Beer | Anyone having set the ISC dhclient successfully on an interface to set IPv6 up in /etc/network/interfaces in chimaera? | 00:13 |
---|---|---|
Beer | Was working wonders in beowulf, but since the upgrade I can't get the DHCPv6 client behave | 00:13 |
Beer | Specifically, it seems it's associateing the IPv6 address to the interface which doesn't seem to work | 00:17 |
linux_n | Does devuan have a deb for virtualbox? | 01:53 |
golinux | linux_n: https://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=virtualbox&x=submit | 01:58 |
golinux | But . . . iirc . . . it didn' t used to provide the guest-additions. | 02:00 |
linux_n | I have devuan chimaera the link you gave me if i use the one for ceres is it compatible or only works for ceres? | 02:07 |
golinux | ceres = sid | 02:07 |
golinux | So unstable. Would be better to backport it if possible | 02:08 |
linux_n | ok thanks | 02:09 |
jason1234 | Please find the compiled version of ioquake3 for amd64 (quake3): https://gitlab.com/openbsd98324/ioquake3-linux for devuan ascii, with firmware also. it works. | 07:18 |
user282069 | my sound has only dummy output. in dmesg snd_hda_codec_realtek autoconfig for ALC887-VD is misconfigured | 11:01 |
user282069 | i dont know whats going on or what im looking at https://pastebin.com/raw/3Msfqg7w | 11:15 |
devuanized | hi guys, I wondered if it would be possible to install the papirus icon theme and match the folders' color with Deepsea | 11:29 |
devuanized | I would also love to use arc theme with Deepsea scheme | 11:34 |
user282069 | blacklisting snd_hda_codec_realtek does not help. the same misconfiguration then occurs with snd_hda_codec_generic | 11:46 |
user282069 | thank you for the help last night, and any ideas today | 11:48 |
ShorTie | can a postinst be made for the live cd install ?? | 13:55 |
mason | user282069: dummy output sounds like a PulseAudio issue | 17:32 |
user282069 | thank you again mason | 17:33 |
mason | user282069: Look at /etc/pulse/client.conf.d/00-disable-autospawn.conf maybe | 17:34 |
tr-amarsh04 | check this thread "PulseAudio always resets to mute" also: http://dev1galaxy.org/viewtopic.php?id=4002 | 17:45 |
mason | user282069: Oh, and maybe make sure your user is part of the audio group, and then log out/in. | 17:46 |
mason | My solution to PulseAudio issues is to not run PulseAudio, FWIW. | 17:47 |
mason | Videochat via web browsers, Steam, everything's happy with no PulseAudio. | 17:47 |
hyrcanus | mason: + | 18:29 |
hyrcanus | a nice project would be per-app volume control via alsa-only | 18:30 |
hyrcanus | can be done | 18:30 |
buZz | a nicer project is to get pipewire working | 18:42 |
buZz | so you can play steamgames with audio without pulseaudio running | 18:42 |
Tenkawa | buZz: whats the difference in pulse having its own daemon and pipe having one? | 18:48 |
mason | buZz: Eh? You missed where I said I run Steam and have audio, but that I don't run PulseAudio. | 18:49 |
Tenkawa | and for your scenario above to work they would have to rewrite a lot of api output code | 18:49 |
mason | Steam doesn't require PulseAudio. It's happy with ALSA. | 18:49 |
mason | I'm as unlikely to run Pipewire as PulseAudio, honestly. | 18:50 |
buZz | mason: -some- work with alsa only , yeah | 18:51 |
buZz | most seem to very hard require pulse | 18:51 |
mason | buZz: Not from what I can see. Absolutely everything I own works with ALSA. | 18:51 |
buZz | not my experience , but fine by me :) | 18:53 |
mason | buZz: Your configuration was probably wrong. | 18:53 |
buZz | could be , just a default devuan install | 18:53 |
mason | Yeah, try using a dmix. | 18:53 |
debdog | so, no one is gonna mention apulse? | 18:54 |
mason | no need | 18:54 |
mason | anyway: https://bpa.st/BCKQ | 18:54 |
Tenkawa | mason: ahh.. i found an example.. firefox | 18:54 |
Tenkawa | mason: firefox requires pulseaudio now | 18:54 |
mason | The capture half of that is because I want my webcam to be my microphone, but the default audio to be how I emit sound. | 18:54 |
mason | Tenkawa: Since when? | 18:55 |
mason | Tenkawa: Firefox is fine emitting audio with no PulseAudio. | 18:55 |
mason | Tenkawa: Debian builds it this way. | 18:55 |
mason | This is all very strange. | 18:55 |
Tenkawa | mason: not anymore according to what I'm reading.. | 18:55 |
mason | Tenkawa: Did buZz write what you were reading? | 18:56 |
Tenkawa | no I'm reading a bug report | 18:56 |
mason | Just kidding. It works. At least as of Chimaera. If there's something in Sid that doesn't, I can't speak to that. | 18:56 |
buZz | ah, i'm not running chimaera | 18:57 |
Tenkawa | https://unix.stackexchange.com/questions/574394/how-viable-is-it-to-ditch-pulseaudio | 18:57 |
Tenkawa | I can't speak to it myself | 18:57 |
mason | Tenkawa: Reply to the bug with "put your user in group 'audio' and log out and back in" | 18:57 |
Tenkawa | but it does say in the report they removed it | 18:57 |
Tenkawa | https://bugzilla.mozilla.org/show_bug.cgi?id=1345661 | 18:58 |
Tenkawa | it can be compiled in by others (that might be what Debian is doing) but Mozilla's version might not have it | 18:59 |
mason | Tenkawa: This stuff is complex and lots of folks are very confused by it. But no amount of gaslighting is going to convince me that I'm not able to get audio is everything with plain old ALSA. | 18:59 |
mason | Right, it's a build-time option, and Debian supports straight ALSA, so as a result Devuan does too. | 18:59 |
Tenkawa | Mozilla does try to keep theirs non distro centric afterall | 18:59 |
mason | Watching this right now, audio works fine: https://yewtu.be/watch?v=Z58j8c3Er7Q | 19:00 |
mason | Firefox as shipped by Devuan. | 19:00 |
buZz | iceweazel? | 19:00 |
mason | The trademark thing went away years ago, didn't it? | 19:00 |
buZz | bit offensive hostname there mason | 19:00 |
hyrcanus | the problem with firefox isn't pulse | 19:01 |
djph | /whois mason | 19:01 |
djph | ... dammit | 19:01 |
mason | ? | 19:01 |
mason | djph: You can get my PGP key ID from the web site if you need private communications. | 19:03 |
djph | ? | 19:04 |
systemdlete | Anyone familiar with this problem: File systems going readonly, suddenly? I've seen this in the past, but I'm talking years ago. Recently (last several days), I've seen this several times on several VMs (but not the host). I've looked at the host logs and I don't see any hard failures (messages with "ata something" | 19:18 |
systemdlete | Rebooting fixes this because it usually does a fsck on the problem fs. | 19:19 |
systemdlete | I'm wondering if this might be related to a recent kernel update. | 19:19 |
mason | systemdlete: dmesg say anything on the VMs? | 19:20 |
systemdlete | I asked this in -dev also, but since I haven't heard back there, I posted here also | 19:20 |
systemdlete | hmmm. I looked in the syslog on those, but not dmesg | 19:21 |
mason | systemdlete: dmesg is the first place to look for things like this, as if the disks go read-only, nothing can log any more | 19:21 |
systemdlete | good point. | 19:21 |
systemdlete | I will def keep that in mind, mason ! | 19:22 |
mason | You might also think about setting up netconsole to capture log on another system in case you can't log in. | 19:22 |
mason | s/log/&s/ | 19:22 |
mason | https://www.kernel.org/doc/html/latest/networking/netconsole.html for instance | 19:22 |
systemdlete | there's also rsyslog which can do that | 19:22 |
systemdlete | (using port 514, etc?) | 19:22 |
mason | more moving parts | 19:22 |
mason | netconsole is ideal for debugging this kind of thing, but you, maybe you can get it via syslog too | 19:23 |
mason | s/but you/but sure/ | 19:23 |
mason | typing-- | 19:23 |
systemdlete | ah, I see. | 19:24 |
mason | The big thing is that netconsole dumps the kernel message buffer directly | 19:24 |
systemdlete | (brain tells fingers what to type, and fingers still do what they want.) | 19:24 |
mason | Exactly. | 19:24 |
systemdlete | I know! | 19:25 |
systemdlete | all the time here | 19:25 |
mason | I feel like maybe I have some bad buffer RAM in my fingers. | 19:25 |
systemdlete | with me, I think it might be the arthritis | 19:25 |
systemdlete | but I've never been very coordinated really | 19:25 |
systemdlete | I am looking at netconsole man page now | 19:25 |
systemdlete | wow. so it sends the messages directly from the kernel (module)? | 19:26 |
mason | Yeah, rather than being funnelled through syslog | 19:26 |
systemdlete | sure, I see why you suggest that | 19:26 |
mason | Are you seeing this with Beowulf or Chimaera? | 19:27 |
systemdlete | beowulf | 19:27 |
systemdlete | although, I do still have one ascii VM (soon to be upgraded now that chimaera is out) | 19:27 |
mason | Hrm. I haven't seen that at all here, with Beowulf/libvirt/kvm as the hypervisor. | 19:27 |
systemdlete | not sure if I've seen this there or not | 19:27 |
mason | (beowulf as both guest and hypervisor here, although that shouldn't matter) | 19:28 |
systemdlete | the host is beowulf also. But no sign of hardware troubles on the host | 19:28 |
mason | qcow as backing storage? | 19:29 |
systemdlete | I would mention that I'm using vbox, but I really don't want to get into the whole bit about "why?" -- I'm using the kvm mode for all the VMs, so other than some logging vbox does, I don't see why it would make much difference | 19:29 |
mason | Oh. | 19:29 |
systemdlete | yeah, oh. | 19:29 |
mason | That's a variable I can't easily reproduce here. I'd tend to suspect that's the root of the issue. I've had good luck with both Xen and KVM in the past, but not so much with VirtualBox. | 19:30 |
systemdlete | I've had few, if any, problems with vbox itself over the years. So I have continued to use it. | 19:30 |
systemdlete | what kind of problems, if I may ask? | 19:30 |
systemdlete | Other people also tell me similar things, but I have not encountered them. | 19:30 |
mason | It's been a long time. Actually, the issues might not matter since IIRC it was a Mac as hypervisor there. Many years ago. Probably not relevant. But high CPU usage and instability on both guest and host. | 19:31 |
systemdlete | what signs of instability, e.g.? | 19:31 |
systemdlete | I mean, how do you know it is vbox causing it and not something else? | 19:32 |
mason | It's long enough go that I don't really have usable data any more, just impressions of past unhappiness. | 19:32 |
mason | ago* | 19:32 |
systemdlete | yeah, that's the sort of deal I hear from other people. Perhaps vbox went through a bad period -- there were instances I recall where they had to put out an emergency fix right after an update. | 19:33 |
systemdlete | There were a couple of times VMs became unusable after an update. | 19:33 |
systemdlete | but these did not occur often or recently | 19:34 |
systemdlete | I usually hold back from upgrading to x.0 when it comes out. I wait until about x.6 or so and then upgrade | 19:34 |
mason | Maybe check the mailing lists. Less luck of the draw getting someone with relevant experience. | 19:34 |
systemdlete | Let other people suffer the new versions... | 19:34 |
mason | Yeah. | 19:34 |
systemdlete | But, mason. I do that with MOST software anyway... | 19:35 |
systemdlete | it's an old "trick" I learned when I worked for a large financial institution years ago. | 19:35 |
systemdlete | e.g., it was months before I ascii->beowulf | 19:36 |
mason | I often wait on lots of stuff, but I wanted Chimaera earlier than usual here because I no longer have to manually build libsane to drive my newer scanner. | 19:36 |
systemdlete | Hate doing that to peers, but I get frustrated and kind of "sieze up" with those sorts of things. | 19:36 |
systemdlete | there are always instances like that. | 19:36 |
mason | It was particularly unfortunate because the package tools changed, so I couldn't just build the -testing package on Beowulf. | 19:37 |
mason | GNU/Stow to the rescue | 19:37 |
systemdlete | What I would do with that, probably, is set up a VM with chimaera to run sane and redirect requests through the VM. | 19:37 |
systemdlete | It's clumsy, but it insulates the rest of my systems and software here so I can continue to visit my bank and other sites in peace without having to worry I will be missing my money and the rest of my life while the problems get sorted out... | 19:38 |
mason | heh | 19:38 |
systemdlete | then the worst that can happen is I can't scan (or whatever it is I needed) | 19:38 |
systemdlete | I guess a REAL programmer would hunt down the bug with everyone else. | 19:39 |
systemdlete | See, after getting screwed a few times by vbox, linux kernel, and other updates, I have learned to avoid the worst downside risk by taking steps like these. | 19:40 |
systemdlete | Again, thanks for the netconsole tip. I will definitely look into that | 19:40 |
mason | sure, hope it helps | 19:41 |
mason | systemdlete: Note that you almost certainly already have the module available and installed. Comes with the kernel. | 19:42 |
systemdlete | I love free stuff like that! | 19:43 |
buZz | i've had no such experiences with virtualbox ever | 20:06 |
buZz | but i also dont use apple hw | 20:06 |
user3614 | After migrating from Debian buster to Devuan chimaera, the udoubleacute key on my keyboard is interpreted as iacute. How can I change the key back? | 20:23 |
buZz | user3614: maybe > dpkg-reconfigure keyboard-configuration | 20:24 |
buZz | also check /etc/default/keyboard | 20:24 |
buZz | unless you ment X , then maybe look at xkbmap stuff, not sure | 20:25 |
user3614 | I have checked it, and it is the same as before. | 20:26 |
FlibberT1 | hi, any special gotchas with moving my 'test' laptop from chimera to daedalus? | 20:26 |
Tenkawa | that sounds like xmodmap | 20:27 |
Tenkawa | user3614: did you check xmodmap? | 20:27 |
user3614 | The keyboard problem is present even on the console, and even if I boot the desktop-live image. | 20:27 |
Tenkawa | ah its both? | 20:28 |
buZz | maybe you moved from ca-fr to fr-fr ? :D | 20:28 |
user3614 | Does xmodmap also affect the console? | 20:28 |
buZz | it does not | 20:28 |
Tenkawa | thats general kb settings | 20:28 |
user3614 | It is "hu" | 20:28 |
buZz | ah, just a guess :) | 20:28 |
* buZz waves to hungary | 20:29 | |
debdog | user3614: stumbled on setkeycodes today. no clue whether that can solve your issue, it might be worth a look. | 20:31 |
user3614 | debdog, thank you, I am reading its man page now. | 20:32 |
FlibberT1 | no wicd on daedalus? what's to use instead please? | 20:36 |
debdog | FlibberT1: chimaera's relese notes lists some alternatives https://files.devuan.org/devuan_chimaera/Release_notes.txt | 20:37 |
user3614 | FlibberT1: The alternatives are connman, cmst, ifupdown, network-manager. | 20:38 |
FlibberT1 | brill, thanks | 20:39 |
FlibberT1 | really wish i'd uninstalled texlive before running dist-upgrade -- more MB than the OS itself :) | 20:44 |
user3614 | I have figured out the setkeycodes command to restore keyboard settings, thanks to all who have answered me. | 20:55 |
user3614 | One more question: where to add this command, so that it is executed at boot time? | 20:56 |
golinux | FlibberT1: Nice to see you! | 21:16 |
golinux | user3614: Maybe it has something to do with going from buster to chimaera and skipping beowulf | 21:17 |
golinux | buster = beowulf and bullseye = chimaera | 21:18 |
user3614 | golinux: Thanks. I am still wondering why the issue happens with the chimaera desktop-live image, too. | 21:24 |
golinux | user3614: That dev is w/o interent atm. I'll review the issue and give him a quick call. He might have an answer | 21:29 |
user3614 | golinux: Thank you! | 21:29 |
golinux | His suggestion is the same as buZz: dpkg-reconfigure keyboard-configuration | 21:36 |
FlibberT1 | hi golinux, nice to see you too! | 21:49 |
Xenguy | buZz, yewtu.be is an offensive name? I thought yew was a type of tree or wood? | 22:00 |
FlibberT1 | hi, seeing 'does not have a Release file [etc...]' for daedalus-security, daedalus-updates and daedalus-proposed-updates... | 22:02 |
FlibberT1 | when I run apt update | 22:02 |
golinux | Is there security for testing? | 22:10 |
golinux | I think only for stable | 22:10 |
FlibberT1 | ah, might explain it! i was going by the entries in https://www.devuan.org/os/packages | 22:11 |
hyrcanus | FlibbertiGibbet | 22:15 |
FlibberT1 | hey hyrcanus | 22:17 |
FlibberT1 | is it something to do with updating the keyring? | 22:18 |
hyrcanus | no, it's a reference to the movie "Joe vs the Volcano" | 22:21 |
buZz | lol | 22:22 |
FlibberT1 | marvellous :/ | 22:41 |
FlibberT1 | :) | 22:41 |
joerg | sorry for the noise | 22:51 |
Xenguy | likewise I'm sure | 22:55 |
systemdlete | what have I done here? Finsh partitioning, hit ok, screen blanks for a moment, then tells me "Attempt to mount a filesystem with type ext4 ... at '/' failed" | 22:55 |
mason | systemdlete: What context? Mounting something on / from a running system is probably going to fail in most circumstances. | 22:56 |
systemdlete | I've web searched a bit, but I only briefly looked at them. They don't call out to me. | 22:56 |
systemdlete | during install of chimaera (fresh install) | 22:56 |
buZz | Xenguy: sorry for the kick earlier, didnt realize you were a teammember or even in here earlier ;) | 22:56 |
mason | systemdlete: Which media? | 22:57 |
systemdlete | devuan chimaera netinstall amd64 | 22:57 |
mason | Ah, I just snagged the netinstall image the other day - legacy install? What partitioning did you select? | 22:58 |
systemdlete | expert install | 22:58 |
mason | systemdlete: And, VM? Hardware? | 22:58 |
systemdlete | VM | 22:58 |
systemdlete | what do you mean by "legacy?" | 22:58 |
systemdlete | Is that a cut about how old I am? | 22:59 |
systemdlete | :D | 22:59 |
mason | systemdlete: I hit an anomaly gathering screenshots the other day and while it haven't tracked it down, I haven't ruled out that it was because of the virtio drivers presenting vd instead of sd | 22:59 |
buZz | non-uefi , i guess | 22:59 |
mason | systemdlete: MBR/BIOS partitioning, as opposed to GPT | 22:59 |
systemdlete | yeah. DOS disk partitioning with one primary partition | 22:59 |
mason | systemdlete: Mostly I want to know what to install to reproduce what you saw. | 22:59 |
mason | kk, that's easy enough then | 22:59 |
mason | I'll give it a try. | 23:00 |
systemdlete | BUt | 23:00 |
systemdlete | I am using lvm | 23:00 |
systemdlete | I've created encrypted partitions including root (/) | 23:00 |
systemdlete | (which seems to be why it is belly-acheing at me) | 23:00 |
systemdlete | I have this same arrangement on beowulf, refracta and even starlinux | 23:01 |
systemdlete | (I've tried them all, just for the heck) | 23:01 |
systemdlete | but those are all beowulf/spock releases | 23:01 |
systemdlete | this is the first time I am trying to install chimaera with this partitioning. | 23:02 |
mason | systemdlete: Oh..... That might be a bit of an issue. Unless something's changed, GRUB needs a boot partition if root is encrypted. | 23:02 |
systemdlete | It has a /boot (unencrypted) | 23:02 |
mason | Oh, you said one primary. | 23:02 |
mason | What's the full layout? | 23:02 |
systemdlete | primary, as in the MSDOS disk partitioning (grub stuff) | 23:02 |
systemdlete | separate partitions of about 2g for var, tmp, home, root, swap (all encrypted) | 23:03 |
systemdlete | and usr 6g unencrypted and /boot about .5g unencrypted | 23:03 |
systemdlete | One 16gb disk will do | 23:04 |
mason | systemdlete: So, /boot unencrypted, and the rest of those inside encrypted LVM? | 23:04 |
systemdlete | no | 23:04 |
mason | Two partitions? | 23:04 |
systemdlete | /usr is unencrypted | 23:04 |
systemdlete | they are all lv's | 23:04 |
hyrcanus | why is /tmp in ram. can we not do that. | 23:05 |
systemdlete | probably 7 or 8 partitions | 23:05 |
Xenguy | buZz, Just seeing this now... | 23:05 |
hyrcanus | i mean i disabled it. but ... as a default | 23:05 |
mason | hyrcanus: You can use tmpfs for /tmp | 23:05 |
Xenguy | I am sorry as well, I did not intend to cause offence | 23:05 |
Xenguy | buZz, ^^ | 23:05 |
hyrcanus | yeah that's bad | 23:05 |
mason | Xenguy: As noted, not your fault. | 23:05 |
hyrcanus | maybe it was a remnant of my previous system | 23:05 |
joerg | while I'm generally for a strikt policy enforcement regarding sustaining a welcoming non-discriminating non racist non-sexist etc pp (younameit) mood in all IRC channels, I'm sure a simple /kick for a question about a word that is three letters and two of them are common with jew as well as with you(tube) is NOT the way to achieve this | 23:05 |
buZz | right, it just appeared like a join and question out of the blue | 23:06 |
buZz | about something that wasnt even a issue initially, so felt aimed at causing issues | 23:06 |
mason | systemdlete: I'm struggling to come up with an adequate mental map of partitions vs logical volumes, and in particular volume groups and PVs for what you're installing. Could I bother you to pop a map onto bpaste.net or similar? | 23:07 |
buZz | all i said was a hostname was a bit offensive | 23:07 |
systemdlete | mason: I would but | 23:07 |
systemdlete | this is in vbox, and I don't have a way to sort of copy and paste it | 23:07 |
mason | systemdlete: No need. I'll give you an example. | 23:07 |
systemdlete | One large Primary partition in fdisk say | 23:07 |
systemdlete | then | 23:07 |
systemdlete | one pv | 23:07 |
systemdlete | one vg on that one pv | 23:08 |
systemdlete | and the rest of above are lv's | 23:08 |
mason | systemdlete: And LUKS? | 23:08 |
systemdlete | 2 are unencrypted, the rest are encrypted | 23:08 |
systemdlete | yes, luks | 23:08 |
Xenguy | buZz, I wasn't sure why, so I just asked to find out. I use that alternative to youtube all the time, so I wondered if it was a 'bad name' or something | 23:08 |
mason | systemdlete: Are you saying you've got LUKS *inside* the LVs? | 23:08 |
systemdlete | (for the encrypted) | 23:08 |
systemdlete | yes | 23:08 |
systemdlete | the installer lets me do that | 23:08 |
systemdlete | there are 2 ways to encrypt, apparently. You can either encrypt the raw disk, or you can encrypt at the lv level | 23:09 |
mason | systemdlete: FWIW, a super common and far simpler layout is: two partitions, one for /boot (ext, unencrypted) and one for LUKS. Inside that LUKS is the PV holding your volume group, and then root, swap, and whatever else you want inside of that. That's well-supported by the installer as well. | 23:09 |
buZz | Xenguy: its a 'bit' because its not 1:1 , its sounding alike, some people experience IRC through a screenreader only | 23:09 |
systemdlete | mason: I've done it that way in the past. But this time, I wanted to be able to do it all from the lvm level. | 23:10 |
mason | systemdlete: The installer will let you set up arbitrarily complex combinations, but they don't always work out as we might wish once it's all installed. | 23:10 |
buZz | nothing wrong with invidious at all , imho | 23:10 |
systemdlete | mason: Yeah. I've noticed... LOL | 23:10 |
mason | Maybe the discussion of Invidious could move to #devuan-offtopic. | 23:10 |
systemdlete | mason: There is one leetle thing I did not tell you... | 23:10 |
mason | systemdlete: The issue with what you've got might revolve around GRUB still. Not sure how happy it is loading up LVM to find /boot. | 23:11 |
systemdlete | I had originally tried the layout you are suggesting, then went back in to the partitioning step with this configuration | 23:11 |
systemdlete | grub is not even installed yet. I'm still at partitioning | 23:11 |
systemdlete | mason: Maybe a reboot will clear the installer's head? | 23:12 |
systemdlete | is there a way to save the state of the install with, say, a thumb drive? | 23:12 |
mason | systemdlete: Well. Yeah, you can pop one in and write out partition, LUKS, and LVM state to files on a thumb drive from a virtual console. | 23:14 |
systemdlete | and then restore them after reboot? | 23:15 |
mason | systemdlete: Recommendation: include this command output: blkid, pvs, vgs, lvs | 23:15 |
systemdlete | mainly, I want to avoid having to enter the same password umpteen times for the encrypted partitions | 23:15 |
mason | systemdlete: Well. I'd think about doing LUKS differently. | 23:15 |
systemdlete | the way I did previously... | 23:15 |
systemdlete | which works and doesn't make the installer choke to death | 23:15 |
mason | And is a bit simpler. | 23:16 |
systemdlete | meh | 23:16 |
mason | Fewer keys to unlock on boot anyway. | 23:16 |
systemdlete | I don't find it much more complicated really. | 23:16 |
systemdlete | actually, it dawns on me I could have one big encrypted LV, and then do bind mounts for the individual filesystems... but again, I don't know if the installer would understand that | 23:17 |
systemdlete | either. | 23:17 |
mason | The installer won't do the bind mounts, but you can just have multiple volumes and that works. | 23:17 |
systemdlete | which is what I have, yes | 23:17 |
systemdlete | I mean, what you are telling me to do is precisly the way I've done it in the past for devuan | 23:18 |
systemdlete | I was hoping to use a slightly different strategy | 23:18 |
mason | That could well be. It's super common. Might even be what you end up with with guided partitioning if you opt for encryption. | 23:18 |
systemdlete | I think I'll just do this whole blasted thing over again. Maybe this time the repartitioning of the disk won't interfere so much. | 23:19 |
systemdlete | nah. I'm stubborn enough. And heaven knows, I've got the time to dick around with this. | 23:19 |
mason | systemdlete: You can do it. Just be aware of boot ordering. | 23:21 |
systemdlete | there is only one OS in this VM | 23:21 |
systemdlete | or do you mean something else? | 23:21 |
systemdlete | one disk and one OS | 23:22 |
mason | systemdlete: No, I mean, the system processes LUKS, then LVM, then MD, all in some order, with a chance for late entries, but it's possible to create a logically valid stack that's not actually supportable with the default software. | 23:22 |
systemdlete | I've never run into an issue with that in a VM, except one time when I was experimenting with RAID in a VM -- to simulate a real-world HARDWARE problem I was having. (btw, I WAS able to replicate it. very handy!) | 23:23 |
systemdlete | oh | 23:23 |
systemdlete | I see. Yeah, that has occurred to me. | 23:23 |
systemdlete | I would likely have to craft a hand-made boot image, etc etc etc etc etc | 23:23 |
systemdlete | mason: My bucket list for partitioning/LVM schemes -- the ability to encrypt a Volume Group, as opposed to individual LV's or underlying harddisk partitions. | 23:24 |
systemdlete | but that's just dreaming | 23:24 |
mason | systemdlete: If you read through the scripts you can see which block device providers can be early, late, etc. You can also give hints with, for instance, the initramfs flag in crypttab. | 23:25 |
systemdlete | thanks, I forgot about that flag | 23:25 |
systemdlete | I've tried to keep this all reasonably simple and probably why it has worked. | 23:25 |
mason | systemdlete: Is it possible to have a PV associated with more than one VG? Or do you mean you want the encryption handled by LVM itself somehow? | 23:26 |
judabuda | hello, can minimal-live be used to install minimal devuan installation? | 23:27 |
systemdlete | IOW, just as LVM currently has the ability to encrypt LV's (but ONLY LV's), if there were some way to extend that logic to the entire volume group. I mean, all it really needs to do is what it currently does, but just allow admins/installe personnel to specify that they want all LVs on the VG to be encrypted. That way all of them could be associated with the same key, etc | 23:27 |
systemdlete | I thought I had actually done that a few months back, experimenting, but I really don't remember now. | 23:28 |
mason | judabuda: yes | 23:28 |
systemdlete | I am not sure if you can associate one PV with multiple VGs. I've never tried it. | 23:29 |
systemdlete | the problem is that there is no device, per se, associated with a VG, only with indivdual LVs. So the cryptsetup has no way to associate a VG to encrypt. I guess that is what would have to change. But then the encryption subsystem has to get smart about LVM, and that is probably why such a thing would never happen. | 23:32 |
systemdlete | unless LVM itself were to implement cryptography. | 23:33 |
systemdlete | mason: Thanks for your feedback and help. You have convinced me to go back to my old approach. Too many cycles being wasted on this. | 23:34 |
mason | cool, happy to help - I need to go feed the chickens some bread, but I'll be back later | 23:34 |
* systemdlete is feeling a bit chicken himself, actually... | 23:35 | |
judabuda | did devs add trinitydesktop.org to devuan installation, i heard they will it add it in advanced installation | 23:42 |
judabuda | ?? | 23:42 |
Hydragyrum | it doesn't appear to be in ceres package repos yet | 23:47 |
judabuda | what is best way to install xfce, install other DE and remove xfce packages so only other DE packages are kept? | 23:53 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!