rwp | Problem: Ceres just updated cron-daemon-common which contains /etc/crontab which sets PATH and just removed /bin from the PATH statement and then called run-parts which is located in /bin on non UsrMerged systems. | 01:43 |
---|---|---|
rwp | Resulting in email "Subject: Cron <root@turmoil> cd / && run-parts --report /etc/cron.hourly" containing "/bin/sh: 1: run-parts: not found". | 01:44 |
rwp | I suppose I should post this to the dev channel. Will do that. But heads up about the problem. | 01:44 |
systemdlete | So I completed a fresh install of daedalus on a testbox here and, overall, pretty smooth. It boots and I don't notice anything unusual. I moved my chimaera and beowulf partitions to new locations in the process, so that presented some challenges, mainly with grub. | 02:08 |
systemdlete | The os-prober (in daedalus) was able to detect chimaera, and after some futzing and mutzing using systemrescuecd, I am now able to boot into both chimaera and daedalus successfully from the grub config created by the deadalus install. | 02:09 |
systemdlete | But I am still struggling with booting into the beowulf install. I can boot into beowulf using the rescue cd, but for some reason, os-prober (from daedalus) doesn't seem to detect the beowulf partition. | 02:10 |
systemdlete | Again, it is my own doing--to some degree--that I chose to move that partition in the process of reconfiguring the system. And granted this really isn't all that important. I doubt I will even need beowulf even for testing going forward. But I thought I would ask just to see if this is a known issue, that sort of thing. | 02:12 |
systemdlete | In sum, very nice work, so I wanted to take a moment to thank you all for pulling this off as smoothly as you have previous releases and for the superb support I always get considering the challenges posed by the haters of "old technology" like sysvinit, and probably the wheel and fire also. | 02:13 |
brocashelm | systemdlete: what kernel is the beowulf on? i think 5.10 is the most you can get on there with backports. have you tried that one? | 02:13 |
systemdlete | Give me a moment... | 02:14 |
plasma_in_EDT | systemdlete: And os-prober isn't disabled in /etc/default/grub, correct? | 02:14 |
systemdlete | plasma_in_EDT, disabled where? On daedalus? | 02:14 |
plasma_in_EDT | systemdlete: yes | 02:15 |
brocashelm | you might have to go into a live session and try to manually reinstall grub to the os disk from a chroot or something (just a thought) | 02:15 |
brocashelm | although with things like luks/encryption, i'm not sure myself (assuming these are unencrypted installs) | 02:15 |
onefang | If it's an EFI system, rEFInd is your friend. It figures out for itself at boot time what are the bootable systems. | 02:16 |
systemdlete | 4.19.0-20 for beowulf | 02:17 |
brocashelm | it would have been more straight-forward if it was a legacy boot system | 02:17 |
brocashelm | uefi is like playing russian roulette | 02:18 |
systemdlete | brocashelm, I've done exactly that already, and like I said, I can boot into it with the aid of systemrescuecd or similiar. | 02:18 |
systemdlete | brocashelm, this is a test system. I feel no need for luks. So we can elminiate that | 02:18 |
brocashelm | probably something in the partitioning that it's not picking up on compared to chimaera/daedalus? are all three operating systems on the same disk? | 02:19 |
plasma_in_EDT | rwp: It looks like this was a change originating in Debian in response to https://bugs.debian.org/1042894. | 02:19 |
brocashelm | systemdlete: if it's efi (like onefang said), try refind (it's in the repo) | 02:19 |
systemdlete | as far as UEFI, the reconfigured partitioning IS GPT, but I have a 1M boot partition, and both chimaera and daedalus are now booting fine. Just not beowulf. | 02:20 |
rwp | brocashelm, UEFI is like trying to pet a feral cat. | 02:20 |
rwp | plasma_in_EDT, So... Is Debian fully UsrMerge now? Would they accept a bug to change that back? Or tell us to go away? | 02:20 |
krumer | Hola que tal como vamos ...¿ | 02:22 |
krumer | me encanta DEVUAN.. | 02:22 |
krumer | Yo era Devianlover ... pero por el systemD ya me fui ... | 02:22 |
systemdlete | No EFI or UEFI, other than the boot bios partition for gpt. But I've got 2 of 3 partitions booting from that disk now. | 02:22 |
rwp | krumer, I am glad you love Devuan! Thanks for the praise. You might find #devuan-es more suitable to Spanish language. :-) | 02:23 |
plasma_in_EDT | rwp: As time goes on it seems it will become harder and harder for Devuan to support non-usr-merged systems given that Debian has declared that UsrMerge will be the only supported configuration going forward. I'd like to continue using the fs layout I have been using, but that may stop being feasible short of custom compiling everything. | 02:24 |
Xenguy | krumer, Gracias amigo | 02:24 |
Xenguy | I have the impression Devuan must follow usrmerge, but don't quote me | 02:25 |
Xenguy | We are concerned with systemd anyway | 02:25 |
rwp | plasma_in_EDT, Right. And it seems inefficient to need to fork cron-daemon-common simply to change one PATH line in the crontab file. | 02:25 |
onefang | My opinion is it's no problem just using a few links to work with usrmerge. It's just not worth fighting it. Make your symlinks, then you can forget about it. It's been more work just saying that over and over, compared to actually dealing with it. | 02:27 |
systemdlete | Burn down all the churches--we don't need no dang religion anymore! | 02:27 |
systemdlete | Just tell the parishioners they can do their prayers on Zoom now. | 02:28 |
onefang | Nah, convert them into housing. Much more climate friendly, and I need anew home. | 02:28 |
* onefang goes back to searching for a new home. | 02:29 | |
systemdlete | Converting them is less traumatic and less resource friendly, that's why. This is America, OK? | 02:29 |
golinux | systemdlete: Have you checked your fstab that the uuids are in order? | 02:29 |
golinux | Didn't read closely enought to know whether you solved thator if I even read the problem correctly | 02:30 |
systemdlete | golinux, I have, but maybe I should double-triple check them... Maybe I need to run update-initramfs or something. | 02:30 |
plasma_in_EDT | rwp: To revert the change, it looks like you'd need to modify the PATH variable in https://sources.debian.org/src/cron/3.0pl1-166/debian/crontab.main/, drop [this patch](https://sources.debian.org/src/cron/3.0pl1-166/debian/patches/fixes/Usr-bin-sbin.patch/) and recompile. | 02:30 |
systemdlete | golinux, I moved my existing chimeara and beowulf partitions to a new gpt partitioning scheme while in the process of installing daedalus. That is probably the issue right there, and I know that Devuan makes no guarantee that if you choose to take big risks like that you are on your own. So I know I am on my own with this. | 02:32 |
systemdlete | And, golinux, this is a testbox, so it is not really that important. | 02:32 |
systemdlete | I have probably used up more oxygen in this room than is warranted. I just came to see if anyone else had similar problems. | 02:32 |
systemdlete | with os-prober. | 02:32 |
Xenguy | systemdlete, You sure have, I'm having trouble breathing here | 02:33 |
Xenguy | hahah | 02:33 |
systemdlete | sorry Xenguy | 02:33 |
Xenguy | I'm kidding, you go go | 02:33 |
systemdlete | (I know) | 02:33 |
brocashelm | can i still opt out of usrmerge if i blacklist that package and usr-is-merged? | 02:34 |
systemdlete | Actually, this experience has been more than a bit instructive. I've come close to killing it all more than once. So I've really come to an understanding of the boot workings, grub, initramfs, and the lvm and mdadm aspects. | 02:34 |
brocashelm | systemdlete: i do wonder, if you still keep that beowulf partition, what it would be like to use lilo in place of grub (since beowulf is the last to ship with it) | 02:35 |
systemdlete | In fact, one thing I should ALSO point out is that the old scheme was largely LVM based, whereas the new gpt partitioning is completely MDADM based. So much so that chimaera now boots off a PARTITIONED raid drive sitting on a raid partition!!! LOL | 02:36 |
onefang | brocashelm: as I said, much easier to just install a few symlinks and be done with it. It's not the big problem everyone thinks it is, unless you decide to fight it. Symlink and move on. | 02:36 |
systemdlete | It's amazing that you can actually do this at all! | 02:36 |
systemdlete | chimaera's root fs was sitting on a LV, and it booted quite nicely. I don't recall now how that came about back when I installed chimaera on that testbox. Now it is booting off the 3rd partition of a raid drive, which itself sits on a raid partition. | 02:38 |
systemdlete | why? | 02:38 |
systemdlete | well, partly to see what it is like to partition a raid drive, something I've noted in the past while working with raid, but never tried. | 02:38 |
systemdlete | Anyway, I will continue to hack at it. I'll probably get there, after several more hours of grinding through it. I mean, I have gotten this far with a rather ambitious (and maybe foolish) migration effort. | 02:39 |
golinux | systemdlete: Yeah when you move stuff around UUIDs can get scrambled | 02:40 |
golinux | and reassigned | 02:40 |
systemdlete | Mostly, I wanted an excuse to fill up the channel with some noise, while also sending praise back to you all for your work on this. | 02:40 |
brocashelm | i will avoid usrmerge as much as i can | 02:40 |
brocashelm | it's blacklisted. if an update tries to force it, then i guess that's that | 02:41 |
brocashelm | i curiously tried it once and it was a disaster; good thing i back up regularly | 02:41 |
systemdlete | a (d)isaster? Isn't that what the "d" in systemd stands for? | 02:42 |
critr | i used to think i was the last person on earth using bios/mbr | 02:42 |
systemdlete | critr, all of my PC hardware is still bios/mbr, except for a laptop which doesn't even work. | 02:43 |
onefang | That's why I name things rather than using UUIDs. | 02:44 |
brocashelm | i use legacy/bios/mbr as much as i can | 02:44 |
brocashelm | i have one laptop that disabled internal legacy boot :( | 02:44 |
brocashelm | so i had to do that bullshit uefi/fat32 partitioning to get my os booting | 02:44 |
critr | labels are a lot easier than uuid's | 02:45 |
critr | up until recently i used a mobo with legacy bios option, but usb and ethernet wouldn't work unless you turned on iommu in bios. | 02:47 |
systemdlete | Here's an actual support question: For mdadm raid, I have been noticing some changes to the way drives can be named over the years. Today, it seems you can't name a drive, say, /dev/raid5 although /dev/md5 seems to be OK. But you can always name them in /dev/md however you want. Trouble is, df does not have any options to list drives by their /dev/md/* names. So I see things like /dev/md127, /dev/md126, etc. which is not very helpful. | 02:49 |
systemdlete | Is there some secret menu item somewhere in, say, sysctl that can be set to address this? It's somewhat minor, but really annoying. | 02:49 |
systemdlete | golinux, acknowledged. But at least I admitted it, right? | 02:50 |
systemdlete | I've been bitten by those UUIDs a few times during this reconfig and install. You are very right. | 02:51 |
golinux | Just make sure that it's in order. UUIDs are very fickle | 02:51 |
systemdlete | And remembering to run initramfs has been a stumblebum as well. If the init stage still has the wrong uuid info, it can be very frustrating. | 02:52 |
rrq | when yous say "can't name a drive /dev/raid5" .. what does that mean? device nodesa are really named by major:minor and the filesystem name is generally not important | 02:52 |
systemdlete | rrq: mdadm allows you to name drives. These show up in /dev/md/* but not in /dev/* | 02:53 |
systemdlete | (where "*" is my notation for whatever the name is) | 02:53 |
systemdlete | My chimaera install, before I did this whole install and reconfiguration with gpt, had /dev/md2 as well as /dev/md/md2. | 02:54 |
rrq | mmm so in theory you could add your /dev/raid5 node if replicating the maj:min of /dev/md/5 ? | 02:54 |
rrq | mmm so in theory you could add your /dev/raid5 node if replicating the maj:min of /dev/md/mf5 ? | 02:54 |
rrq | mmm so in theory you could add your /dev/raid5 node if replicating the maj:min of /dev/md/md5 ? | 02:54 |
systemdlete | I can't seem to do this with the new scheme, at least not yet. | 02:55 |
systemdlete | What do you mean by "replicating?" (sorry) | 02:55 |
systemdlete | When you create or assemble a raid array, you can pass it a name, like "raid5" | 02:56 |
rrq | do: "ls -l /dev/md/md5" first, to see what its major:minor are, then do "mknod /dev/raid5 b $major $minor" | 02:56 |
plasma_in_EDT | rwp: Looks like this has been reported upstream at https://bugs.debian.org/1049969 | 02:56 |
systemdlete | the system will create /dev/md/raid5 which is a link to /dev/md127 (or whatever it ends up as) | 02:56 |
systemdlete | rrq: I am not sure that won't conflict with mdadm's scheme for managing arrays. | 02:57 |
rrq | I suppose the software wants you to change your mind about what's good for you :) | 02:57 |
systemdlete | rrq: Yep. | 02:57 |
systemdlete | Kind of like another chunk of software we know | 02:57 |
rrq | yeah | 02:58 |
rrq | there's consulting fees attached to every source of confusion | 02:58 |
* systemdlete switches his kvms back over to his testbox and continues hacking away... | 02:59 | |
rrq | systemdlete: with software raid, isn't your runtime system defined by the kernel+inird that spins of the raid.. then it's just a question of switching to the desired release version filesystem ? | 03:06 |
rrq | or, you want each releae version inird to spin up that same raid (in their own ways) ? | 03:13 |
rwp | systemdlete, It is possible to rename md devices in mdadm, just not trivially easy. | 03:39 |
rwp | Use --update=super-minor to do it. Like this: mdadm --assemble /dev/md1 --update=super-minor /dev/sda1 /dev/sdb1 | 03:40 |
rwp | Which usually means booting rescue media since it can only be done when the array is assembled. | 03:40 |
rwp | Then after making that change the initramfs needs to be updated. I update all with: update-initramfs -u -k all | 03:40 |
onefang | I listen to an Internet radio station all day using VLC, but when me regular incoming rsync jobs run (backing up my servers) I get constant dropouts. rsync IS running under ionice. What other programs might cope better for listening to somaa.fm, or are there other ways I could tame rsync's server? | 04:13 |
gnarface | i've never had a problem with audacious in the background that i can recall, but it seems like you just need to increase the default buffer size in vlc | 04:16 |
bb|hcb | nice/ionice will not affect networking. Your option is to create a qdisc, e.g. htb with 2 queues. Priority itself will not solve the problem, because it doesn't know how much bw you have - the congestion point is not your egress interface. IMHO, the option is to limit rsync traffic to whatever bw you have minus 1mbit... | 04:17 |
gnarface | oh, maybe i read this wrong... yea if you have a asynchronous connection with less upload bandwidth than download bandwith, it's really easy to choke download speeds by saturating the upload side | 04:18 |
gnarface | you probably honestly don't need a whole 1mbit for the tcp acks though | 04:19 |
gnarface | a few kb is probably enough | 04:19 |
systemdlete | rwp: Thank you so much for that. | 04:19 |
bb|hcb | right, if you know the exact limiter on the ISP end ;) | 04:19 |
bb|hcb | But since you don't know (they may use 1000 for 1k or 1024 for 1k, may include or exclude packet headers, etc, etc), 1mbit is a safe bet | 04:20 |
onefang | I'm not seeing any buffer size controls in VLC. "create a qdisc, e.g. htb with 2 queues" is words I've not seen before, what is that? | 04:20 |
onefang | Limiting rsync bw sounds good. | 04:21 |
bb|hcb | ah, "man tc" - you do that on eth0, if internet comes there | 04:21 |
gnarface | onefang: been a while, but as i recall you have to go into advanced options where it exposes the raw command-line field and just edit it directly | 04:21 |
gnarface | not sure if will actually work though now if bb|hcb's theory is correct about the actual problem though | 04:22 |
gnarface | my theory was that it was actually *cpu* priority that was hosing your stream | 04:22 |
gnarface | so regular nice would work in that case too, possibly | 04:22 |
gnarface | but if your upload bandwith is saturated then the download is just gonna choke out no matter what | 04:23 |
onefang | We are using my ancient WiFi router, coz it's too far for me to string Ethernet. So not the fastest. HFC to the rest of the world. Server end is my sledjhamr.org which everyone here keeps telling people is the best mirror, no bw issues there. B-) | 04:23 |
gnarface | (well, unless maybe you switch everything to udp or something) | 04:23 |
bb|hcb | cable/docsis says it all ;) | 04:23 |
gnarface | onefang: ok so, just in case, in vlc: media->open network stream->show more options | 04:24 |
onefang | Or, just get lucky when I move sometime before October and hope to get better Internet. And hope this old client pulls his finger out and actually gets me to do some work, so I can afford a better WiFi router. | 04:25 |
gnarface | the "edit options" field is just for literal command-line additions | 04:25 |
gnarface | but on my version there's a separate caching field there too | 04:25 |
onefang | Yep, I see that caching field now. Is there a way to do that for a playlist with several streams in it? | 04:26 |
gnarface | uh, not sure. like i said i usually use audacious for this. | 04:26 |
onefang | Ah right click and Advanced Open. | 04:26 |
bb|hcb | gnarface: Normally on docsis 2.0 you don't get more than 2mbit upload, on 3.0 you may get 10-20. That can never compare to cpu throttling | 04:26 |
onefang | I'm doubling the caching to 2000 ms. See if that helps. | 04:27 |
onefang | Now I gotta wait for the next rsync. | 04:28 |
onefang | Thanks everyone. | 04:28 |
gnarface | good luck | 04:28 |
bb|hcb | So the 'proper' way of doing that is to measure the real bandwidth in the rsync direction, then limit it to something less. I'd bet that rsync chokes the upload, then tcp acks get dropped and the radio stalls. Increasing the buffer would most probbaly not help | 04:29 |
onefang | If that works, one less frustration in my currently shit life. lol | 04:29 |
gnarface | well the cheap wifi router could be a bottleneck too | 04:29 |
onefang | It's not cheap, just ancient. I bought it loooong ago. | 04:29 |
onefang | And then didn't need it until we moved into this house four months ago. | 04:30 |
rrq | and rsync has a bwlimit option as well | 04:30 |
onefang | Well, I needed it when I bought it. | 04:31 |
bb|hcb | tc qdisc add dev eth0 root handle 1:0 htb default 1 | 04:38 |
bb|hcb | tc class add dev eth0 parent 1:0 classid 1:1 htb rate 100mbit ceil 100mbit | 04:38 |
bb|hcb | tc class add dev eth0 parent 1:0 classid 1:2 htb rate 5mbit ceil 5mbit | 04:38 |
bb|hcb | tc filter add dev eth0 protocol ip parent 1:0 u32 match ip dst 185.38.15.81 flowid 1:2 | 04:38 |
bb|hcb | onefang: ^^^ | 04:39 |
onefang | Ah tc is traffic control settings. Now I have a man page to read. Thanks. | 04:39 |
onefang | "qdisc is short for 'queueing discipline'" | 04:39 |
bb|hcb | and to undo all of the above: tc q del dev eth0 root | 04:40 |
onefang | And HTB is "Hierarchy Token Bucket implements a rich linksharing hierarchy of classes with an emphasis on conforming to existing practices" | 04:42 |
onefang | Something complex to learn (which I'm great at), but I'm busy with RL enshitification, so let's see how well the VLC caching change helps first. | 04:43 |
fluffywolf | do they also leverage synergies to enhance business innovation while driving market immersion with an action bias? | 04:43 |
onefang | I suspect real estate agents do, which might be why we are having a country wide rental crisis. | 04:44 |
systemdlete | fluffywolf!!! *eyesroll* | 05:57 |
systemdlete | fluffywolf, you must be back from a corporate seminar on optimizing your benefits, or something like that | 05:58 |
fluffywolf | lol | 05:58 |
fluffywolf | the thing onefang pasted just struck me as a bunch of buzzwords... | 05:59 |
systemdlete | golinux, I carefully double and triple checked the uuids of the md raids and all else. I added a custom.cfg in /boot/grub and reran update-grub (all on daedalus). Now I get an entry on boot to call beowulf, and it works! | 05:59 |
systemdlete | So I conclude that this problem is squarely upon os-prober. | 06:00 |
systemdlete | For some reason, it is not seeing what is a perfectly legit boot partition (since it works). | 06:00 |
systemdlete | In the past, I have heard a few moans and groans about os-prober, tbh. And docs even indicate that it is a potential security hazard. | 06:01 |
systemdlete | fluffywolf, where did you learn to talk like that? | 06:02 |
systemdlete | and onefang, you really need to watch yourself | 06:02 |
* systemdlete thinks: Such Language! | 06:03 | |
systemdlete | fluffywolf, tell me one thing: Did it hurt to say all that? A little bit even? | 06:03 |
fluffywolf | lol | 06:03 |
systemdlete | I used to attend quality seminars at AT&T and they supplied plenty of printed material also telling you all about (1) what quality is, (2) how to produce quality, and (3) what YOU can do to improve quality. Yeah, I learned a lot from all that S.H.I.T. | 06:05 |
systemdlete | (special high intensity training) | 06:05 |
systemdlete | Not once did it mention telephony, computer networking, software technology, or anything else we actually worked on. So I totally agree. | 06:06 |
systemdlete | golinux, anyway, I will consider the matter settled. If there are further problems with future releases, maybe I'll try filing a bug report. | 06:07 |
systemdlete | I have enough outstanding bug reports that have been ignored, sometimes for years, and this one is hardly important esp since there is an easy enough workaround. | 06:08 |
onefang | That buzzword overkill was quoting from a man page. lol | 06:24 |
onefang | "I was quoting" | 06:25 |
onefang | As for watching myself, I've been dealing with real estate agent web sites. Their bullshit is in danger of rubbing off on me. | 06:26 |
systemdlete | sounds like it | 06:26 |
systemdlete | what ticks me off the most is that people get PAID to write that crap | 06:27 |
fluffywolf | some people's jobs should be replaced by AI... and then we'll hopefully realize the job they were doing shouldn't be done at all. | 06:28 |
onefang | lol, though #offtopic. | 06:28 |
fluffywolf | I consider that a pretty good test, really... if a job really could be done by chatgpt, does that job actually do anything useful anyway? | 06:37 |
fluffywolf | bbl, bedtime | 06:47 |
clemens3 | argh, first time installing daedalus.. | 10:48 |
clemens3 | when asked in the installer where to install grub, it used to be a selection, now you have to type in /dev/sda or so | 10:48 |
clemens3 | i just entered and now it can't boot, i would expect it to choose the one and only partition available.. | 10:49 |
clemens3 | so let's try again | 10:49 |
clemens3 | maybe i did the automatic install first and then it went astray.. | 10:53 |
clemens3 | but choosing between automatic and expert, what if i am not an expert, and just want a stone age everage manual install? | 10:55 |
clemens3 | it signals devuan is just for experts | 10:55 |
rrq | last time I tried the dialog had the options "manual" or a device; I typically select that device | 10:56 |
clemens3 | i am still not through, but definitely it is a step back compared to the last two releases | 10:58 |
clemens3 | i am sure it is debian, but just sayin | 11:00 |
clemens3 | why i need to choose a kernel? If I choose none? | 11:00 |
clemens3 | seriously | 11:00 |
rrq | that grub step is like it's been since jessie | 11:00 |
rrq | bad day, today? | 11:01 |
clemens3 | the grub i came from auto and then it contionued without an auto file in some manual mode | 11:01 |
clemens3 | and finally i had to fill a text field | 11:01 |
clemens3 | rrq, well, is there a good reason to choose a kernel at install? | 11:01 |
clemens3 | i thought you release a new version to improve things | 11:02 |
rrq | if there are options there might be | 11:02 |
jaromil | Ahoy! in case anyone is around at the Chaos Computer Camp, some devuanistas are having a meeting there! meet Adam and Plentyn at 14:30 in the Aachen village (just opposite of Heaven) https://map.events.ccc.de/camp/2023/map/#18.56/53.0307173/13.3053098/m=53.030781,13.305234 | 11:16 |
jaromil | see also https://toot.community/@devuan/110909726848442110 | 11:18 |
Neilo | Hello everyone, I have downloaded the new Devuan 5 netinstall image (devuan_daedalus_5.0.0_amd64_netinstall.iso), but when I try to check the GPG keys, it doesn't work. | 12:17 |
Neilo | I followed the instructions at https://www.devuan.org/os/documentation/install-guides/daedalus/install-devuan. | 12:17 |
Neilo | I get the following error message: | 12:17 |
Neilo | $ gpg --verify SHA256SUMS.txt.asc devuan_daedalus_5.0.0_amd64_netinstall.iso | 12:17 |
Neilo | gpg: Signature made on Mon, Aug 14, 2023 18:57:56 CEST | 12:17 |
Neilo | gpg: with RSA key 619933B4CD8A97408A3C47E2680B5A1F661ECDBC | 12:17 |
Neilo | gpg: WRONG signature of "Ralph Ronnquist <rrq@rrq.au>" [unknown] | 12:17 |
Neilo | Does this mean that the file is corrupted ? | 12:17 |
onefang | rrq should be able to help, there was some sort of issue and you might need to find his latest key. | 12:18 |
onefang | I don't recall the details. | 12:18 |
rrq | mmm we have the keys at keyring.devuan.org .. I guess that file https://files.devuan.org/devuan-devs.gpg isn't updated | 12:21 |
rrq | i.e. Nello you might want to import from https://keyring.devuan.org instead | 12:23 |
xrogaan | meta, asking what to do. Somebody on debian side made a QA update to compton (which is defunct) and pointed the new sources towards picom, which is the fork. What would be the right move forward? Do I reply to the bug (908533) or is it legit, or something else? | 12:25 |
xrogaan | https://salsa.debian.org/debian/compton/-/commit/303f5e6a03d8414bbd670dd8f4c47b6d585de096 | 12:25 |
xrogaan | picom exists as is in debian. | 12:26 |
Neilo | I need password ? I don't understand where I upload the file to the site :/ | 12:27 |
rrq | do: gpg --keyserver https://keyring.devuan.org --search-keys 619933B4CD8A97408A3C47E2680B5A1F661ECDBC | 12:27 |
rrq | that will import that key to your own keyring | 12:27 |
rrq | then the signature check (next step) will not complain | 12:28 |
rrq | (I'll tyr to get my key into that devuan-devs.gpg file.. but it'll take a few breaths) | 12:28 |
Neilo | That's strange, I have the same error | 12:35 |
Neilo | gpg: WRONG signature of "Ralph Ronnquist <rrq@rrq.au>" [unknown] | 12:36 |
Neilo | "gpg: error searching keyserver: Operation cancelled | 12:38 |
Neilo | gpg: key server search failure : Operation cancelled" | 12:38 |
rrq | I'm not sure what's happening :( have ti dig into it.... | 12:46 |
xrogaan | rrq: trying the gpg command, gpg tells me: https://dpaste.com/3UTV7M4QU | 12:53 |
Neilo | Yes, i have same problem | 12:54 |
xrogaan | running gpg (GnuPG) 2.2.40 | 12:55 |
xrogaan | rrq: devuan keyring server isn't working correctly. Using the default (http://keys.openpgp.org:11371) imports the key correctly. | 13:00 |
djph | rrq: you break your key again? | 13:06 |
rrq | not sure; maybe gpg has changed protocol so the server is outdated ? | 13:31 |
djph | more I remember there being a discrepancy between the website and something you were signing | 13:32 |
djph | sidenote, I really gotta find a local group or some such that's running a keysigning party ... | 13:32 |
rrq | that error (gpg: no valid OpenPGP data found) comes from gpg when I trye to import the key | 13:33 |
xrogaan | yes | 13:34 |
xrogaan | Thought that was obvious :) | 13:34 |
rrq | I have no idea why | 13:34 |
xrogaan | 'cause whatever the server is sending isn't valid per gpg | 13:34 |
rrq | and why did that start to happen now? | 13:34 |
rrq | the server has been serving keys some year or so | 13:35 |
djph | huh, it worked here | 13:35 |
djph | well, 'worked' ... it just showed me "key 1 of 1 Ralph [...]" | 13:36 |
rrq | then as you selet it you get the error | 13:36 |
rrq | select | 13:36 |
djph | hm, I only had "next" or "quit" | 13:37 |
djph | oh wait, nvm, "select the number" got eaten the first time through | 13:37 |
xrogaan | djph: but the import fails | 13:38 |
djph | indeed, | 13:39 |
xrogaan | okay, found the issue | 13:39 |
xrogaan | the server doesn't send the key, but the index page | 13:39 |
xrogaan | you need to get this: https://keyring.devuan.org/pks/lookup?op=get&search=0x680B5A1F661ECDBC | 13:39 |
xrogaan | but you get https://keyring.devuan.org/ | 13:40 |
rrq | ok | 13:40 |
clemens3 | Neilo: imho you got wrong gpg command line, you need to compare the signature file with the SHASUMS file | 13:40 |
xrogaan | rrq: here's the debug-all output: https://dpaste.com/GVYGAFZX7 | 13:40 |
clemens3 | gpg --verify SHA256SUMS.txt.asc | 13:41 |
clemens3 | if good signature | 13:41 |
clemens3 | then look into SHASUMS file | 13:41 |
clemens3 | and sha sum your iso image | 13:41 |
clemens3 | and compare | 13:41 |
rrq | xrogaan: thanks .. trying to work out why | 13:41 |
clemens3 | i used another keyserver, worked out all just fine | 13:44 |
clemens3 | could follow the web of trust | 13:44 |
xrogaan | rrq: what's the software used on the server side? | 13:45 |
rrq | onak via nginx | 13:45 |
rrq | hmm works with hkps | 13:49 |
rrq | gpg --keyserver hkps://keyring.devuan.org --search-keys 680B5A1F661ECDBC | 13:49 |
rrq | isn;t that just the same protocol but on port 11371 ? | 13:51 |
rrq | .. or even: gpg --keyserver keyring.devuan.org --search-keys 680B5A1F661ECDBC | 13:54 |
rrq | all down to my bogus advice I suppose :( | 13:54 |
djph | huh, well, that's fun... signed the key and now my web of trust decided to rebuild. I don't even TRUST any of these keys, *sigh* GPG ... | 13:55 |
xrogaan | rrq: so nginx being a reverse proxy, it doesn't understand gpg request? | 13:57 |
rrq | well, the request is passed to onak regardless so I guess onak's lookup binary makes a distinction | 13:58 |
rrq | or the nginx conf is wrong | 14:00 |
rrq | port 443 has a prior "location / {" .. maybe that kicks in instead? | 14:01 |
rrq | though is "location block" order in config important? | 14:02 |
djph | dunno :( , I don't run nginx here (well, except that integral one with gitlab, but it's got an apache proxy in front) | 14:03 |
rrq | probably safest to leave it for bb|hcb ... I'll probably just mess things up | 14:05 |
djph | rrq: at least you're cognizant of your limits? | 14:06 |
rrq | ... some, sometimes :) | 14:06 |
djph | hah | 14:17 |
Neilo | Ok, now it's work | 14:18 |
rrq | thanks .. hope you enjoy | 14:18 |
Neilo | with gpg --verify SHA256SUMS.txt.asc | 14:19 |
rrq | yes. | 14:19 |
Neilo | And thank you all for your help :) | 14:19 |
djph | *sigh* why have search engines gone to such lengths to be terrible. Where's a good key-signing etiquette guide these days? | 14:25 |
Neilo | Maybe ChatGPT can give you a guide | 14:29 |
djph | no. bad. | 14:34 |
djph | it'll probably just tell me how to sign a file or some other useless "guide" | 14:34 |
bb|hcb | rrq: the working command is: gpg --keyserver hkps://keyring.devuan.org --search-keys 619933B4CD8A97408A3C47E2680B5A1F661ECDBC | 15:57 |
bb|hcb | note hkps instead of https | 15:57 |
bb|hcb | Or leave the protocol for gpg to decide, like: --keyserver keyring.devuan.org | 15:57 |
bb|hcb | about files.d.o/devuan-devs.gpg and devuan-archive-keyring.gpg, I did update them on 15 Aug | 15:58 |
bb|hcb | wget https://files.devuan.org/devuan-devs.gpg && gpg --import devuan-devs.gpg | 16:00 |
bb|hcb | rrq: Your new 4096 bit key is already there | 16:00 |
xrogaan | Bad protocol! Sounds like a music. | 16:10 |
xrogaan | https://brokenbit.bandcamp.com/track/bad-protocol | 16:13 |
joerg | xrogaan: awesome :-) | 17:06 |
al1r4d | #devuan-bridge | 17:36 |
al1r4d | :) | 17:36 |
golinux | Whats it this #devuan-bridge channel that someone just posted? Can anyone just create a new channel? I'm hesitant to even go there . . . | 17:39 |
golinux | joerg: ^^^ | 17:42 |
joerg | new channels just exist as soon as >0 users decide to join | 17:43 |
djph | /join #golinuxsnewchannel | 17:47 |
golinux | Please delete that | 17:56 |
djph | delete what? | 17:57 |
golinux | Whatever you just used my nic for without permission | 17:58 |
djph | Kinda hard to delete something once it's been said on IRC :/ (there isn't, to my knowledge, a channel by that name, unless someone else created it after i mentioned it.. ) | 18:00 |
al1r4d | golinux: someone on telegram made that channel. so devuan telegram group unofficial <> irc | 18:13 |
xrogaan | golinux: in cases of harassment, you can reach the ircops. In this case, it's just some nonsense that will go away. | 18:14 |
joerg | please everybody stop discussing this topic in here. Now. | 18:15 |
xrogaan | I don't recall if liberachat supports namespaces for projects. So that nobody can create a new channel with a project name in it. | 18:15 |
xrogaan | joerg: gotcha | 18:15 |
sakrecoer | Greetings Devuan! | 18:22 |
sakrecoer | There's now a channel called #devuan-bridge that is bridged to telegram. This is not intended to replace anything. It is a mean to get telegram user closer to the source (and break some silos) | 18:23 |
al1r4d | Nice sakrecoer | 18:24 |
sakrecoer | You may join it or ignore it, but rest assured it is moderated and monitored with care and attention. | 18:24 |
al1r4d | Thank you for your works, sakrecoer | 18:25 |
sakrecoer | For anyone interested, the user "Pontifex" is the bridge bot. It listens to what is written here on IRC and over on https://t.me/DevuanGnuLinux and relays it. | 18:26 |
sakrecoer | well "here"... it only listens to #devuan-bridge | 18:26 |
sakrecoer | thank you al1r4d! | 18:27 |
al1r4d | sakrecoer: hmm, sorry, since english is not my native. What is meaning of "thank you al1r4d"? You supposed to be say "your welcome", right? | 18:27 |
sakrecoer | al1r4d: you wrote "thank you or your work" and i meant to thank you for appreciating it | 18:28 |
al1r4d | sakrecoer: ok, nice | 18:28 |
golinux | A heads up would have been nice. | 18:42 |
golinux | Social networking on devuan is anathema to me. It is a sad day. | 18:44 |
sakrecoer[bpist] | agreed! i did it in the contortionist branch of agile :sweat: | 18:45 |
sakrecoer[bpist] | well... it's sad that eveyrone migrated from IRC, yes. Is it a good idea to create contact between bubbles? i don't know but i hope so | 18:45 |
sakrecoer[bpist] | after all, IRC is the OG social network.... albeit from a time where socializing meant something different from today, but chang eis the only constant | 18:47 |
golinux | I don't want to touch the bubble od social networking and never will. Sad that it might be infecting one of the remaining havens of sanity | 18:48 |
joerg | who's in control of pontifex? | 18:48 |
golinux | Who is pontifex? | 18:48 |
sakrecoer[bpist] | joerg: jaromil and me | 18:49 |
joerg | >>[18 Aug 2023 18:26:24] <sakrecoer> For anyone interested, the user "Pontifex" is the bridge bot.<< | 18:49 |
joerg | good | 18:50 |
sakrecoer[bpist] | golinux: sadness is a good emotion. then again, social media is like IRC with a corporate dialect. I'd like to think it is a happy day fro communication: the major rempart of individualism being exposed to the virtues of altruism | 18:51 |
* golinux experiences fear and dread . . . | 18:52 | |
sakrecoer[bpist] | also, if it turns out to be a bad experience, we WILL kill it :) | 18:52 |
sakrecoer[bpist] | just to clarify, this bridge will not make it possible for telegram users to reach out yo IRC users in any other setting than inside #devuan-bridge | 18:53 |
golinux | What bothers me most is you guys just swooping in and doing whatever YOU want without an public notification. Pretty arrogant . . . | 18:56 |
golinux | Makes it feel like an invasion not a collaboration which doesn't work for me . . . | 18:58 |
golinux | I am probably not alone. Time will tell and we all know where the door is . . . ;) | 18:59 |
sakrecoer[bpist] | well, maybe it's a disresepctful thing to think, but increasing accessibility is something i firmly believe in. | 18:59 |
critr | are bridged users moderatable in the same way as regular irc users? | 19:00 |
golinux | You don't just swoop into someone's house and start rearranging the furniture because you think it's a good idea . . . | 19:01 |
sakrecoer[bpist] | no, i wouldn't. That's why it is happening in a neighbouring empty house:) | 19:01 |
sakrecoer[bpist] | critr: no, that's why it requires moderators to be in both channels. | 19:03 |
critr | in that case, imo, bridged users shouldn't be allowed on irc unless the irc channel op is willing and able to moderate in both channels. | 19:05 |
golinux | Hopefully they will be good neighbors that I never have to see or hear.' | 19:06 |
golinux | I don't want a pickleball court nextdoor | 19:07 |
golinux | But who cares about that. All things do change so let the floodgates open! I surrender to the absurdity and chaos and can always pull the plug. | 19:09 |
sakrecoer[bpist] | critr: that's the thing, the telegram channel is moderated, and the lone channel that recieves telegram users message is too. | 19:09 |
sakrecoer[bpist] | golinux: yes, unles you join #devuan-bridge none of it can reach you. | 19:09 |
critr | would you let someone into your house who you can't moderate? | 19:10 |
sakrecoer[bpist] | no, that's why we are only letting people in that can be moderated :) | 19:10 |
sakrecoer[bpist] | actually, it might be harder to join the telegram channel than it is to join #devu | 19:11 |
sakrecoer[bpist] | #devuan-bridge | 19:11 |
critr | moderated by the irc channel op(s)? | 19:11 |
sakrecoer[bpist] | critr: you can't join #devuan-bridge from telegram without passing the telegram moderators. Can there be damage done? well.... a better questions is: does internet only do good? Let's try to cultivate a constructive channel :) | 19:13 |
critr | so you'd allow people into your house because the guy down the street said it was ok? | 19:23 |
sakrecoer[bpist] | critr: no, but if the house down street was empty with a great street number, i would totally invite you to play the best music you've heard in it. if you played bad music, i'm sure everyone would boo you out anways | 19:25 |
critr | well, any user who comes into an irc channel should be 100% moderatable by the operators of that irc channel. | 19:27 |
sakrecoer[bpist] | critr: IRC is retroactive moderation.(thank dogs, we are still innocent until proven otherwise) So, to answer your concern, someone coming in from telegram will be 100% moderated. Albeit retroactively, yes | 19:34 |
joerg | sakrecoer[bpist]: :-) | 19:39 |
djph | heyo joerg. "good job" and all that for daedalus :) | 19:43 |
djph | boo, only 1 see... oh, no there we go, that's better | 19:44 |
joerg | djph: I did literally zilch to contribute to Deadalus, so I'll pass on those kudos to the really awesome devels | 19:47 |
djph | joerg: heh, I can't keep tabs on who all is actually a dev, I should get better about that, what since having used it now since Jessie... | 19:48 |
djph | joerg: condolences in advance then for having to point out it's not "deadalus" | 19:49 |
joerg | sorry, typos is me :-D | 19:49 |
ted-ious | What's the name of the next spelling challenge version? :) | 19:54 |
djph | Probably whatever comes after 'Excalibur' | 19:55 |
ted-ious | I vote for Eyjafjallajökull. | 19:55 |
djph | ... that doesn't start with an 'F' though. | 19:55 |
golinux | Freia | 19:56 |
ted-ious | Oh ok if it's too late for e then my vote for f will be https://en.wikipedia.org/wiki/Fëanor. | 19:56 |
golinux | We already have an E | 19:56 |
golinux | ted-ious: Are you really that clueless/ All Devuan releases come from the list of Minor Planets | 19:58 |
golinux | https://minorplanetcenter.net//iau/lists/MPNames.html | 19:58 |
ted-ious | Sorry for trying to have fun for a few seconds. | 20:01 |
ted-ious | I'll leave you to be miserable in peace. | 20:02 |
golinux | Have fun in offtopic. | 20:02 |
golinux | This is a support channel not see how clever I am channel | 20:02 |
systemdlete | golinux, +1, maybe +2 or +3 | 20:21 |
jaromil | ahoy! I am happy we have an **opt-in** and **separate** channel called #devuan-bridge because the Devuan community in 2023 is larger than IRC, for instance the english speaking telegram channel now counts >200 people and growing fast. This will avoid disconnection, esp. for those IRCs here who like to share some wisdom with the sinners :^D | 20:59 |
jaromil | I for one am a sinner and using telegram on my mobile, so in case there is anything urgent I can be reached immediately through #devuan-bridge mentioning my nick, it will pop a notification on my handheld. Whoever likes that can use it too that way. Also you can lurk or even participate in discussions happening over telegram. | 21:01 |
jaromil | if you don't need that, I envy you :^) and it's fine to ignore | 21:02 |
joerg | jaromil: :thumbs-up: :-) | 21:16 |
systemdlete | I just can't wait for the first grammatically incomprehensible txt to come across. | 21:18 |
systemdlete | (Emphasis usually on the use of the medium rather than what the medium is used for.) | 21:19 |
golinux | jaromil: No one "needs" it. It is a choice unless someone points a gun to your head. | 21:46 |
golinux | As long as it doesn't invade this channel let those who wish go play in dystopia | 21:47 |
jaromil | of course, no way anyone from telegram can join this channel | 21:50 |
fifi | I believe that some people don't use smartphones? :-) | 21:54 |
golinux | I do not have one | 21:55 |
jaromil | the very best people :^)))) use maemo-leste (devuan based) like a boss https://maemo-leste.github.io/ | 21:55 |
jaromil | BTW they just released a new firmware based on Chimaera, 2 days before the Daedalus release :) porting on mobile takes time... | 21:56 |
jaromil | anyway, since less than a week one can use devuan on a Nokia N900 and make "real" phonecalls :^) | 21:59 |
fifi | wow! | 21:59 |
fifi | but how? | 22:00 |
jaromil | it's hacky, details are here https://leste.maemo.org/Nokia_N900 | 22:01 |
jaromil | there is also a #maemo-leste channel here on liberachat | 22:01 |
golinux | I thought this was a support channel . . . | 22:03 |
golinux | not a promotional one | 22:03 |
fifi | yup | 22:06 |
brocashelm | not sure how to feel about the telegram bridge channel. as long as it doesn't take away from the official support here | 22:10 |
brocashelm | seeing more channels openly allowing mastodon/matrix (whatever that is) users on, which you can tell by the [m] appended to their nicks | 22:11 |
brocashelm | and oftentimes, they send weird message formats that you have to "click to read the rest"... | 22:11 |
brocashelm | so i agree with what golinux said: it should have been discussed first | 22:12 |
hagbard | In my experience, in general, more channels of communication don't lead to more communication, but for more division of communcation into more channels. | 22:16 |
* joerg doesn't see any issues for now, with the bridge channel. | 22:30 | |
xrogaan | jaromil: https://xkcd.com/1782/ | 23:16 |
joerg | $xkcd 1782 | 23:22 |
joerg | oooh, wrong channel, sorry | 23:22 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!