| fsmithred_ | eyalroz, 'chattr +i /etc/resolv.conf' after you make it the way you want, so that network-manager can't write to it. | 00:30 |
|---|---|---|
| fsmithred_ | oh right, it's a symlink now | 00:31 |
| rrq | brocashelm: there's wpa_gui from the wpagui package for totally using ifupdown .. get rid of the other ones :) | 01:44 |
| brocashelm | rrq: i'll check it out, ty | 01:45 |
| brocashelm | i know there was also dhcpcd-ui or whatever | 01:45 |
| rrq | https://www.devuan.org/os/documentation/install-guides/daedalus/network-configuration.html | 01:51 |
| rrq | ... and "man wpa_action" is very good (better) | 01:54 |
| rrq | well, I wrote the first one, so it's pretty good ;) | 01:55 |
| Xenguy | re: /etc/resolv.conf: I remember many years ago realizing that various programs were messing with my settings in /etc/resolv.conf (which never used to happen)... | 03:15 |
| Xenguy | I took a look at the program resolvconf, and it just made no sense to me, so after some research, I just went with uninstalling it, and using chattr to make the file immutable... | 03:16 |
| Xenguy | Like, hands off weirdos | 03:16 |
| Xenguy | Still works to this day, so I remain willfully ignorant of why they decided to touch that file dynamically | 03:16 |
| Xenguy | In the first place | 03:17 |
| Besnik_b | Hello! I’m struggling with high processor temperatures, an AMD Phenom X6(?). I’ve changed the thermal paste 10 times in the last 15 months or so. I first thought my Arctic Silver(?) had gone dry and bought a new MX-6. After booting a new Devuan installation and opening mail, qjackctl, vlc and Firefox with 18 tabs but with umatrix enabled, from 44C it goes up to 55C degrees. I’m happy with that. I suspend the machine at nigh | 09:34 |
| Besnik_b | t and bringing it back in the morning yields 80C degrees!!! I just rebooted and opened my usual aplications and I see 52C degrees. Is there any known issues with Suspend? | 09:34 |
| gnarface | Besnik_b: not something i've heard of, but it also seems possible the sensor just isn't accurate for a few moments after resume... | 10:09 |
| gnarface | i'd look around for some known issue but i'd maybe also try pointing a fan at it | 10:09 |
| Besnik_b | gnarface, I have a box with three big fans, one not connected and a Corsair liquid cooling for the processor, The two fans work, the liquid cooling is few years old… | 10:21 |
| Besnik_b | I’ve cleaned them. | 10:21 |
| Besnik_b | Maybe I should connect the third fan with a Y-split… | 10:23 |
| linux-friend-123 | hey :) does anybody know how safe it is to replace sysvinit with runit? I have changed my mind after installation and don't want to have to do a full reinstall | 12:41 |
| leitz | https://dev1galaxy.org/viewtopic.php?id=6105 | 12:44 |
| leitz | Note that I have not tried that, just goog'd it. | 12:44 |
| linux-friend-123 | thanks | 12:45 |
| linux-friend-123 | he never replied maybe it destroyed his machine lol | 12:45 |
| brocashelm | you don't have to do a full reinstall due to runit / sysvinit | 12:48 |
| brocashelm | i've been using runit for years and it's not bad | 12:48 |
| linux-friend-123 | is it safe to just apt install runit then? or are there any other changes I need? | 12:49 |
| brocashelm | apt should automatically track dependencies if you run this command: apt install runit runit-init runit-services | 12:50 |
| brocashelm | pay close attention to the prompt; it should ask if you want sysvinit and its related packages removed for the init change to take effect (must reboot afterwards) | 12:51 |
| brocashelm | you don't have to configure anything else beyond package management | 12:51 |
| brocashelm | with runit-services, you'll be able to manage a couple of services with sv/runsv instead of service | 12:52 |
| brocashelm | runit has been available since beowulf (devuan 3), so it's been well-tested over the years and should not disappoint | 12:54 |
| linux-friend-123 | cool thank you! I will give this a go tonight :) | 13:01 |
| brocashelm | no problem | 13:05 |
| brocashelm | if you have any questions, just ask them here | 13:06 |
| brocashelm | the transition to runit was pretty seamless, even back in the beowulf (when it went stable) days | 13:06 |
| cousin_luigi | So, which init would you recommend? | 17:46 |
| n4dir | sysv seems like the best known one, in case you run into problems (in case you have questions). I use runit, but for no reason. | 17:49 |
| rkta | a | 17:57 |
| rkta | ^H | 17:57 |
| cousin_luigi | Yeah, I'll have to be conservative in this install. | 18:01 |
| cousin_luigi | By the way, will I see differences between devuan and debian on a headless machine? | 18:01 |
| cousin_luigi | Both using sysv | 18:01 |
| n4dir | probably not, to put it different: i wouldn't expect it. | 18:02 |
| Besnik_b | gnarface, would a BSD kernel (thus installation) run cooler? | 18:02 |
| n4dir | in general i don't see much of a difference | 18:02 |
| Besnik_b | (and as consequence a BSD installation) | 18:03 |
| n4dir | debian KfreeBSD probably isn't a thing anymore | 18:04 |
| Besnik_b | I meant a pure BSD installation… | 18:05 |
| cousin_luigi | Thing is, I'm seeing weird routing problems on a debian machine. Granted, it has been updated a number of times and it runs custom scripts, so who knows. That's the reason why I wanted to try devuan | 18:05 |
| Besnik_b | Or a custom kernel? | 18:05 |
| cousin_luigi | No, not in several years. | 18:11 |
| cousin_luigi | basically I want to use a non-default route for dnsmasq and I see absolutely no traffic with my old install. | 18:12 |
| cousin_luigi | But it works flawlessly with debian+systemd. But the latter, in turn, has serious trouble with ifupdown. | 18:13 |
| cousin_luigi | I've tried ifupdown-ng too, which also broke. And it doesn't support hotplug. | 18:13 |
| cousin_luigi | So I've tried installing the custom scripts package on a regular devuan, but it won't work properly and I couldn't understand why. | 18:14 |
| fsmithred_ | cousin_luigi, do the interface names in the scripts match the names in the system? | 18:52 |
| fsmithred_ | eth0 or enpcrap? | 18:52 |
| cousin_luigi | fsmithred_: Oh no, the scripts merely set up tmpfs directories to have / mounted as ro | 19:00 |
| cousin_luigi | then copy back and forth on startup/shutdown. | 19:00 |
| cousin_luigi | My concern is that I might have removed some essential package | 19:00 |
| cousin_luigi | Or that perhaps dnsmasq requires some systemd-dependent package to operate properly. | 19:01 |
| cousin_luigi | And after 4 in-place upgrades, I suspect crud may have accumulated. | 19:02 |
| fsmithred_ | I'm not familiar with dnsmasq | 19:02 |
| fsmithred_ | did you migrate from debian to devuan, or did you start with a devuan install? | 19:02 |
| fsmithred_ | dnsmasq is not a forked packaqe. Depends: netbase, dnsmasq-base, init-system-helpers (>= 1.18~), lsb-base (>= 3.0-6), runit-helper (>= 2.14.0~) | 19:05 |
| fsmithred_ | Suggests: resolvconf | 19:05 |
| cousin_luigi | Oh no, I haven't tested devuan as a router yet. The problem I had is on my old debian install | 19:11 |
| cousin_luigi | I'm installing devuan on a thumb drive to see if that makes a difference. | 19:11 |
| cousin_luigi | Then I wonder what it might be. People in #netfilter were puzzled that I couldn't see any hit in the RAW table either. | 19:12 |
| hagbard | It seems that currently some firmware packages put their stuff in /lib, while others put in in /usr/lib/. Could get my machines to properly boot again with the help of a bunch of hard and symlinks, but oh boy is that ugly... | 19:29 |
| hagbard | So should i just install the usrmerge package and then forget about the whole thing, or is there more to watch out for? | 19:30 |
| hagbard | What is the default for new devuan installations? | 19:30 |
| fsmithred_ | default is still not merged | 19:35 |
| fsmithred_ | If you choose expert install you will be asked if you want it or not. | 19:36 |
| n4dir | i already forgot about the whole thing without having done anything. Is there something to think about? | 19:36 |
| hagbard | fsmithred_: thx | 19:37 |
| fsmithred_ | n4dir, probably not. | 19:37 |
| fsmithred_ | or not yet. | 19:37 |
| n4dir | okay, i would assume so | 19:37 |
| hagbard | Also didn't care about the whole topic, until stuff started breaking today. | 19:37 |
| n4dir | what did break? | 19:38 |
| fsmithred_ | hagbard, are you on daedalus? | 19:38 |
| hagbard | Firmware wasn't found during boot. | 19:38 |
| hagbard | yes | 19:38 |
| gnarface | Besnik_b: i couldn't be sure, but i would doubt it because in general they have less hardware support and power management is patently bad for server management so it's probably the last thing they bother with on hardware they do have some support for (at least judging by my brief foray into trying to put it on ARM devices) | 20:05 |
| gnarface | s/server management/server stability/ | 20:05 |
| gnarface | you never know though, you might learn something trying either way | 20:05 |
| gnarface | the first thing i would want to do is make sure that's a real figure, not just some sensor hysteresis | 20:06 |
| gnarface | if it's really 80C while sleeping you'd be able to feel it with your hand | 20:06 |
| Besnik_b | Thank you, Sir! | 20:06 |
| gnarface | i would be radiating obvious thermals | 20:07 |
| gnarface | it would heat the room | 20:07 |
| gnarface | if the mobo is cold to the touch it's probably just sensor glitch | 20:07 |
| gnarface | if it's not... i'd point a fan at it | 20:07 |
| Besnik_b | I’ve touched one of the tubes, the one gets hot and I feel it above my body temperature for sure | 20:07 |
| gnarface | how hot though? 80C hot? to be sleeping it has to keep the RAM powered, but it shouldn't be that much power | 20:08 |
| Besnik_b | I don’t know… the readings was 80C… I’ll order a Y cable to split and feed the dormant fan first | 20:11 |
| gnarface | you don't have a regular floor fan you can just point at it? | 20:11 |
| Besnik_b | The case woud let me add 2 more fan on top and one to the bottom | 20:11 |
| gnarface | if pointing a box fan at the open case side doesn't make the apparent wake temperature budge i'd generally assume sensor glitch. if it drops to 79C i'd worry. | 20:12 |
| gnarface | full disclosure; i never sleep my machines | 20:13 |
| Besnik_b | Oh I see what you mean… silly me… no I don’t (there is a tall one around here though…) | 20:13 |
| gnarface | that would probably work | 20:16 |
| gnarface | it's just about creating enough airflow | 20:17 |
| rwp | A new-ish feature of htop allows F2 Setup -> Display options -> Also show CPU temperature, and Frequency. It's useful to see real time temps and freqs if debugging this type of thing. | 20:17 |
| Besnik_b | Should I distrust “sensors”? | 20:17 |
| gnarface | rwp: the issue is that it wakes up from sleep hot, but the mystery deepens that it goes back to normal almost immediately | 20:17 |
| rwp | Distrust? Why distrust it? | 20:18 |
| gnarface | Besnik_b: it's fine, but there's separate hardware sensor drivers for every device, sometimes multiple, and you can't always trust them | 20:18 |
| rwp | I can't imagine why a machine would be hotter sleeping than not sleeping, unless someone has closed the laptop and put it in the bag with no airflow. | 20:19 |
| gnarface | specifically, you can't always trust the calibration data they're shipped with, which is sometimes incomplete or just wrong | 20:19 |
| rwp | Which would lead me to wonder if sleeping is turning off fans but not reducing power from the cpu, then it would get hot from lack of airflow while still producing heat. | 20:19 |
| gnarface | yea, if it's really going to 80C at sleep, that's a problem that needs to be taken seriously, for sure | 20:19 |
| DelTomix | Besnik_b: If physically confirmed that the temp during hibernate is actually getting that high - it could be that with a liquid block there is a condition where it suspends, but cpu still generates some idle heat, if during suspend the water is not circulating, and the block on the cpu has no fins on it like regular heatsink it will build up | 20:19 |
| gnarface | ...but it also suggests that it might not actually be sleeping right, so it might be a problem with that driver instead of the sensors... | 20:20 |
| Besnik_b | DelTomix, no fins. | 20:21 |
| rwp | This has a water cooler and it is getting hot? | 20:21 |
| Besnik_b | Yes | 20:21 |
| gnarface | hmmm, the plot thickens | 20:21 |
| fsmithred_ | Have you tried going in through ssh to make sure it's really asleep? | 20:22 |
| gnarface | heh, yea that might be worth a try | 20:22 |
| Besnik_b | Do I need ssh for a machine 10 inch away from my hand? | 20:22 |
| DelTomix | a regular fan+heasink will disspiate the idle temp to ambient, but a liquid block will not unless the fluid is circulating. | 20:22 |
| fsmithred_ | not sure how to tell if it's really asleep by looking at it | 20:22 |
| gnarface | Besnik_b: it's just an easy way test. any network access should be sufficient to prove it's not actually asleep. | 20:23 |
| gnarface | Besnik_b: but with ssh you could also for example, check sensors | 20:23 |
| rwp | Some network cards will still answer to ping and wake-on-lan though. | 20:23 |
| gnarface | yes, but that's a hardware feature i've never seen enabled by default | 20:24 |
| gnarface | well, i guess i'm old but i usually had to change something in the bios and connect an extra cable to the network card from the motherboard | 20:24 |
| gnarface | a laptop or something modern, i guess who knows? | 20:24 |
| DelTomix | even if the machine is actually in suspend the cpu may be in a low power mode - which isn't enough to require fan active with a regular heatsink - but could still generate enough heat to build up in the liquid block case - since it accumulates | 20:25 |
| rwp | The old NICs used to need the wire for WoL but for responding to ping that's all software no wire needed. | 20:25 |
| gnarface | hmmm, interesting | 20:25 |
| rwp | Is there a way to keep the water cooler pumping all of the time? Or at least cycling on and off at some low level? | 20:26 |
| DelTomix | ^ that would solve it | 20:26 |
| fsmithred_ | power it from another computer? | 20:26 |
| * fsmithred_ hides | 20:26 | |
| DelTomix | there may also be BIOS settings for suspend mode | 20:28 |
| Besnik_b | The readings from htop and sensors are the same. 63-63C degree right now | 20:29 |
| Besnik_b | DelTomix, there is a setting for suspend mode in BIOS, I chose S3, if I not mistaken | 20:30 |
| InitSomethin | @rwp have you experimented with any software like Coolero? | 20:30 |
| InitSomethin | it's on Flathub https://flathub.org/apps/org.coolero.Coolero | 20:30 |
| DelTomix | re S3: yes - thats it i forget which are which but you should be able to google what setting does what mode | 20:31 |
| DelTomix | I have a Phenom II 1190T and even with the stock cooler 63 was like max heat under load - but it has some funny behaviors in standby conditions. (I hever use suspend/hibernate either) | 20:32 |
| InitSomethin | @rwp note: coolero was moved to CoolerControl located https://gitlab.com/coolercontrol/coolercontro | 20:33 |
| rwp | InitSomethin, I have not played with coolero. Never heard of it before. I don't have anything with a water cooler though. | 20:35 |
| Besnik_b | DelTomix, the one in question is a Phenom too | 20:36 |
| Besnik_b | I bought the case and the liquid cooler to lower noise from the machine | 20:36 |
| rwp | What?! GitLab wants me to sign in before I can see things there?! Insert many swear words here. Not happening. | 20:36 |
| DelTomix | yeah the stock cooler is gets like a get engine. :) | 20:37 |
| DelTomix | ^get/jet | 20:38 |
| InitSomethin | Gitlab wants you to log in? are you on a strange network? | 20:38 |
| rwp | Besnik_b, Have you tried hibernate-to-disk instead of suspend-to-ram? That would be a full power off state. Should be no heat problem then. | 20:39 |
| rwp | InitSomethin, Yes. It was forcing me to log in. "(!) You need to sign in or sign up before continuing." When trying the link you posted. | 20:39 |
| Besnik_b | rwp, no, because of some bad memories under my previous installation. I might try it. | 20:40 |
| rwp | I use hibernate on my laptop quite often. I have a real swap partition though. To store the ram image. Large enough for the entire machine state. | 20:41 |
| Besnik_b | DelTomix, It’s a Phenom ii x6 1100t, and I have an Nvidia 4GB RAM, 8GB Ram. | 20:41 |
| rwp | Since it is encrypted I do have to enter the decryption passphrase at boot time since it is a full power up boot which is not required for resume from suspend-to-ram | 20:41 |
| Besnik_b | I think I do have one too, real swap, rwp | 20:41 |
| Besnik_b | rwp I’ll try this night, without applications left open and see how it goes… | 20:42 |
| rwp | For hibernate config I will mention /etc/initramfs-tools/conf.d/resume with RESUME=/dev/... pointing to the swap partition so that it is in recent memory in case that information is needed. | 20:43 |
| rwp | The most common reason I need it is that I have moved disks or virtualized a machine or whatever and the swap is now not on LVM where it was before and that must be adjusted. | 20:44 |
| Besnik_b | I’ll ping you, with your permission, if help is needed. | 20:44 |
| rwp | You can always highlight my nick and my irc client is always connected. That's no problem. | 20:45 |
| rwp | But I might not be at the keyboard. Like now, a skillet is heating up on the stove and it is calling my name. "Bob... You are hungry. Eat Bobby, eat!" :-) | 20:46 |
| rwp | If you do have problems with hibernate just ask them. There are many here who would be helpful. | 20:46 |
| Besnik_b | Thank you! And Bon appétit! | 20:47 |
| rwp | Sometimes hibernate to disk is more reliable than suspend to ram because the graphics driver problems of resume from suspend are avoided. | 20:47 |
| spine-o-saurus | i lost my native resolution for my desktop on my excalibur installation | 23:09 |
| spine-o-saurus | after i ran update i had to reinstall firmware-linux-nonfree but it still didn't get resolution | 23:10 |
| brocashelm | besnik_b: i use openbsd. even with the best settings with apmd enabled, intel cpu temps are at least 8-12 C warmer than linux. this is due to the kernel not being as "advanced" at multitasking/hyperthreading/etc. when compared to linux, so power-saving settings do not come close | 23:13 |
| Besnik_b | Thank you, brocashelm! | 23:16 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!