of them. | 18:51 |
| freem | it's still weird since x86 is the 1st in the list and should thus be the one proposed, I think? (and I doubt that recommends will be disabled by default, that would break the "include the world for ease of use" stance of debian) | 21:19 |
| fsmithred | I've seen more people worried that they'll include Suggests by default. | 21:29 |
| freem | that... would really suck yes | 21:36 |
| freem | I don't even understand what the "suggest" thing is about | 21:37 |
| freem | it always felt very... say, artistic, to me | 21:37 |
| freem | I do not dislike art, mind you, but my computers are tools, I prefer them efficient, rather than beautiful, elegant, or suggesting philosophy questions :D | 21:38 |
| freem | s/questions/problems/ | 21:38 |
| rrq | it's worth to note that all qemu-system-blah are emulators of the blah architectures, to run on the host architecture | 23:24 |
| rrq | (just in case someone thought differently) | 23:24 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!