| freaxeh | i solved the problem earlier with devuan not booting on a ryzen 7 2700 system, turned out to be an nvidia graphics card holding up the boot process. i swapped that out with an ati rx580 card and it works fine now | 04:12 |
|---|---|---|
| freaxeh | the nvidia graphics card in question is a 1030 | 04:13 |
| gnarface | cool | 04:13 |
| gnarface | weird that it would hang, but maybe it was hanging at launching the graphical login? | 04:13 |
| gnarface | some models might do that if you're using the nouveau drivers | 04:13 |
| freaxeh | no i didn't install a graphical login, just ssh server and base packages | 04:13 |
| freaxeh | this is a file server not a workstation | 04:14 |
| gnarface | oh, weird that it would matter then | 04:14 |
| freaxeh | yes | 04:14 |
| gnarface | hmm, i wonder if putting "nomodeset" in the kernel command-line parameters would have worked | 04:14 |
| freaxeh | probably | 04:14 |
| gnarface | maybe it was booting successfully and only the console was going blank? | 04:14 |
| gnarface | that's also a possible issue with some driver/model combos | 04:15 |
| gnarface | well, worth trying if you want to be sure, but you're probably better off with the rx580 either way | 04:15 |
| freaxeh | not in a file server, lol, i'll buy another amd card to put in there. | 04:16 |
| freaxeh | better than buying another motherboard that wont solve the issue | 04:16 |
| gnarface | indeed | 04:17 |
| freaxeh | anyway just wanted to let you guys know, time to get to work on reassembling this system | 04:18 |
| freaxeh | my devuan box thinks its 2017 | 05:47 |
| freaxeh | can someone whos well versed with "date" spit out the correct command to set it | 05:48 |
| freaxeh | my bios doesn't have the option to set the date | 05:48 |
| fluffywolf | apt-get install ntpdate? :) | 05:48 |
| freaxeh | it thinks the repositories are out of date | 05:48 |
| freaxeh | i tried apt update earlier | 05:48 |
| freaxeh | E: Release file for http://deb.devuan.org/merged/dists/daedalus-updates/InRelease is not valid yet (invalid for another 2780d 2h 56min 6s). Updates for this repository will not be applied. | 05:49 |
| fluffywolf | date -s "Sun Aug 11 08:49:05 PM PDT 2024" | 05:49 |
| freaxeh | apt install ntpdate worked | 05:49 |
| freaxeh | cheers | 05:49 |
| freaxeh | all fixed :) | 05:50 |
| drizzt | hello ! | 08:20 |
| drizzt | back with the devuan upgrade problem on aarch64 (arm64) | 08:21 |
| gnarface | fyi there's also #devuan-arm, though it's even slower over there... | 08:22 |
| drizzt | I strongly think that the problem comes from some programs segfaulting | 08:22 |
| drizzt | before the critical part of the update I still can install other packages, but I already have libc updated, and some programs already segfault | 08:24 |
| drizzt | xz, vim | 08:25 |
| drizzt | I'll check #devuan-arm | 08:28 |
| eyalroz | bash completion is acting up on Excalibur... | 11:22 |
| eyalroz | https://unix.stackexchange.com/q/781758/34868 | 11:22 |
| freaxeh | is it ok if i idle in here? i'm eager to learn more about devuan | 12:04 |
| rrq | certainly | 12:05 |
| rrq | (most of us do :) | 12:05 |
| fsmithred | You'll be like most of the folks in here - silent. | 12:05 |
| freaxeh | oh i doubt that fsmithred | 12:05 |
| freaxeh | i'm a bit of a talker | 12:05 |
| fsmithred | Feel free to either ask or answer a question. | 12:05 |
| freaxeh | and have plenty of questions about my devuan box | 12:06 |
| drizzt | we are here to help (and get some help too from time to time) | 12:42 |
| adhoc | indeed. | 13:00 |
| ibanja | I have a strange problem. There have been two instances where I’ve modified scripts in two different directories and the next day the have reverted back to the original script from several modifications earlier in the one case. | 13:37 |
| ibanja | I am using inotifywait to monitor the directory, but is there a way to monitor what process is causing the changes? | 13:37 |
| drizzt | make the script read-only and check the logs for errors ? | 13:38 |
| drizzt | (or rootfs read-only if you fear that the modifications are made by a root process) | 13:39 |
| ibanja | good idea. | 13:40 |
| drizzt | or a combination of inotify + lsof, but I don't know if the lsof will comme in quickly enough to see who has the file openned | 13:41 |
| ibanja | I just found a tool called auditctl (auditd package). Maybe that will tell me something. | 13:44 |
| buZz | could also perhaps be some aggressive and incorrect caching? | 13:46 |
| ibanja | buZz: how is that? | 14:10 |
| ibanja | What might cause that? I am running smartctl on the drive to check for errors. | 14:10 |
| ibanja | smartctl showed no errors. | 14:12 |
| buZz | i dont know what filesystem you're using or anything | 14:21 |
| buZz | but i've seen overeager caching sometimes not write changes away, for instance | 14:21 |
| drizzt | if you fear for a caching problem, you can run "sync" after your modifications | 14:25 |
| drizzt | but this should not happen as the kernel and VFS should grant you that you always access the latest version of the file | 14:26 |
| buZz | drizzt: yeah but maybe its a write-back cache on a hw device that doesnt listen to 'sync' | 14:30 |
| buZz | wouldnt be the first time i see such misconfigged caches, anyway ;) | 14:30 |
| ibanja | I'm using ext4, but I'm in the process of trying to get devuan installed on a zfs mirror. | 14:47 |
| plasma41 | ibanja: Where are the scripts within the filesystem? | 16:25 |
| ibanja | plasma41: they were in ~/dev/cpp/projects/fpa-black-scholes and ~/dev/scripts/bash/projects/fpa-general-scripts | 16:30 |
| al1r4d | onefang, hi sir, do you get new server rental? | 16:33 |
| plasma41 | ibanja: Hmm, both within your home directory. That pretty much rules out some weird interaction with the package manager. | 16:35 |
| ibanja | plasma41: It's a mystery. I haven't changed anything config-wise that I can think of. | 16:40 |
| ibanja | I found the issue of why my file keeps reverting to an old version. I automate packaging my scripts into deb packages and a typo in the build script overwrote my new source files with old ones. It was a matter of a missing semi-colon. OOPs | 19:24 |
| djph | ibanja: it's alwasya a semicolon! | 19:43 |
| ibanja | djph: Isn't it though. :) | 20:01 |
| mathew | Hello. I want to set up syncthing on my homeserver. the program comes with an systemd file on install. but on my homeserver runs devuan chimaera. Now i need to write an init script to start the syncthing servie at boot. did someone here did it already? | 23:20 |
| rrq | mathew: you might know; the start point for diy init script is "man init-d-script" | 23:37 |
| mathew | thx! i will read this! | 23:38 |
| mathew | Now i did find a script on the internet. i will use it and change the necessary aspect with man init-d-script in mind. https://gist.githubusercontent.com/typomedia/3fe16c0e8214ce60b0b2/raw/23595ba29ce853da85e26e6b376ad52fb9f0b861/init.syncthing.sh | 23:40 |
| rrq | looks fine to me; maybe something to submit to the orphan-sysvinit-scripts package; I'm not sure about their process for adding scripts, but maybe a bug report is good | 23:51 |
| rwp | I also gave it a quick look-over and it looks okay to me too. Nothing of note jumped out at me. | 23:52 |
| rrq | mathew: i.e. an email to submit@bugs.devuan.org, where first line is "Package: orphan-sysvinit-scripts" and second is "Version: ceres", and then up to you for the rest | 23:53 |
| rwp | The only thing I would have added would be the optional inclusion of /etc/default/$NAME which if present could override things like TIMEOUT and so on. But that's really a minor thing. | 23:53 |
| rrq | mmm maybe it's to submit@bugs.debian.org rather, since its not a forked package | 23:55 |
| mathew | one problem is that one has to add an user "syncthing" which is not done at apt install. so the scirpt failes if no user "syncthing" is present. | 23:55 |
| mathew | but if the user is present, the script runs fine so far. | 23:56 |
| rwp | Users are normally added in the package postinst script. | 23:56 |
| rrq | that's a bug for installing syncthing rather | 23:56 |
| rrq | (what rwp said) | 23:56 |
| rwp | A pet peeve of mine is when postinst scripts unconditionally add a user. Which works great on a fresh install. But on an upgrade then complains that the user already exists. | 23:56 |
| rwp | postinst scripts should test first and add user if the user does not already exist. That also handles a case where the local admin customized things. | 23:57 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!