Xenguy | I think he's on the wrong track entirely | 00:03 |
---|---|---|
Xenguy | He's promoting the idea of infecting the BSD ecosystem with systemd now... | 00:03 |
Xenguy | Ironically LP was entirely happy to throw the BSDs under the bus in his coding of systemd | 00:04 |
Xenguy | The BSDs should realize that systemd in linux is enouraging people to explore the BSDs, people that might not have bothered otherwise | 00:05 |
buZz | i do keep hearing devuan mentioned more and more in general user channels | 00:05 |
buZz | good times | 00:05 |
buZz | (like ##hardware) | 00:06 |
Xenguy | Right glad I am to hear it | 00:06 |
DrDeke | i admit that _some_ of my dislike of systemd is attributable to resistance-to-change or not seeing why the change is beneficial. but as i use it more, there are also plenty of things about it that i actually dislike for reasons. | 00:11 |
buZz | my biggest annoyance is the lack of startup order | 00:15 |
buZz | which kills even many fresh install of it from recovering properly from suspend/hibernate | 00:16 |
buZz | sooo many of my /etc/rc.local just got 'sleep 60' at the top to make SURE it runs last | 00:16 |
buZz | ffs | 00:16 |
roo^y | does booting the new Refracta 9's boot menu, & using the install feature, have more install options than starting Refracta live desktop, & using Menu > System > 'Refracta Installer' feature? or is the only difference that from the live desktop, there will be less hardware resources available? | 05:36 |
fsmithred | roo^y, what install feature in the boot menu??? | 05:38 |
fsmithred | there is only refractainstaller | 05:38 |
fsmithred | what options do you want? | 05:39 |
roo^y | ..well it was late last night, the pc booted refracta's boot menu, & i skipped past what i assume was install options, & chose the US live boot (**not those specific words) but i was happy i got to try it live 1st, when i didn't know if it had a live desktop boot option or not | 05:41 |
roo^y | ) | 05:41 |
fsmithred | no install option in the menu | 05:42 |
fsmithred | any changes you make in the live system will be copied in the install | 05:42 |
fsmithred | but adding software in live is not a good idea - you can run out of ram fast | 05:43 |
roo^y | ok. i haven't shut it down this live session to see it's boot options again | 05:43 |
fsmithred | bed time here. Talk fast, or I'll catch you in the morning (mine) | 05:44 |
roo^y | i'll leave you alone. goodnight. | 05:44 |
fsmithred | there's an install guide with screenshots | 05:44 |
fsmithred | I'll get the link | 05:45 |
roo^y | i didn't know if the ISO i installed to USB uses rEFInd or Grub, but it works | 05:45 |
furrywolf | meep, a roo in here? | 05:45 |
roo^y | yup, don't shoot. allergic to bullets | 05:45 |
fsmithred | https://devuan.org/os/documentation/install-guides/live-gui.html | 05:45 |
furrywolf | I'm far, far more likely to randomly pet and/or hug roos. :P | 05:46 |
roo^y | thankyou again. snore hard | 05:46 |
fsmithred | iso will boot isolinux on bios, grub on efi | 05:46 |
roo^y | ok | 05:46 |
fsmithred | grub-efi-amd64 is installed. If you boot bios, the installer will find the grub-pc package and ask you if you want to copy the files. Yes, you do. | 05:47 |
roo^y | well if you see a roo in the wild. i hope it isn't a big male, or any agressive roo that claws, bites, kicks, & tears you to shreds | 05:47 |
fsmithred | g'night and good luck | 05:47 |
roo^y | thx | 05:47 |
roo^y | I have a problem with refracta's installer. i've chosen to manage partitions myself, but it wants me to have already done this, when I wanted to use gparted to put a fat32 boot partition infront of the main ext4 "sda1" (it let me choose sda1 for boot, then let me choose it for refaracta install, but said i can't use it for both *well open gparted & i'll fix that :P * ) | 13:24 |
fsmithred | roo^y, you can't use a separate partition for /boot and for the rest of the OS, too. | 13:28 |
fsmithred | why fat32 boot partition??? | 13:28 |
fsmithred | what exactly are you trying to do? | 13:29 |
roo^y | i already setup ext4 for sda1 & swap for sda2. i was going to change it with gparted to add a partition. sda1 would now be a small UEFI FAT partition. a slightly reduced sda2 would be the main ext4, with a unchanged swap "sda3" | 13:31 |
fsmithred | so why is that a problem? Should be easy. | 13:32 |
fsmithred | or is the system already installed? | 13:32 |
roo^y | during installation i chose yes, hit me up with gparted & i'll get to work (so to speak) instead it gave me options of choosing where to put boot & main install, not going to gparted | 13:33 |
fsmithred | gui or cli installer? | 13:34 |
roo^y | gui | 13:34 |
fsmithred | can you paste the log somewhere so I can see? | 13:35 |
roo^y | i closed the installer | 13:36 |
roo^y | can gparted be accessed from the live desktop | 13:36 |
roo^y | ? | 13:36 |
fsmithred | look at /var/log/refractainstaller.log | 13:36 |
fsmithred | yes | 13:36 |
fsmithred | oy | 13:37 |
fsmithred | oy | 13:37 |
fsmithred | givl | 13:37 |
fsmithred | it's in System or Preferences in the menu | 13:37 |
roo^y | this is not spam https://www.irccloud.com/pastebin/UxOhszzW/refractainstaller.log | 13:37 |
fsmithred | if you do all your partitioning before running the installer, you can check the 'do not format' box. | 13:38 |
roo^y | that might be my problem | 13:38 |
roo^y | i may have prematurely ticked Do Not Format | 13:39 |
fsmithred | that should not stop you from running gparted | 13:41 |
ejr | what's your average cpu temperature when you are running devuan on your systems without any too cpu-extensive processes? (I might also ask: is 60-70°C normal?) | 13:43 |
r3boot | that kind of depends on the type of cpu, the workload running on there, but also the airflow inside of the casing + the ambient temperature + humidity :) | 13:44 |
fsmithred | ejr, that sounds high. It varies depending on the cpu model. Right now I'm at 24 and 16. | 13:44 |
fsmithred | I have a laptop that runs up to 60-70 if I'm watching video. | 13:45 |
r3boot | I have a hppa2 that runs 90C celcius @ idle; I've got atoms that run ambient room temperature when idle. Video is a good way to stress cpu's | 13:46 |
r3boot | ejr: what is your system doing atm? How many cores do you have in use? | 13:46 |
ejr | i think it has to do with my thinkpad ultrabase, because the temperature only goes that high when it's plugged into that (or when i am doing cpu-intensive stuff like, as you said, watching videos) | 13:46 |
r3boot | Does that block airflow in any way? | 13:47 |
ejr | 2 cores, x60 thinkpad (i.e. rather old) and nothing running except tmux, ssh client and vim | 13:47 |
r3boot | Could also be that your cpu is running in performance mode (instead of powersaving mode) as soon as you plugin your AC adapter | 13:47 |
r3boot | but, I'd check which processes are running, and how much they are currently are consuming | 13:47 |
ejr | it also happens with other x60's with the same architecture and system, so I think it's really due to the ultrabase mostly. | 13:47 |
roo^y | i can find gparted in synaptic package manager. i can't run it on live desktop | 13:48 |
r3boot | if your box is really idle, i'd suspect airflow inside of the casing (dust, etc) | 13:48 |
r3boot | (get a can of compressed air to clean that out) | 13:48 |
ejr | yeah, i think i should do that | 13:48 |
r3boot | also, it can help to re-add thermal paste to your heatsinks, that dries out over time | 13:48 |
r3boot | (but, do that step only if the compressed air step doesnt help) | 13:49 |
ejr | ah right, have heard about that. thanks for the advice | 13:49 |
roo^y | "less thermal paste is more" i've done a couple of repastes. only need the tiniest amount | 13:50 |
ejr | this is where youtube tutorials are actually useful (or images) i guess | 13:50 |
r3boot | ejr: mja, like roo^y mentions, repasting can be a bit tricky, so only do that if all else fails :) | 13:52 |
r3boot | with the compressed air: first, loosen up material by blowing agaist the direction of the airflow inside the case, and follow up by blowing in the direction of the airflow. Should have it clean in less then 5 mins or so | 13:53 |
roo^y | i busted the plastic fan clamp/heatsink clamp once | 13:53 |
r3boot | Meh :) | 13:53 |
* r3boot once upgraded a cisco pix (hardware firewall, slow as molasses, but enterprise grade) with the biggest p3 processor I could find, so I could run ASA firmware on there (the successor to pix, also enterprise grade). Took me quite a few attempts to get the amount of paste correct before the system ran stable | 13:55 | |
roo^y | Installer: 'Partitioning' clicking on button [Run Gparted] goes to 'Select EFI partition' clicking on button [partitioner] loops back to start of process 'Partitioning' | 14:07 |
roo^y | making Gparted a non-existing ghost, in this instance | 14:08 |
fsmithred | roo^y, I'm about to try it. It's never done that before. | 14:15 |
roo^y | clicking [Gparted] button again here, goes to 'Format EFI' "Format new efi partition?" button [Yes] goes to '/boot partition' "Select a partition for /boot" | 14:15 |
roo^y | So as I'm failing to run gparted from the desktop, i can go back to my prev. OS & use gparted to add a boot partition, or i could add the bootloader to this USB i've booted live desktop off of (it has a sda1 i unnecessarilly made with the unneeded rEFInd) | 14:18 |
roo^y | & put the bootloader where it needs to be at a later time | 14:19 |
roo^y | s/needs to/should/ | 14:20 |
fsmithred | check sha256sum of the iso to make sure the download was good | 14:21 |
fsmithred | you're using refracta iso with Feb 6 date in the filename? | 14:21 |
fsmithred | 20190206 | 14:21 |
fsmithred | why do you need a separate boot partition? | 14:22 |
fsmithred | in live session, if gparted won't start from menu, run 'sudo gparted' in terminal | 14:23 |
r3boot | fsmithred: she/he has a uefi-based system, which requires a dedicated boot partition | 14:23 |
roo^y | i've only assumed UEFI/EFI needs it's own FAT partition for my mele pcg35 opo (no BIOS) | 14:24 |
roo^y | < - penis-person | 14:24 |
* r3boot also has a separate /boot (MBR-based, on an usb stick), with my full sdd luks encrypted. | 14:24 | |
roo^y | external appendage | 14:25 |
fsmithred | r3boot, a boot partition is a separate partition for /boot | 14:25 |
fsmithred | not the same as the efi partition | 14:26 |
r3boot | *Linux boot partition then ;) | 14:26 |
r3boot | I was talking about the hardware boot partition.. but lets not get stuck on semantics | 14:26 |
roo^y | sudo gparted didn't work (I had my douts, as it didn't work from the root terminal) | 14:26 |
fsmithred | well, he's using an installer that has a checkbox for using a separate /boot partition | 14:27 |
fsmithred | something odd is going on | 14:27 |
Wonka | since efi-grub, a "BIOS boot partition" is no longer necessary for EFI systems to be able to boot... | 14:28 |
Wonka | just have EFI boot efi-grub, which then can boot Linux | 14:29 |
fsmithred | roo^y, what iso are you using? | 14:29 |
Wonka | needs an EFI System Partition though (FAT) | 14:29 |
roo^y | is there a cmd that will list it's build? | 14:31 |
fsmithred | yeah | 14:32 |
roo^y | i assume it was the same listed the other day on distrowatch | 14:32 |
fsmithred | ls /lib/live/mount/medium | 14:32 |
fsmithred | should show a package list directory that has the date in it | 14:32 |
roo^y | isolinux pkglist_refracta9_xfce_amd64-20190206_1724 | 14:34 |
fsmithred | ok, that's the latest one | 14:34 |
fsmithred | you put it on usb or cd? | 14:35 |
roo^y | usb | 14:35 |
fsmithred | how? dd/cat or something else? | 14:36 |
roo^y | dd to partition 2 (as i wrongly thought i needed refind | 14:36 |
roo^y | which i'd installed to partition 1) | 14:37 |
fsmithred | ok, I've never tried it that way, and I didn't know it would work | 14:38 |
fsmithred | normally, you dd it to the whole device | 14:38 |
roo^y | Please CLOSE all running applications NOW (you silly bald aussie, n we'll install to SSD & boot to USB like r3boot) :P | 14:39 |
fsmithred | I keep refind on an old 256mb usb stick by itself | 14:39 |
roo^y | i see | 14:39 |
fsmithred | you're running refractasnapshot? | 14:40 |
roo^y | i don't believe i've used that feature | 14:40 |
fsmithred | ok, I don't recall the installer saying to close all running apps | 14:40 |
fsmithred | snapshot does say that | 14:41 |
roo^y | that installer guide link you gave me even has close all apps (must be modern instructions) | 14:41 |
fsmithred | yeah, it's in the summary window. I forgot that was there. | 14:44 |
roo^y | it's killall -9 firefox o'clock here | 14:44 |
fsmithred | lol | 14:44 |
roo^y | i forgot i have irccloud. i'm not going anywhere | 14:44 |
roo^y | firefox: no process found | 14:45 |
roo^y | than you shall stay running | 14:45 |
roo^y | install completed, but i don't think i got boot installed to USB. i may have to boot back into Easy Pyro OS & use gparted in there to create a boot partition ..then install boot from the live desktop without installing refracta again | 15:10 |
roo^y | well i see grub-efi-ia32_2.02~beta3-5+deb9u1_amd64.deb in the filesystem which might do it :s | 15:16 |
fsmithred | roo^y, what does fdisk -l show? | 15:17 |
fsmithred | does your hard drive already have a bootloader and another linux on it? | 15:18 |
fsmithred | and if so, does that other linux boot in uefi mode? | 15:19 |
fsmithred | and if so, does your hardware need a 32-bit uefi bootloader? | 15:21 |
roo^y | I closed pc down. booted to boot menu of refracta on usb. entered grub. pressed c for cmd line | 15:33 |
roo^y | was a fresh install on new ssd. no previous bootloader | 15:34 |
specing | roo^y: installing GRUB means writing GRUB's first stage loader into MBR and the second stage loader (~500KB) in the space between MBR and first partition (not sure about the second part) | 15:36 |
specing | and the third stage (modules) into /boot partition | 15:36 |
fsmithred | except on uefi where installing grub means putting a bootloader in the efi partition | 15:37 |
specing | in case you have UEFI, I think the above simplifies into just installing a lot of files into the EFI partition, but I have no UEFI hardware to see this myself | 15:37 |
fsmithred | roo^y, what's your partition scheme | 15:37 |
* specing GRUB in bios flash ftw | 15:38 | |
roo^y | all of ssd is partitioned to ext4, except for the last 4GB, SWAP | 15:39 |
roo^y | originally, I gave up trying to leave a gap before the main partition, as it was messing up the other partition sizes, of leaving 4GB after it | 15:44 |
fsmithred | you need at least two partitions, three if you want a swap partition. | 15:45 |
fsmithred | one efi partition, fat32 with esp and boot flags | 15:45 |
fsmithred | one linux partition | 15:45 |
fsmithred | optional swap partition. (The installer will make a 256mb swapfile if you don't have a swap partition) | 15:45 |
fsmithred | since you already installed, you could shrink the existing partition to make room for efi partition. | 15:46 |
fsmithred | 100mb is more than enough | 15:47 |
roo^y | once made, what's the best way to install to efi partition? | 15:48 |
fsmithred | boot from grub command line into the installed system | 15:49 |
fsmithred | run 'grub-install' (no device!!!) | 15:49 |
fsmithred | then update-grub | 15:49 |
specing | 256 MB swapfile haha | 15:49 |
specing | this is like those other extremes, such as 1+GB /boot | 15:49 |
fsmithred | I've seen as much as 80mb of swap get used (with 2G ram) | 15:50 |
specing | fsmithred: just 80MB? | 15:50 |
fsmithred | yeah | 15:50 |
specing | I've seen all off 8GB swap used with 8GB RAM | 15:51 |
fsmithred | other than hibernating? | 15:51 |
specing | there is no upper limit | 15:51 |
specing | fsmithred: yes | 15:51 |
fsmithred | what are you doing? | 15:51 |
specing | compiling firefox/libreoffice on tmpfs tends to do that | 15:51 |
specing | or QT | 15:51 |
fsmithred | oh, I'm using 88 right now | 15:52 |
roo^y | I'm on grub cmd now. can I boot refracta on ssd from here? | 15:52 |
fsmithred | yes roo^y | 15:52 |
Wonka | jftr, there's the grml-rescueboot package which adds entries to grub's menu for all existing GRML images in /boot/grml/ | 15:52 |
fsmithred | set root=(hd0,gpt1) (if it's the first partition) | 15:52 |
Wonka | GRML is a very nice live linux system | 15:52 |
fsmithred | linux /vmlinuz ro root=/dev/sda1 | 15:52 |
fsmithred | initrd /initrd.img | 15:52 |
fsmithred | boot | 15:52 |
Wonka | specing: and GRML is the reason why I have 2GB /boot. | 15:53 |
fsmithred | yeah, I've put a live iso in a big /boot partition before. It's handy. | 15:53 |
Wonka | for rescuing stuff when the main system is f*cked up somehow... | 15:54 |
Wonka | I have recovered my system that way several times... | 15:54 |
roo^y | error: '/vmlinuz' not found | 15:56 |
specing | Wonka: GRML? | 15:57 |
specing | Wonka: I have old systems available (btrfs snapshots) | 15:57 |
Wonka | specing: grml.org | 15:57 |
specing | fsmithred: that swap is in RAM btw (zswap, 40% reduction with compression) | 15:57 |
Wonka | specing: well, my notebook is fully encrypted (excluding /boot, of course), so if the initramfs is broken, I can't just revert to an earlier snapshot because I can't open the crypto | 15:58 |
specing | Wonka: my notebook is fully encrypted (including! /boot, of course) | 15:59 |
Wonka | specing: using GRML, I can cryptsetup LUKSopen (or whatever), mount, chroot, repair initramfs, reboot, done | 15:59 |
Wonka | specing: where's your kernel and initramfs? | 15:59 |
specing | Wonka: in /boot | 15:59 |
specing | Wonka: GRUB can read LUKS since a decade ago | 15:59 |
Wonka | specing: mhkay... nice. I'd still want someplace unencrypted for GRML... it's just too handy. | 16:00 |
Wonka | also, I don't actually boot using grub, but rEFInd | 16:01 |
specing | as you wish, my computer is setup to only boot my GPG-signed kernels and initramfs, so having an unencrypted system would be counterproductive | 16:02 |
Wonka | so, your kernel is behind your crypto, and you still sign it... hmm. | 16:02 |
Wonka | also, does your EFI check that your grub is unmodified? | 16:03 |
specing | Wonka: it is more a measure ensuring no other kernel is booted than ensuring that the encrypted ones are real | 16:04 |
specing | I don't have EFI | 16:04 |
Wonka | even worse ;) | 16:04 |
specing | why | 16:04 |
jonadab | Not having EFI is a good thing. | 16:04 |
Wonka | does your BIOS do anything to secure your grub is correct? | 16:04 |
jonadab | But it's getting to be a real pain to find hardware with a decent BIOS these days. | 16:05 |
Wonka | is there any Secure Boot without EFI? | 16:05 |
specing | Wonka: no, because it being incorrect would mean the whole "BIOS" has been compromised | 16:05 |
jonadab | No, Secure Boot was specifically designed to make using open-source software difficult. | 16:05 |
specing | Wonka: my GRUB is part of the "BIOS", it is not installed on a hard drive or ssd | 16:05 |
Wonka | specing: hm, coreboot? | 16:06 |
specing | Wonka: better, Libreboot | 16:06 |
Wonka | specing: well, my HW has that fruit with a bite on it's lid, unlikely I'll get something like that working there... | 16:07 |
fsmithred | roo^y, what's going on? | 16:08 |
Wonka | nearly 8 years old, that thing... and I hope it'll last some more. there's nothing much on the market to replace it without losing features. | 16:09 |
roo^y | I'll pack it in for today. set root cmd didn't throw error in live usb's grub. then next cmd error of file /vmlinuz not found | 16:12 |
fsmithred | oh, maybe it's hd1 | 16:13 |
fsmithred | you can use tab-completion on grub command line | 16:13 |
fsmithred | can help you find stuff | 16:13 |
fsmithred | set root=(hd1, | 16:13 |
specing | Wonka: if your fruitbite hardware is old enough, it might actually be supported | 16:13 |
fsmithred | then hit TAB and it should list possible choices | 16:13 |
Wonka | specing: Early 2011 | 16:13 |
specing | core2duo macbooks are supported in Libreboot | 16:14 |
specing | Wonka: you can buy used thinkpads off ebay and similar places | 16:14 |
specing | Wonka: if you don't need a laptop or x86, you can look into talos workstations | 16:14 |
Wonka | specing: i7-2620MQ, and I don't want to go back to something slower anymore. Iff some notebook with something current was supported by libreboot, I might look into trying it | 16:16 |
specing | Wonka: there are in-production supported arm notebooks, but it'd be a downgrade | 16:17 |
specing | Wonka: only talos would be an upgrade, but it is not a laptop | 16:17 |
Wonka | <- waiting for reasonably fast RISC-V notebooks... | 16:17 |
Wonka | *dream* | 16:17 |
specing | it won't happen | 16:18 |
specing | we'll get POWER9 notebooks sooner | 16:18 |
specing | heck, maybe IBM sees our thirst and releases a notebook POWER9 | 16:18 |
specing | cpu | 16:18 |
MinceR | POWER-thirst? | 16:18 |
specing | otherwise it'll be 4core/16thread 90W TDP CPU handicapped to 35W for laptop use | 16:19 |
specing | MinceR: high-end POWER9 are over 200W per socket iirc | 16:19 |
specing | max 4 sockets per motherboard | 16:19 |
specing | full open source firmware (including microcode) | 16:19 |
roo^y | it is hd1, but still no /vmlinuz | 16:19 |
MinceR | https://www.youtube.com/watch?v=qRuNxHqwazs | 16:19 |
sxpert | specing: comes with a tesla powerpack as the battery ? | 16:19 |
MinceR | i don't expect anything good from ibm since they became card carrying apple shills, bought red hat and didn't stop any of their evil dealings | 16:20 |
specing | sxpert: no need for that, 1kg of li-ion would hold it up for hours | 16:21 |
specing | sxpert: 35W-capped 4core CPU in laptop use case, that is, servers use usually hooked to large lead acid UPS | 16:21 |
specing | sxpert: 35W-capped 4core CPU in laptop use case, that is. servers are usually hooked to large lead acid UPS* | 16:22 |
sxpert | specing: we replaced the lead-acid with Li-ion UPS here | 16:22 |
sxpert | lasts much longer for cheaper ! | 16:22 |
specing | MinceR: they open sourced all firmware | 16:22 |
sxpert | specing: including CPU microcode ? | 16:22 |
specing | MinceR: you can even license the design and make your own processors yourself | 16:22 |
specing | sxpert: YES | 16:23 |
MinceR | well, that's something | 16:23 |
specing | China is licensing OpenPOWER (POWER9) and making their own, as they do not trust intel/amd | 16:23 |
specing | Russia are making their own ARM-based SoCs nowadays | 16:24 |
Wonka | I'd prefer LiFePO4 batteries though. A little less capacity, but much less BOOM when damaged | 16:24 |
specing | Wonka: don't damage them :P | 16:24 |
specing | Wonka: cylindrical li-ion is very safe | 16:25 |
Wonka | specing: can't always avoid it when mobile... | 16:25 |
specing | I have high energy density cylindrical cells all over the place, exposed | 16:25 |
Wonka | not all notebook batteries are made from cylindrical cells... | 16:25 |
sxpert | specing: russia is designing their own RiscV processors for their newfangled govt-use cellphones | 16:26 |
specing | samsung tests random samples of their cells to withstand impact, fall, oven,... | 16:26 |
sxpert | Wonka: EDONTCARE, thing is outside | 16:26 |
Wonka | the DELL at my last workplace had it's battery cushion out a little | 16:26 |
specing | they will leak, not explode | 16:26 |
specing | the only danger is if you keep cylindrical li-ion in waterproof containers (e.g. high quality flashlights) | 16:27 |
sxpert | specing: what happens then ? | 16:27 |
MinceR | LiFePO4 is also longer lived, afaik | 16:27 |
specing | but a normal plastic laptop battery will just burst at the seams if enough pressure builds up | 16:27 |
specing | sxpert: if you are lucky the lens/button will give out | 16:28 |
specing | sxpert: if not, it'll explode like a pipe bomb | 16:28 |
Wonka | specing: and if you don't keep them watertight, you risk short circuits when you expose your device to water. like rain. when outside. | 16:28 |
specing | MinceR: it is, but the low energy density makes it not worth it for person-mobile applications | 16:28 |
specing | MinceR: but it is fine for e.g. electric bus use. It can also be recharged much faster | 16:29 |
sxpert | specing: still less dangerous than cop's grenades... | 16:29 |
specing | sxpert: yes, li-ion flashlights are not meant to kill people | 16:29 |
specing | haha | 16:29 |
specing | well, the $3 china zoomies might be, but fortunately those are not waterproof | 16:30 |
MinceR | iirc in a few years ordinary Li-Ion's energy density drops below that of LiFePO4 | 16:30 |
sxpert | specing: 2 dead and 7 lost hand/foot and counting... | 16:31 |
Wonka | from french cop flash bangs? | 16:31 |
sxpert | yeah | 16:31 |
specing | MinceR: if they are abused as usual, meaning charged to >4V and heated by laptop cpu | 16:32 |
specing | MinceR: otherwise no | 16:32 |
specing | MinceR: electric vehicle and grid balancing li-ion packs are designed to last decades | 16:32 |
Wonka | uuuh, overcharing is evil | 16:32 |
Wonka | +g | 16:32 |
specing | Wonka: yes. But everyone does it | 16:32 |
specing | going from 4V to 4.2V gives device manufacturers another runtime hour in reviews | 16:33 |
specing | but nobody does battery life reviews after a year of use | 16:33 |
Wonka | hmf. introduce a setting for charge stop voltage? | 16:33 |
sxpert | Wonka: bad BMS devices are very cheap... | 16:34 |
Wonka | "the higher, the more runtime, but less lifetime"? | 16:34 |
specing | Wonka: some devices have it, e.g. thinkpads | 16:34 |
specing | Wonka: indirectly, via stop percentage | 16:34 |
sxpert | specing: most good quality phones / tablets | 16:34 |
specing | sxpert: this is not due to BMS | 16:34 |
MinceR | how do i set that? | 16:34 |
r3boot | Not really a review, but the thinkwiki does have a lot of info wrt battery maintenance / care https://www.thinkwiki.org/wiki/Battery | 16:35 |
sxpert | specing: no, I said "bad BMS", those with funky constants | 16:35 |
specing | MinceR: echo 80 > /sys/devices/platform/smapi/BAT0/stop_charge_thresh | 16:35 |
specing | sxpert: the BMS does not set charge voltage | 16:35 |
specing | sxpert: the BMS only interrupts charging if the voltage rises too much, which is usually 4.25V. But this should never happen during normal operation. | 16:36 |
MinceR | specing: thx | 16:36 |
specing | r3boot: the german one should have even more. | 16:36 |
sxpert | specing: I'm talking about combination charger / bms you get for 2$ from aliexpress | 16:36 |
specing | sxpert: thats it not a bad BMS, that is a bad buyer | 16:37 |
specing | clueless buyer gets hardware appropriate for clueless buyers ;D | 16:37 |
specing | sxpert: and I haven't seen combo charger+BMS on single PCB being sold outside DIY powerbank | 16:39 |
sxpert | specing: tre | 16:42 |
specing | though if you ask me, a complete Xiaomi powerbank is cheaper than DIY | 16:43 |
sxpert | specing: true, and you can find lots of those in Huaqiangbei | 16:43 |
specing | you can also use a thinkpad battery as powerbank by attaching a buck converter to its power pins | 16:45 |
specing | Unfortunatelly all my thinkpad batteries are a decade old and have lost 50% of their capacity. Will have to re-cell them some day. | 16:48 |
specing | and reverse engineer the BMS in them so a FOSS one can be made | 16:48 |
specing | given that there is no high-performance open harwdare laptop on the horizon, I'll probably be stuck with core2duo thinkpads for the next decade at least | 16:54 |
cosurgi | guys, where can I find package xserver-xorg-core-dbgsym ? | 17:08 |
cosurgi | I need to install it from repository, so that it will work also for currently running Xorg sessions. | 17:09 |
cosurgi | I tried "deb http://pl.mirror.devuan.org/merged-debug/ ascii main non-free contrib" but that didn't work. | 17:09 |
cosurgi | I have xserver-xorg-core 2:1.19.2-1+deb | 17:10 |
KatolaZ | cosurgi: cosurgi you need to get it from the debian repo | 17:10 |
cosurgi | And I have even compiled my own xserver-xorg-core-dbgsym_1.19.2-1+deb9u5_amd64.deb, but if I install it, then I guess that it won't work for a currently running session. | 17:11 |
KatolaZ | cosurgi: we are not merging -debug repos | 17:11 |
KatolaZ | we actually should | 17:11 |
KatolaZ | but wre aren't atm | 17:11 |
cosurgi | So this one: deb http://debug.mirrors.debian.org/debian-debug/ stretch-debug main ? | 17:12 |
KatolaZ | http://ftp.us.debian.org/debian/pool/main/x/xorg-server/ | 17:13 |
KatolaZ | just get it from the pool | 17:13 |
KatolaZ | oh it's not there sorry | 17:14 |
cosurgi | ok. A package xserver-xorg-core-dbgsym is available and I installed it. | 17:14 |
KatolaZ | yep | 17:14 |
cosurgi | I added this to /etc/apt/sources.list | 17:14 |
cosurgi | but I need to verify that it's binary compatibile with my presently running Xorg | 17:15 |
cosurgi | how would I do that? | 17:15 |
KatolaZ | cosurgi: we haven't forked xorg | 17:15 |
cosurgi | aptit | 17:15 |
KatolaZ | so the xorg package comes directly from Debian | 17:15 |
cosurgi | (oops, not this terminal ;) | 17:16 |
KatolaZ | hence, the corresponding dbgsym should be binary compatible | 17:16 |
cosurgi | ok. so it just should be compatible, you say? :) | 17:16 |
KatolaZ | it should | 17:16 |
cosurgi | ok, thanks :) | 17:17 |
cosurgi | KatolaZ: please let me know if you start mergind -debug repos, and what is the deb address for /etc/apt/sources.list | 17:33 |
cosurgi | I add this one from debian for now. | 17:33 |
KatolaZ | cosurgi: we will surely circulate that information widely | 17:33 |
KatolaZ | only, it has not really been a priority so far | 17:34 |
cosurgi | sure, ok :) I just switched to nouveau drivers, and experience crashes sometimes. Guys from #nouveau ask for stack trace :) | 17:34 |
FlibberTGibbet | has anyone encountered 'blank red screen' syndrom on waking a laptop using ascii and i915 drivers? happens with lightdm and either xfce or i3 here. | 17:35 |
FlibberTGibbet | s/syndrom/syndrome | 17:36 |
_ev | I've just begun using Devuan this morning. Flawless thus far. Which manpage do I read for sysvinit? "man init" and "man update-rc.d" ? | 18:56 |
greenjeans | Is there a a beowulf download available or would I need to do ascii and upgrade? | 21:08 |
fsmithred | greenjeans, http://pkgmaster.devuan.org/devuan/dists/unstable/main/installer-amd64/current/images/netboot/mini.iso | 21:09 |
fsmithred | change amd64 to i386 if you want. Be careful, both isos have the same name. | 21:09 |
greenjeans | you da man FSR, thanks! | 21:09 |
fsmithred | I gotta go. See you later. | 21:09 |
specing | fsmithred: are there gpg signatures for those? | 21:53 |
specing | Okay, so I've decided to try out the beowulf installer in a VM and it looks just as horrible as debian's installer looked like 10 years ago when I last tried it | 22:38 |
specing | namely: 1) The minimum iso is 45MB and yet it still has to download packages even before partitioning can begin. Don't you think this .iso is a bit too minimal? | 22:39 |
specing | 2) It takes a damn while for transitions between different screens | 22:39 |
golinux | I think it is done that way so that the latest pks can be included and you won't have to dl twice | 22:39 |
golinux | pkgs | 22:39 |
specing | 3) It won't let me encrypt whole disk? That is, no partition tables at all | 22:43 |
gnarface | i think the beowulf installer you used is just a work-in-progress version of the "businesscard" installer. debian still has these too, last i checked. you want the "netinstall" image, of which there probably just isn't one prepared for beowulf | 22:44 |
specing | 4) It asks for user/root passwords, timezones and whatnot even before touching the disks, prolonging the install process | 22:44 |
gnarface | yea, that's not a bug | 22:44 |
golinux | gnarface: There isn't a netinstall image for beowulf | 22:45 |
golinux | only the mini.iso | 22:45 |
gnarface | golinux: yea, there isn't one yet i meant to say. there still will be one eventually though, won't there? | 22:45 |
golinux | Of course. The mini.iso is for testing | 22:45 |
golinux | some options were added to the installer | 22:46 |
gnarface | just making sure that's clear for specing... i think he's under the impression this is intended to be beowulf's final form | 22:46 |
golinux | and the [kgs are in flux. | 22:46 |
golinux | pkgs | 22:46 |
specing | gnarface: I'm hoping that my listing of issues can help it not being the final form | 22:47 |
golinux | specing: Don't you need root to download pkgs up front? | 22:47 |
golinux | You are critiquing something that is used for internal testing. | 22:48 |
golinux | Which is pretty non-productive. | 22:48 |
specing | golinux: no, why would I? | 22:48 |
golinux | Try the ascii netinstall and see if that more to your linking the changes the soureces. list to beowulf | 22:49 |
specing | does ascii have apparmor enabled and enforcing? | 22:49 |
golinux | Oy lots of typos/. Sorry | 22:49 |
golinux | I don't think so. But it will after an upgrade to beowulf iiuc | 22:50 |
specing | When will the upgrade to beowulf be available? | 22:50 |
golinux | You can do it now but ymmv | 22:51 |
golinux | There are lots of thread son the dev1galaxy.org forum | 22:51 |
specing | 5) THe installer defaults to "erase data" in the disk encryption menus, which is just as slow and much less effective than overwriting with pseudorrandom data | 22:51 |
gnarface | specing: it's available right now, it's just got broken stuff. you were referred to the mini.iso previously because you seemed like you wanted to help with the actually broken stuff. none of what you're complaining about right now is actually broken. | 22:52 |
gnarface | there's things like package dependency conflicts to deal with | 22:53 |
KatolaZ | gnarface: which? | 22:53 |
KatolaZ | (it's a genuine question) | 22:53 |
gnarface | KatolaZ: i don't have any specific examples in mind | 22:53 |
gnarface | KatolaZ: am i already out-of-date on that? are those all fixed now? | 22:53 |
gnarface | i thought there was still an ongoing issue with polkit/elogind | 22:54 |
gnarface | or some stuff related to that | 22:54 |
KatolaZ | dunno gnarface | 22:54 |
KatolaZ | if there is anything broken, just shout | 22:54 |
gnarface | and i assume a myriad of ant-systemd patches to window managers and stuff that got into ascii that still need to be re-applied/merged with beowulf | 22:55 |
KatolaZ | policykit should be fixed now | 22:55 |
gnarface | oh, good to know | 22:55 |
KatolaZ | there is some working going on with slim | 22:55 |
KatolaZ | and there might be a few things to iron out | 22:55 |
KatolaZ | but please if you (or anybody) finds something that does not work, just shout | 22:55 |
KatolaZ | open a bug on bugs.devuan.org | 22:55 |
KatolaZ | we need to know asap | 22:56 |
specing | 6) using the partitioner feels like walking around an undirected cyclic graph | 22:57 |
golinux | specing just posted a list of complaints about the installer itself | 22:57 |
specing | 6+) ..instead of a tree as it should be | 22:57 |
specing | 7) compression and other flags are missing in the mount options for btrfs | 23:00 |
specing | 8) the small menu right after you select use for a block device is very confusing, as it is basically asking you for confirmation of this action. But this is not necessary as no changes have been applied and it can be changed to something else quickly | 23:02 |
KatolaZ | specing: what are you talking about exactly? | 23:05 |
specing | KatolaZ: after you assign a device for LUKS, it basically again asks you what to use it for | 23:07 |
specing | 9) When I create a LUKS device, the created "logical device" pops up above the real device and in the same indentation level instead of being indented a few to the right of the parent device | 23:07 |
specing | 9+ the LVM lvolume in that LUKS device popped up between the virtual LVM device and the physical LUKS host. This is so utterly confusing | 23:08 |
KatolaZ | specing: we haven't touched the partition tool | 23:09 |
KatolaZ | so it comed directly from Debian | 23:09 |
KatolaZ | s/comed/comes | 23:09 |
specing | 10) There is no support for btrfs subvolumes?! | 23:09 |
specing | KatolaZ: I figured. It is exactly the same as I remember it from a decade ago | 23:10 |
KatolaZ | specing: I doubt that, since btrfs was included in stretch/ascii :) | 23:10 |
KatolaZ | so two years ago | 23:10 |
KatolaZ | but I get your point | 23:11 |
specing | the looks/feels of it, I'm sure some new feature makes it in per decade | 23:11 |
watchcat | you can get apparmour in ascii by installing kernel 4.19 from ascii backports. | 23:18 |
gnarface | specing: please, learn to segregate your distaste for the aesthetics of the interface from actual complaints about missing/broken functionality. i'm not even the one working on it and i even find it offensive the way you're doing that. i'm not sure it's intentional though, which is why i'm bringing it up like this. | 23:20 |
specing | watchcat: kernel support is 0.01% of functional apparmor/selinux | 23:21 |
specing | watchcat: 99.99% are good security policies | 23:22 |
watchcat | and...? | 23:23 |
specing | watchcat: well, are there good apparmor security policies in ascii? | 23:24 |
specing | gnarface: I'm just bitter because it has many of the same problems that I remember from a decade ago | 23:25 |
gnarface | specing: everything you've complained about other than it being ugly was a feature that didn't exist a decade ago | 23:25 |
gnarface | so there's that | 23:25 |
gnarface | (ignoring stuff that works you just don't like how) | 23:26 |
gnarface | it's ok to be bitter when it's well-directed | 23:26 |
gnarface | but some of your problems are clearly self-inflicted | 23:27 |
gnarface | (we've all been there before too) | 23:27 |
specing | It proceeded to start installing after partitioning is complete. | 23:27 |
specing | I'm positively surprised that it did not complain about /boot on btrfs-lvm-luks combobreaker | 23:27 |
gnarface | and i used to hate the debian installer's "ugly" ncurses look&feel too. fyi, it grows on you | 23:28 |
specing | The UI's looks are fine, it is the feels that are wrong | 23:28 |
specing | I like nice textual user interfaces | 23:28 |
specing | Is there a way to interrupt the actual package download and install process, so I can do the subvoluming myself? | 23:29 |
gnarface | lots of people run into complications using the auto-partitioning feature with encryption and LVM stuff | 23:29 |
gnarface | just like 10 years ago, you should still be able to get a shell with alt+F4 | 23:30 |
gnarface | do whatever you want to the install | 23:30 |
gnarface | i don't think they even changed the working mount point (/target if i recall?) | 23:30 |
gnarface | you should be able to chroot into it and manipulate the install at that point | 23:31 |
gnarface | i admit it's been almost 10 years since i've had to though | 23:31 |
gnarface | the installer is really in a lot better shape than it was back then. it doesn't look any different but it's better. | 23:32 |
specing | I am in shell, it is just that it will not let me unmount doe to the installer still touching it | 23:32 |
gnarface | hmm. you might have to catch it before or after the actual package download | 23:33 |
gnarface | but from that shell you should be able to interrupt the download by killing it manually, too | 23:33 |
specing | I need to remount to 1) change root filesystem from / to /devuan/root.subvol 2) add compress=zlib,nossd,autodefrag to the mount flags | 23:33 |
specing | gnarface: is it easily restartable? | 23:33 |
gnarface | specing: i assume just go back to alt+f1 and select that menu entry again | 23:33 |
specing | killing programs is easy, picking up the pieces usually less so | 23:34 |
gnarface | it should in theory just error back to the main menu, from where you can re-do any of the selections | 23:34 |
gnarface | *from which | 23:34 |
KatolaZ | specing: just go back to one of the steps to re-do it | 23:34 |
gnarface | i feel like there's gotta be an easier way to do what he's doing. isn't there some optional installer component you can load that does this for you? | 23:36 |
KatolaZ | there must be | 23:36 |
specing | busybox ps not displaying a hierarchical process tree is making it difficult to select proper kill targets | 23:36 |
gnarface | cryptsetup or something like that? isn't that what this is for? | 23:37 |
KatolaZ | specing: are you doing an expert install? | 23:37 |
specing | cryptsetup is the LUKS encryption utility | 23:37 |
specing | KatolaZ: I believe so | 23:37 |
nemo | first official devuan instance at work \o/ | 23:53 |
gnarface | grats | 23:53 |
specing | Is there a document somewhere on how to make a lightweight mirror? Basically a package proxy with caching capabilities | 23:57 |
specing | apt-mirror creates a full one, correct? | 23:58 |
gnarface | specing: there's already apt caching proxies in the repo for you to use. i liked apt-cacher-ng, which i think needed a patch for devuan but i think it's got it by now... | 23:58 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!