| danuan | gnarface thanx for help the other day with slow apt repos, even though i still get sent down to europe majority of the time, i guess its deb.devuan.orgs way of throttling me. | 03:33 |
|---|---|---|
| gnarface | no problem | 03:35 |
| gnarface | check out apt-cacher-ng, it might help a little | 03:35 |
| danuan | setting apt-cacher was great , now most updates go at lan speed, and for the oddball time it needs new files now i might , start rednsing deb.devuan.org to fast servers as not to mess with the local mirror that all points to it now | 03:35 |
| danuan | ya | 03:35 |
| danuan | and here is ROOT on zfs HOWTO for two thirds done but enough to get going devuan https://dev1galaxy.org/viewtopic.php?pid=24411#p24411 | 03:39 |
| golinux | Did you fill it in? When I saw it, it was placeholders which was a bit strange | 03:41 |
| danuan | ye just did not want posts in the middle | 03:41 |
| danuan | last part i will add tomorrow , will deal will optimising things, making it faster, other options | 03:42 |
| danuan | but now that i check something happened to part 2 , need to edit again | 03:44 |
| danuan | multiple windows must have submited wrong one , ehh | 03:49 |
| danuan | ok found it in closed tabs and resubmited | 03:52 |
| danuan | part 1 and 2 there | 03:52 |
| golinux | I deleted my comment. Formatting still seems inconsistent | 04:00 |
| danuan | part3 is temp, otherwise anything specific ? | 04:02 |
| danuan | or another howto with a better formatting for ref? | 04:09 |
| golinux | Terminal output in code tags is a good rule. I see you added some of them. | 04:31 |
| Cgliz | Is this where I can ask questions related to Devuan? | 06:29 |
| cronolio | yes | 06:30 |
| Cgliz | Great just a small question: I was wondering why -by default when using the netinstall method of installation and only selecting to install base components- it installs adwaita icons? | 06:32 |
| gnarface | Cgliz: i don't actually know, but it might be because there needs to be some icon theme so someone chose that to be the default | 06:34 |
| gnarface | Cgliz: (as in, merely an aesthetic choice) | 06:34 |
| gnarface | like i said though, i don't actually know | 06:34 |
| Cgliz | ah, okay thanks for the help! | 06:35 |
| rwp | I see Cgliz has already left but I think they must have installed Xfce desktop, and adwaita-icon-theme is in the dependency chain of it. Seems to be for me anyway. | 07:57 |
| rwp | It's too bad that dependencies are so messy these days that apt-cache dotty generates such large plots. Too big for my large monitors! | 07:59 |
| rwp | Makes me wish I had an easy to use dependency explorer tool. Just to answer these types of questions. | 07:59 |
| rwp | But actually "apt-get purge -s adwaita-icon-theme" works pretty well to identify if something is actually Depends: or just Recommends: . | 08:01 |
| xinomilo | apt install $package --no-install-recommends | 08:40 |
| rwp | xinomilo, If that was in response to Cgliz's question asking why adwaita-icon-theme had been installed on a default Devuan installation then I think it doesn't work since the adwaita-icon-theme package is already installed upon the first login. | 08:53 |
| rwp | For my tastes I always install a minimal installation first. And then configure the system to my liking with APT::Install-Recommends "0"; and then install what I want from there. | 08:55 |
| xinomilo | ok, i didn't see the "default installation" part... | 08:57 |
| gnarface | well he said only "base components" which i assumed he meant base system utilities, as opposed to just checking nothing at all. maybe i should have mentioned that he could have checked nothing at all. maybe i should have clarified if he meant that he actually picked a desktop... | 09:48 |
| rwp | It's impossible to think of everything that a petitioner may not have said when they ask a partially specified question. | 09:57 |
| fonky | hi, for the lazy can somebody paste ceres repos if possible and not to much | 13:07 |
| fonky | using deb-src http://auto.mirror.devuan.org/merged ceres main | 13:08 |
| fonky | deb http://auto.mirror.devuan.org/merged ceres main | 13:08 |
| fonky | atm | 13:08 |
| fonky | but when apt-update i get pkgmaster.something | 13:08 |
| fonky | are any new ones or any link to where i can check | 13:08 |
| fonky | ty in advance | 13:09 |
| debdog | try deb.devuan.org instead | 13:12 |
| debdog | at least that's what Beowulf's release notes says http://files.devuan.org/devuan_beowulf/Release_notes.txt | 13:13 |
| debdog | though I am not familiar with ceres | 13:13 |
| xinomilo | deb http://deb.devuan.org/merged ceres main contrib non-free | 13:16 |
| fonky | ty | 13:46 |
| fonky | oh i have not been looking at new releases very much or checking out any feeds or souch as happy camper, without updating for around 3-4 months now | 13:47 |
| fonky | :) | 13:47 |
| fonky | rock stable, not sure how many people hacked the pc as i dont care to check anymore | 13:47 |
| fonky | will do that after fresh install though | 13:48 |
| fonky | anyways thank you all for the data | 14:04 |
| fonky | have a nice day | 14:05 |
| fonky | exit | 14:05 |
| Xelraa | what happened with the partimage package for devuan? | 20:49 |
| Xelraa | why is only the docs available | 20:49 |
| fsmithred | it's in jessie, stretch, bullseye and sid, according to package.debian.org. So it's not in buster/beowulf. I don't know the reason. | 20:51 |
| golinux | I was just about to say, ask Debian | 20:52 |
| Xelraa | not even in the contrib and non-free :( | 20:55 |
| fsmithred | after reading the changelog, I still don't know the reason. | 21:00 |
| fsmithred | and checking the bug list for partimage and partimage-server doesn't tell me any more. Most of the bugs are old, newest is 3 years. | 21:00 |
| Xelraa | Looks like an RC bug was reported and they didn't have time to fix it before the release: https://bugs.debian.org/cgi-bin/bugrepo ... =%23915987 | 21:08 |
| Xelraa | thats just great... | 21:08 |
| fsmithred | you could try backporting it | 21:12 |
| fsmithred | ok, that was easy. It build, it installs. I can't try it to see if it actually works, but it probably does. | 21:17 |
| fsmithred | Xelraa, do you know how to do that? | 21:25 |
| Xelraa | Im just doing it but its breaking Broken partimage-build-deps:i386 Depends on libssl1.0-dev:i386 < none @un H > | 21:29 |
| Xelraa | good old libssl1.0-dev | 21:30 |
| fsmithred | oh, I did it on amd64 | 21:30 |
| Xelraa | its an old 32bit only box | 21:30 |
| fsmithred | Xelraa, it builds ok here | 21:44 |
| fsmithred | Xelraa, did you get it? | 22:13 |
| user11 | hi i found a message from 2018-11 in the logs by katolaz mentioning initial mini.iso in the works for openpower/power9...was this effort halted? i don't recall seeing any ppc64le isos | 22:29 |
| fsmithred | yes, there is ppc64le mini.iso | 22:30 |
| fsmithred | https://pkgmaster.devuan.org/devuan/dists/beowulf/main/installer-ppc64el/current/images/ | 22:31 |
| fsmithred | user11, mini.iso is in netboot dir | 22:31 |
| user11 | oh nice since 3.0 huh | 22:32 |
| user11 | i saw it thnks | 22:33 |
| ham5urg | Is a simple Intel NUC + Devuan + Kodi a good choice for my TV? | 22:47 |
| user11 | do they actually manage to fit ME into the NUC, really? | 22:49 |
| user11 | managementengine | 22:49 |
| ham5urg | Yes, the BIOS | 22:49 |
| ham5urg | or EFI I must say is a fully OS and who knows what it is doing | 22:50 |
| user11 | ham5urg: i just searched the channel log for kodi...i didnt see nuc mentioned, but kodi likely works fine on devuan, and the nuc is popular with kodi users (on other distros) iirc | 23:08 |
| systemdlete2 | perf: interrupt took too long (2503 > 2500), lowering kernel.perf_event_max_sample_rate to 79750. | 23:20 |
| systemdlete2 | That happened last night, apparently, and was the last message in the log prior to reboot; I had to reboot my system. | 23:21 |
| systemdlete2 | The system was alive, as I could telnet/ssh to it, but it was not responsive. | 23:21 |
| systemdlete2 | Note that I just rebooted the system a few minutes ago; apparently, the ssystem became lame sometime since last night when that message appeared. | 23:22 |
| ham5urg | user11, yes, I used NUC for improvising server installation at emergency situations. small and robust. | 23:25 |
| gnarface | ham5urg: i remember someone having a string of problems with one due to (iirc) apparent refusal to patch the kernel or install via ethernet, but that might have been back during jessie or ascii | 23:25 |
| systemdlete2 | Searches on the web tell me it is intermittent, or that it is an advisory warning. Yet others indicate it freezes their system, such as happened to mine. | 23:25 |
| systemdlete2 | btw, I have not toyed at all with the performance subsystem, or the kernel. | 23:26 |
| gnarface | systemdlete2: i see them sometimes under load when there is high I/O wait, usually with writing big images to cheap shitty usb flash keys | 23:26 |
| systemdlete2 | However, I do have syslog-ng installed now, and I am wondering if that is related. I haven't had this prior to installing syslog-ng | 23:26 |
| systemdlete2 | gnarface: Hmmm. | 23:27 |
| systemdlete2 | I don't remember writing any large files to USB flash drive yesterday. I did do that, but on the other machine. | 23:27 |
| gnarface | systemdlete2: if it precedes destabilization for you, check into hardware devices potentially hanging or just being too slow (i think the default threshold for this warning is 120 seconds of "where you at dawg?") | 23:27 |
| systemdlete2 | There had been absolutely no messages for hours before that final message. | 23:28 |
| gnarface | systemdlete2: i thought i would see this usually with slow flash but it might be possible that any hardware device can trigger it if it doesn't respond in time | 23:28 |
| systemdlete2 | In fact, it looks like nearly 13 hours. But I am not sure if the switch to syslog-ng was within that period or not. | 23:28 |
| systemdlete2 | Let's see... | 23:30 |
| systemdlete2 | I installed syslog-ng at 11:00am yesterday. | 23:30 |
| gnarface | systemdlete2: also maybe i'm conflating this with another error, but if you were running ffmpeg at the time, try turning up the thread queue size | 23:30 |
| systemdlete2 | The last message before the final message prior to the freeze was at 10:29am yesterday. | 23:31 |
| systemdlete2 | I was running ffmpeg (vlc?) but -- again -- on the other machine. | 23:31 |
| systemdlete2 | So I have been doing the sorts of operations you say could lead to a freeze, but not on the machine that froze. | 23:32 |
| systemdlete2 | And no, I am not forwarding syslog-ng messages... at least, not yet. | 23:32 |
| systemdlete2 | (that could certainly cause a delay, esp. if the syslog-ng server were overloaded, e.g. But I don't have it configured that way... yet) | 23:33 |
| systemdlete2 | So it looks like I was running the default syslogger until 11:00am, according to the history.log when I installed syslog-ng. | 23:34 |
| systemdlete2 | Then there were no messages in the kern.log until 11:23 last night, when the perf error occurred. And there were no further messages after that in the kern.log until after the reboot. | 23:36 |
| systemdlete2 | I should look at some other logfiles... | 23:36 |
| systemdlete2 | Looks like there were successful messages from cron about half an hour before I rebooted. | 23:37 |
| systemdlete2 | So, maybe the freeze was very recent to the time I tried to scp from the other machine. | 23:38 |
| systemdlete2 | That perf: error might just be a fluke? | 23:38 |
| gnarface | hmmm, or a red herring | 23:48 |
| gnarface | not sure | 23:48 |
| gnarface | i think i even get those messages from my soundcard sometimes while using wine | 23:48 |
| systemdlete2 | But that still doesn't say much about the freeze-up. | 23:49 |
| systemdlete2 | Before this investigation began today, I was working on the other machine and wanted to scp a file to the frozen up one. That's when I realized I needed to investigate. | 23:49 |
| systemdlete2 | scp? | 23:49 |
| gnarface | no, it doesn't. they may or may not be related. my guess is chances are though it's a symptom of the same cause rather than the cause itself. | 23:50 |
| systemdlete2 | I suppose that could cause a freeze-up | 23:50 |
| gnarface | anything in the logs about a cpu "soft lockup" ? | 23:50 |
| systemdlete2 | no. | 23:51 |
| systemdlete2 | I checked kern.log for that one. | 23:51 |
| systemdlete2 | I've been noticing these cpu soft lockups though, esp in my virtual machines on the other box. But I wasn't running any VMs on the box in question. | 23:52 |
| systemdlete2 | In fact, the only thing I'm doing on my testbox atm is syslog-ng | 23:53 |
| systemdlete2 | working on it, I mean. | 23:53 |
| gnarface | when this all coincided with predictable lockups for me, the ultimate culprit turned out to be a bios setting called "C1E power saving" or something like that | 23:53 |
| gnarface | as well as temperature sensors that didn't respond on time all the time | 23:54 |
| systemdlete2 | I am wondering if syslog-ng itself could be the source of the problem. This could make sense given that syslog-ng does have gearing for managing remote queues, which might be invoked even if those are not configured | 23:54 |
| gnarface | but the combination of the two would cause soft lockups that would occasionally freeze the system entirely if there was disk load at the moment (like if i was scping something to it) | 23:54 |
| systemdlete2 | Yeah, I keep forgetting about that C1E biz. I need to look at it, on both my boxes. | 23:54 |
| gnarface | i thought it was a AMD feature so it may not be relevant to you, but Intel might have something similar just named differently | 23:55 |
| systemdlete2 | Both of these are AMD. | 23:55 |
| gnarface | oh this isn't the Intel NUC | 23:55 |
| gnarface | i'm getting two conversations confused in my head right now | 23:55 |
| systemdlete2 | (btw, an aside: I'm still having issues with the Intel 2-in-1, but we can deal with that later) | 23:55 |
| systemdlete2 | np | 23:55 |
| gnarface | yea look into that C1E thing, take temperature polling out of all your monitoring jobs, and set the cpu frequency governor to anything static | 23:55 |
| systemdlete2 | people lean on you for a lot of info, gnarface. | 23:55 |
| systemdlete2 | I wonder if I should wait until/if this happens again. I mean, it might be one of those total freaks. | 23:56 |
| gnarface | for me, it could go weeks without causing trouble | 23:57 |
| systemdlete2 | I hate these kinds of bugs the most. | 23:57 |
| gnarface | the second time in a row it froze while i was actively scping something to the drive though i started to get suspicious | 23:57 |
| systemdlete2 | But I guess everyone does. | 23:57 |
| systemdlete2 | scp! | 23:57 |
| systemdlete2 | what I was trying to do when this occurred. I wonder... maybe should just try it again... | 23:58 |
| systemdlete2 | If I can replicate it, maybe then remove syslog-ng on the testbox and see if it still can be replicated. | 23:58 |
| systemdlete2 | that's an easy experiment. | 23:58 |
| gnarface | it was often during disk i/o that the freeze up would occur, but it was hard to trace because it needed to be during the conjunction of a cpu soft-lockup and an unlucky delayed temperature sensor response | 23:58 |
| systemdlete2 | this connection may go south, but I'm still on as systemdlete on the main box. | 23:59 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!