LeePen | Does anybody have a beowulf system with wifi that they could test wicd on, please? | 12:11 |
---|---|---|
LeePen | See #464 | 12:12 |
fsmithred | LeePen, you want to know if wpa-aes works? | 12:24 |
fsmithred | or some other test? | 12:24 |
LeePen | fsmithred: Well, he says no ESSIDs are found and WPA-AES is not supported. | 12:31 |
LeePen | I don't have any harware to test either. | 12:31 |
LeePen | Is it broken or is he missing something? | 12:33 |
fsmithred | I thought I've seen WPA-AES listed in wicd, but I don't see it now. | 12:33 |
LeePen | Looking at https://tracker.debian.org/pkg/wicd, it is in a bad way in Debian. | 12:40 |
LeePen | Missing from testing. Stuck on python2 | 12:40 |
LeePen | There is an experimental build from December. | 12:42 |
fsmithred | ok, here's the report... | 12:45 |
fsmithred | my router is set to WPA2-PSK (AES) | 12:45 |
fsmithred | my wicd is set to use WPA1/2(Passphrase) | 12:46 |
fsmithred | it works (fsr can enter this channel but can't speak) | 12:47 |
LeePen | And is wicd or wpa mentioned in /etc/networking/interfaces? | 12:47 |
fsmithred | checking... | 12:49 |
fsmithred | just lo | 12:50 |
fsmithred | I have to find "configuring an ad hoc network" | 12:50 |
onefang | Anything in /etc/networking/interfaces.d? | 12:51 |
fsmithred | no | 12:51 |
fsmithred | ok, I found that window | 12:51 |
fsmithred | only one checkbox for encryption, and the choice is WEP | 12:52 |
fsmithred | what is an ad hoc network? It has a default ip address of 169.x.x.x | 12:52 |
LeePen | Hmmm, so it only does WEP for ad hoc. | 12:53 |
daniel_hozac | it's a direct link between two computers, without an AP. | 12:54 |
LeePen | Ad hoc is between any 2 wifi stations, without one being a master. | 12:54 |
LeePen | I wonder if he has wpa-supplicant | 12:54 |
LeePen | fsmithred: Can you try to create a template for it? http://wicd.sourceforge.net/templates.php | 12:57 |
fsmithred | template for what? | 13:00 |
LeePen | fsmithred: nw. | 13:04 |
LeePen | Just looked at the source for wicd. It assumes WEP for ad-hoc networks. | 13:04 |
LeePen | Anbody fancy trying to fix it? | 13:05 |
fsmithred | not sure it's broken | 13:05 |
fsmithred | he might just need to tell wicd that wlan0 is the default wireless interface | 13:06 |
fsmithred | then he'll see a wireless network and be able to find the encryption settings | 13:06 |
fsmithred | still won't see AES, but it does work for regular wireless connection, which is probably what he wants. | 13:07 |
LeePen | fsmithred: Thanks. Can I leave you to reply to the bug? I don't have any suitable hardware to be able to comment, test or give instructions. | 13:07 |
fsmithred | if I ever get the email for that bug, I'll reply | 13:08 |
fsmithred | 461 is the last one I got | 13:08 |
LeePen | It is because Katolaz is still listed as maintainer! | 13:09 |
fsmithred | it should still show up in the bug mailing list, shouldn't it? | 13:09 |
LeePen | You can get it here as a mbox: https://bugs.devuan.org/cgi/bugreport.cgi?bug=464;mbox=yes;msg=5 | 13:09 |
LeePen | It should still come on devuan-bugs, but I am not convinced that everything does. | 13:12 |
LeePen | Jaromil: Are you able to look at the devuan-bugs mailman config? Some bug traffic doesn't appear on the list. | 13:14 |
fsmithred | reply to 464@bugs.devuan.org will work? | 13:16 |
LeePen | And include the OP, please. | 13:16 |
fsmithred | ok | 13:23 |
LeePen | Thanks. | 13:25 |
LeePen | The same poster also had a problem with iwlwifi firmware in the installer iso. | 13:27 |
LeePen | #462 ^^^ | 13:27 |
LeePen | fsmithred: was that the same firmware you had problems with? | 13:28 |
fsmithred | yes | 13:28 |
LeePen | But you found a woraround? | 13:29 |
fsmithred | also noticed this morning that if I install it when running, I have to remove/insert module before it works | 13:29 |
fsmithred | it's possible for me to get wireless on that computer in the installer, but it's a big pain in the ass to do it | 13:29 |
fsmithred | alt-f2, delete some files in /tmp, then rerun hardware detection | 13:30 |
fsmithred | I think that's what it was | 13:30 |
LeePen | Can you reply to #462 and explain that, please? | 13:31 |
LeePen | If it is just that firmware, should we add something about it to the Release Notes. | 13:32 |
LeePen | At least that way we can say 'we told you'? | 13:32 |
fsmithred | don't know if it's just that one | 13:51 |
fsmithred | I did also test a box that uses firmware-atheros, and that one worked | 13:52 |
fsmithred | no clue about others | 13:52 |
fsmithred | also not sure if iwlwifi fails for everyone | 13:53 |
LeePen | This seems like a second, at lease. | 13:53 |
LeePen | ..least. | 13:53 |
gnu_srs | I have laptop running Beowulf with a working wlan0. In /etc/network/interfaces: auto wlan0; iface wlan0 inet dhc; wpa-driver wext; wpa-ssid ... ; wpa-psk ... | 14:32 |
LeePen | gnu_srs: That is good to know. Thanks. | 14:55 |
fsmithred | beta3 netinstall won't let me do wireless installation. | 17:13 |
fsmithred | it's back to being stuck in the loop asking for essid, encryption type and password | 17:13 |
fsmithred | This is on the toshiba with realtek wireless, not on the thinkpad with iwlwifi that had problems before. | 17:13 |
MetaYan | golinux: On https://beta.devuan.org/os/packages "Default configurations" "Devuan 3.0.0 Beowulf (stable)" beowulf-updates is listed, but further down, under "Additional repositories" it says "Add stable updates for fast release packages (default: no)" | 17:29 |
MetaYan | Also, all the examples use http, while - at least when building with arm-sdk - https is put in /etc/apt/sources.list | 17:37 |
mason | fsmithred: What's a URL for the beta3 netinstaller? I can give it a spin on a couple boxes here, at least far enough to see if it can get the network going. | 17:43 |
fsmithred | they're all here: https://files.devuan.org/devuan_beowulf/installer-iso/ | 17:50 |
fsmithred | MetaYan, what repo is put in the arm sources.list? If it uses deb.devuan.org, there will be problems. | 17:51 |
fsmithred | not all the mirrors support encryption | 17:51 |
mason | fsmithred: Oh, duh. Right. Downside to only ever doing debootstrap installs, I guess. | 17:53 |
fsmithred | lol | 17:53 |
fsmithred | upside is that you avoid installer bugs | 17:53 |
mason | Yeah. | 17:53 |
MetaYan | fsmithred: arm-sdk put https://pkgmaster.devuan.org/merged | 17:53 |
fsmithred | btw, for the live isos I tried some things with udevadm to get mouse and keyboard working | 17:54 |
mason | Any luck? | 17:54 |
fsmithred | MetaYan, pkgmaster will work with https, but we discourage its use so the server doesn't get overloaded | 17:54 |
fsmithred | udevadm control -R | 17:55 |
fsmithred | udevadm trigger | 17:55 |
fsmithred | udevadm settle | 17:55 |
fsmithred | did nothing | 17:55 |
mason | Most of my mucking with udev has been for storage handling, and is made more painful by the ubiquity of udisks2, which I've grown to hate. | 17:55 |
MetaYan | fsmithred: All right. Will use http. | 17:55 |
mason | hrm, sorry to hear it - was hoping | 17:55 |
fsmithred | yeah, I'm pretty frustrated with it | 17:55 |
mason | I'll feel more free to help debug things when I've gotten the new server stuff (twiki, mailman3) off my plate. | 17:56 |
fsmithred | ok, thanks | 17:56 |
mason | So, T420 using beta3 netinstaller, first thing I see is a note about missing firmware for iwlwifi, and an offer to load it if I have it on removable media. And this rings a bell - Debian can't install over wifi without my using the "firmware" ISOs for just this reason. | 18:02 |
mason | fsmithred: ^ | 18:02 |
fsmithred | that shows up early | 18:02 |
mason | Yes. | 18:02 |
fsmithred | that's for when you have the firmware files on usb | 18:02 |
fsmithred | say no | 18:02 |
mason | Ah, thought you were saying it was surprisingly early. Trying again so I can note what was just before it. | 18:03 |
fsmithred | this isn't debian - we put the firmware on the iso | 18:03 |
mason | But I suspect it'll fail - Debian couldn't proceed without the firmware. | 18:03 |
fsmithred | RMS can hate us for not hiding non-free | 18:03 |
mason | Oh, so it's there but doesn't notice it? | 18:03 |
mason | heh | 18:03 |
mason | Yeah, pragmatism has to win out sometimes. | 18:03 |
fsmithred | that window is useful if you have certain broadcom wireless | 18:03 |
fsmithred | and have to install without ethernet | 18:04 |
fsmithred | some broadcom firmware packages only contain an installer which must download files from the internet | 18:04 |
fsmithred | catch 22 | 18:04 |
mason | Oh, argh, and I brainlessly tapped through again. But I'll proceed saying "no" - pretty sure it was just language selection before this. | 18:05 |
fsmithred | it should ask you to insert media, and you should be able to bypass that or back out | 18:05 |
mason | tries eth0 first, seems to be trying dhcp without seeing a link | 18:05 |
fsmithred | yeah | 18:05 |
mason | network autoconfig failed, and it's acting as though it doesn't see wlan0 | 18:06 |
fsmithred | it didn't try to connect to wlan0? | 18:06 |
fsmithred | ok, want the workaround? | 18:06 |
fsmithred | alt-F2 then Enter | 18:07 |
fsmithred | rm /tmp/* | 18:07 |
fsmithred | alt-F1 and Go Back to network detection | 18:07 |
mason | Yeah, and now it sees wlan0. | 18:08 |
mason | Interesting. | 18:08 |
mason | Sees my access point, so it's loaded the firmware properly. | 18:08 |
mason | So, we need to load firmware earlier? Maybe seeing how the Debian firmware* images do it would be useful. | 18:08 |
fsmithred | good. It acted the same way twice. I like consistent bugs. | 18:08 |
mason | Yeah. | 18:09 |
fsmithred | debian don't load any non-free firmware | 18:09 |
mason | That's what the firmware- images do, isn't it? | 18:09 |
mason | Explicitly I thought. | 18:09 |
fsmithred | oh yeah, I forgot they had those | 18:09 |
mason | https://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/10.4.0+nonfree/amd64/bt-cd/ | 18:09 |
mason | and similar | 18:09 |
mason | So, we're no more reprehensible than Debian here if we include it. :) | 18:10 |
fsmithred | you can go to some strange places looking at the free thing | 18:10 |
fsmithred | e.g. libre kernel restricts your freedom | 18:11 |
mason | Hm, let me try this on my work laptop too, for comparison. X1C6 | 18:11 |
fsmithred | cool, thanks | 18:11 |
mason | At risk of going off-topic, another great example is the FSFer stance on ZFS, where they fight tooth and nail to keep people from using software that's not just free, but also copyleft. | 18:11 |
mason | Oh, heh, I need to actually reinstall this X1C6 anyway. Have an old RHEL on there to debug a customer issue that needed hardware. | 18:13 |
mason | Ah, and we don't boot on it out of the box. Got to turn off SecureBoot. | 18:14 |
fsmithred | ??? | 18:14 |
fsmithred | it should boot with secure boot | 18:15 |
mason | Hrm. It doesn't though. Honestly I didn't realize it was supposed to. | 18:15 |
mason | I know Buster picked up SecureBoot support. I guess it stands to reason we'd have it transitively. | 18:16 |
mason | language, language, language, detect cd, firmware message | 18:16 |
fsmithred | that sounds right | 18:16 |
mason | saying "no", waiting to go back to menu | 18:16 |
mason | F2, clear state from /tmp, back to detect, and ... | 18:17 |
mason | now it sees wlan0 | 18:18 |
fsmithred | cool | 18:18 |
mason | So, there's a third confirmation. | 18:18 |
fsmithred | I like consistent workarounds, too | 18:18 |
mason | Sorry for the dramatic pause. It took that long. :P | 18:18 |
fsmithred | it's ok, I'm doing more than one thing here | 18:18 |
mason | And, there, it sees my bssid. | 18:18 |
mason | bbiab, checking packages that just arrived - I think the latest was sliced cheese from a restaurant supply store | 18:19 |
fsmithred | oh, it only has the one efi bootloader - bootx64.efi - it should have some others whose names I've forgotten | 18:19 |
fsmithred | yum | 18:20 |
mason | Yeah, there's grub, then the shim loader, etc. | 18:20 |
mason | shimx64.efi, possibly mmx64.efi | 18:21 |
mason | brb anyway | 18:21 |
fsmithred | I just checked several isos going back to Feb, and I can't find one that supports secure boot. But I know I saw those files in some iso around a couple months ago. | 18:27 |
gnu_srs | BTW: On Beowulf the grub flash screen shows Debian GNU Linux ... Not Devuan GNU Linux. | 18:46 |
fsmithred | gnu_srs, yes, we changed to ID itself as debian for a couple reasons | 18:50 |
fsmithred | it allows secureboot to work | 18:50 |
fsmithred | it allows certain third-party software to install, because not everyone knows we exist | 18:50 |
fsmithred | and we might change it back (maybe already have) in ceres/chimaera to see what breaks | 18:51 |
mason | I was more bothered by that until I decided that the best future for Devuan is to be a level that shifts Debian and maybe others. | 18:51 |
mason | lever* | 18:52 |
fsmithred | I hope we can have some influence on those poor, misguided souls. | 18:52 |
mason | Being successful and reliable is the way to do it. | 18:52 |
fsmithred | yup | 18:53 |
mason | My biggest worry is that the project be perceived as a community of "I hate thing". Needs general appeal to thrive. | 18:53 |
mason | But that's political so I'll not go on at length. :P | 18:53 |
mason | Oh, random note, I remembered what my last install on the X1C6 was for - it was to work through a udev bug, ironically. | 18:54 |
fsmithred | mason, have you looked at the discussion I'm having on the forum about this issue? | 18:59 |
mason | fsmithred: No, I dislike forums enough that I avoid them unless pressed. | 19:00 |
fsmithred | There may be some clues there that I'm not seeing. | 19:00 |
mason | I can make an exception and look. | 19:00 |
mason | fsmithred: Which forum / topic? | 19:01 |
fsmithred | https://dev1galaxy.org/viewtopic.php?pid=21671#p21671 Start here, ignore the part about lvm and the repeated warnings - that part is fixed. Trying to get mouse and keyboard on desktop is an issue. | 19:02 |
mason | kk - and wow, the forums | 19:02 |
fsmithred | thanks | 19:02 |
fsmithred | I won't tell anyone you were there. | 19:03 |
mason | Might be too late. I think I left some of my breakfast on a couple of the posts. Which image do I want to use to observe this stuff? | 19:03 |
mason | I see devuan_beowulf_3.0.0_beta2_amd64_desktop-live.iso mentioned early on - I'll assume the beta3 equivalent is good to check. | 19:04 |
fsmithred | hang on - I made another one | 19:05 |
mason | kk | 19:05 |
fsmithred | I made two | 19:05 |
mason | Where can I find them? | 19:06 |
fsmithred | this one is a refracta iso: ssh server is running in case you want to access from another computer and also has runlevel 3 set for no display manager: http://distro.ibiblio.org/refracta/files/experimental/refracta10-beta3_xfce_amd64-20200515_1335.iso | 19:06 |
mason | kk, snagging | 19:06 |
fsmithred | this one is devuan desktop-live: https://get.refracta.org/files/experimental/devuan_beowulf_3.0.0_beta5_amd64_desktop-live.iso | 19:07 |
mason | Same issue either way? | 19:07 |
fsmithred | yes | 19:07 |
mason | And is it just USB keyboard/mouse? | 19:07 |
fsmithred | yeah | 19:07 |
mason | kk, I'll fish out some hardware | 19:08 |
fsmithred | test with something that doesn't have a trackpad | 19:08 |
fsmithred | with trackpad, if you remove/insert psmouse module, it works | 19:08 |
mason | Oh, I can test on my four-year-old's system. That'll be perfect for this. | 19:08 |
fsmithred | if only usb inputs, no change | 19:08 |
fsmithred | unplug/replug mouse or keyboard also fixes it | 19:08 |
mason | Hm, looks like a bit more than half an hour to finish the download - not seeing stellar speeds, but I'll write it out once it's done. | 19:09 |
fsmithred | yeah, I put one on ibiblio because the refracta.org upspeed was really slow when I did it. | 19:10 |
fsmithred | was going to take 14 hours | 19:10 |
fsmithred | but that one is usually faster | 19:10 |
fsmithred | refracta.org is usually faster | 19:10 |
mason | Started off seeing 600K/s, bouncing between 400-550 now. | 19:11 |
mason | Might watch a movie while it chugs along, and then it'll take no time at all! :) | 19:12 |
fsmithred | oh, and note that the problem occurs when booting from DVD, not usb | 19:13 |
mason | Ooh. | 19:17 |
mason | Ah, kk, system in question has an optical drive, and I think I have some DVD-Rs around. | 19:18 |
mason | fsmithred: So, it hasn't downloaded, but that's odd if it's just DVD media and not the same thing on USB. Makes me think it has to be something revolving around persistance, where some things can be written on on USB but not DVD. Maybe not, but that's a really odd proviso. | 19:31 |
fsmithred | can't write to usb if you image with isohybrid - it's read-only | 19:32 |
fsmithred | also problem doesn't occur if you boot toram | 19:32 |
fsmithred | also might be related to framebuffer, but I don't understand that stuff | 19:33 |
mason | I'd think that'd be the same. And the "toram" thing makes me think ability/inability to write things matters, but... I'll see if I can reproduce something once this finishes. | 19:35 |
fsmithred | thanks | 19:35 |
fsmithred | I also get it sometimes when booting the iso in qemu | 19:36 |
mason | Oh, that'd be ideal. | 19:36 |
fsmithred | display size is wrong and mouse and keyboard don't work | 19:36 |
mason | Maybe I'll try that before committing plastic to the process. | 19:36 |
fsmithred | then I kill the window, boot it again and it works | 19:36 |
fsmithred | but that's not as consistent as optical media | 19:36 |
mason | Yeah. But also quicker to cycle. :) | 19:37 |
fsmithred | and I put a hack fix in both of those isos - modprobe -rv psmouse && modprobe -v psmouse | 19:37 |
fsmithred | in /etc/rc.local | 19:37 |
mason | Alright, I'm down below 200K/s now. Going to stop watching the download, but I'll come back and try it until I get it to fail, and then explore. | 19:37 |
fsmithred | fixes it on laptop, but only because of the trackpad | 19:38 |
fsmithred | enjoy movie | 19:38 |
mason | Oh, might also be worth exploring xinput | 19:38 |
mason | fwiw | 19:38 |
mason | I use it to disable/enable things based on xscreensaver triggers here. Might be worth seeing if it still sees the devices when they're unresponsive. | 19:39 |
mason | anyway, bbiab | 19:39 |
fsmithred | thanks | 19:39 |
golinux | Evilham: rrq set up the new-beta website 12 or so hours ago. But the commits that both he and I made are again not being passed to the live site. | 19:46 |
golinux | Could you please have a look? | 19:47 |
golinux | The new site is now being served on beta.devuan.org | 19:47 |
fsmithred | https://termbin.com/z4gt6 mason, it sees the hardware | 19:50 |
Evilham | the thing was ran with the wrong user so it had the wrong permissions all over the place | 19:57 |
Evilham | there is a README and a Makefile explaining how things work on the server | 19:59 |
golinux | Now the site is a 404 | 20:03 |
bgstack15 | ditto | 20:03 |
bgstack15 | was just coming here to say that | 20:03 |
golinux | It was working for rrq and myself I guess because we are both admins | 20:03 |
golinux | (Evilham is the best at triage) | 20:05 |
Evilham | it's a beta site | 20:06 |
Evilham | it's allowed to be foobared for a bit | 20:06 |
golinux | Indeed. This is a major change. rrq said something about the python being another rat's nest that you could likely simplify. | 20:07 |
Evilham | it's already simple enough, nobody has had to think about how middleman works in years | 20:08 |
Evilham | is there a reason why there is now a new-beta branch? | 20:09 |
Evilham | am I understanding correctly that new-beta contains HTML that should be served as-is? | 20:11 |
Evilham | sorry, that change is out of my understanding, if new-beta has a directory that should be served as is, then the python thing is useless because it was to not have to deal with middleman ever (it was very successful at that) | 20:14 |
Evilham | which means that the deployment is basically git pull | 20:14 |
Evilham | in particular that also means that there are a bunch of files that should be deleted from that branch | 20:15 |
Evilham | I have restored previous behaviour (branch beta.devuan.org --> beta.devuan.org, branch master --> devuan.org) | 20:16 |
Evilham | and for the new page it'd be a matter of ensuring that the repository is updated and the page switched in an atomic fashion | 20:16 |
Evilham | so, it's a totally different deployment | 20:17 |
fsmithred | what command can I run that will make the computer believe that I just unplugged and re-plugged my usb mouse or keyboard? | 20:26 |
fsmithred | and while I'm here, I just want to say that running xinput over ssh to disable the keyboard did NOT do what I expected it to do. | 20:35 |
fsmithred | In case anyone was wondering why I rebooted. | 20:36 |
fsmithred | and as a warning to mason not to do the same thing | 20:36 |
fsmithred | 'modprobe -r usbhid && modprobe usbhid' works. I did it over ssh, once while the desktop was running and once after stopping display manager. It works either way. | 20:42 |
golinux | Evilham: Sorry been afk in the kitchen. | 20:45 |
golinux | Soooo . . . I went to beta.devuan.org and it is the "new-beta" site with the changes in place afaict | 20:47 |
Evilham | as I mentioned: I have restored previous behaviour (branch beta.devuan.org --> beta.devuan.org, branch master --> devuan.org) | 20:48 |
Evilham | new-beta is not being used, which seems to have been the intent | 20:48 |
Evilham | but that's a totally different pipeline and fully by-pass the current deployment, because that deployment is to make middleman bearable | 20:49 |
golinux | Perhaps rrq made those changes also on the beta branch. | 20:49 |
golinux | He'll be up in a few hours . . . | 20:50 |
golinux | This understanding is beyond my pay grade. | 20:50 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!