fsmithred | devuoom, there's no 78 in the repo yet. | 00:05 |
---|---|---|
suavedandy | Guys. I have elogind installed but for some reason ACPI events aren't handled. Do I also need acpi-support? | 00:40 |
suavedandy | Also, how do I see init's logs? I noticed something red while SysVinit was rolling. | 00:48 |
devuoom | suavedandy: bootlodg for pre-rsyslogd boot logs | 00:55 |
devuoom | bootlogd | 00:56 |
devuoom | the actual log file show the actual color if you cat or vim it... maybe someone knows how to parse the color in a playback fashion | 00:57 |
suavedandy | bootlogd? Is that a file? Is that a service? | 00:58 |
suavedandy | Sounds like a service. | 00:58 |
suavedandy | bootlogd is not running. | 01:00 |
suavedandy | FAILED! | 01:00 |
suavedandy | MISSION FAILED | 01:00 |
suavedandy | How do I start bootlogd… on startup? | 01:05 |
systemdlete | Interesting. After hours and hours -- no, really, days! -- of fighting with this, I have discovered something. I'm on ascii with kernel 4.9.0-13 and I could not add a partition on a new disk (same size) to an existing array running on two existing disks, BUT | 02:26 |
systemdlete | I can create a NEW array with the same partition on the new disk (forcing a single member) and then add one of the partitions of the existing array (after removing it first) to the new array. | 02:27 |
systemdlete | In the first (more normal I think) scenario, the kern.log gets flooded with all kinds of nasty messages from the disk driver and the system honks every so often. | 02:29 |
systemdlete | In the 2nd scenario, nothing bad happens. In both scenarios, the array gets sync'd though. | 02:29 |
systemdlete | ???? | 02:30 |
systemdlete | btw, I'm using the mdadm tools (not mdraid, just to be clear) | 02:33 |
systemdlete | also, all of my raids are level 1 and are at version 1.2 | 02:33 |
systemdlete | I should also clarify that I can indeed, "add" the partition in both cases, but in the first case, I get all kinds of horrible noise and errors. | 02:34 |
gnarface | so are you suspecting a bug in the harddrive controller? | 02:34 |
gnarface | in the driver for the harddrive controller, i mean? | 02:34 |
systemdlete | why? | 02:35 |
gnarface | i'm asking you if that's what you're saying | 02:35 |
systemdlete | If it were REALLY, TRULY, a hardware controller or disk issue, I'd expect the errors to occur in either scenario. | 02:35 |
systemdlete | right? | 02:35 |
systemdlete | oh, you mean the kernel code | 02:35 |
gnarface | hmmm... they're the same models of disks? | 02:35 |
systemdlete | no | 02:35 |
gnarface | yea, i mean in the kernel code | 02:35 |
systemdlete | I purposely do NOT use the same models | 02:36 |
gnarface | maybe a regression | 02:36 |
systemdlete | Let me give you the layout | 02:36 |
gnarface | or, well, maybe it's been fixed by now. they're well past version 4.9 | 02:36 |
systemdlete | I am feeling great pressure to move on to beowulf. | 02:36 |
systemdlete | the sync (2nd scenario) just finished. no issues. | 02:37 |
systemdlete | but I am going to give the new array some "work" to do. | 02:37 |
systemdlete | Just to see if there might be some actual problems w h/w, etc | 02:38 |
gnarface | good idea | 02:39 |
systemdlete | I just found "iometer" I will try that | 02:39 |
systemdlete | iogen also | 02:39 |
gnarface | also, even if you don't want to go to beowulf yet, you might still be able to try the ascii-backports kernel, which i think is similar in version | 02:40 |
gnarface | it's not a guarantee that there's a problem with the kernel, but it seems highly plausible | 02:40 |
systemdlete | neither iometer nor iogen are available in the repo, but I found fio, which I think does similar | 02:43 |
systemdlete | (I agree. It has to be the kernel, since it is pretty clear it is NOT the new disk, or the old ones for that matter) | 02:44 |
gnarface | well, could be the harddrive controller itself, but usually when they go they don't get subtly flaky like this. that's possible, but rare. usually they either die completely, or at least individual ports die completely while the rest works fine. | 02:46 |
gnarface | kernel regressions on rare old hardware that shares drivers with vast swathes of common new hardware though... that's common | 02:46 |
systemdlete | holy cow. The man page for fio is, like, 1000 pages long. | 02:47 |
systemdlete | I should study and memorize it, and then rent myself out at $50/hour to help people with file and disk i/o problems. | 02:47 |
gnarface | systemdlete: you're just looking for a disk i/o meter? try iotop | 02:47 |
systemdlete | I don't just want to measure gnarface; I want to put a load on the md filesystem | 02:47 |
gnarface | oh, if you're using for an i/o tester though i dunno. i'd maybe try bonnie++ or just hdparm -T/hdparm -t | 02:48 |
systemdlete | i.e., give it some work to do | 02:48 |
gnarface | s/using/looking/ | 02:48 |
gnarface | hdparm -T is more for benchmarking raw harddrive speeds, bonnie++ is for benchmarking filesystems | 02:48 |
gnarface | well, harddrives too, but with filesystem action, as opposed to hdparm | 02:49 |
systemdlete | speed is less the issue. What I want to do, really, (I think) is exercise i/o and hit a few "corner" cases | 02:49 |
gnarface | so, one corner case i've had semi-recently is "also accessing temperature sensors at the same time" | 02:49 |
gnarface | basically i had numerous issues with post 4.x kernels destabilizing on older boards but it seemed related directly to heavy disk i/o at the same time as temperature sensor polling | 02:50 |
systemdlete | how did that one go for you? | 02:50 |
systemdlete | I see! | 02:50 |
gnarface | well i struggled with it for a long time, but after a lot of deductive testing and help from some kernel experts, i disabled the temperature polling in munin, and then disabled "C1E" power saving mode in the bios. rock solid after that. | 02:51 |
gnarface | only burns about 3-5 more watts | 02:51 |
systemdlete | I would guess this is somewhat mainboard-specific? | 02:52 |
gnarface | yes, we ended up concluding it must have been a BIOS-level bug that was never fixed | 02:52 |
systemdlete | how long ago? which hardware era? | 02:53 |
gnarface | but also must never have been noticed at the time, just by luck, the old kernels didn't trip whatever problem it had | 02:53 |
gnarface | pretty old now... one of the earliest Athlon64 boards | 02:53 |
gnarface | and it had an onboard "marvell" brand sata/pata controller that at the time had problems with stability of it's own unless you forced the driver to limit it to PATA mode manually | 02:54 |
gnarface | transitional era hardware with a lot of cursed parts i guess | 02:54 |
gnarface | but maybe similar to what you have actually? | 02:54 |
gnarface | not sure | 02:54 |
systemdlete | yeah, I think it might be. | 02:55 |
systemdlete | I originally had a Thuban x6, but I've since upgraded to a fx-8350 8x | 02:55 |
gnarface | if so, actually it's a good enough reason for you to scour the bios for that "C1E" power saving setting. chances are it's not saving you much | 02:55 |
systemdlete | thuban is athlon I think | 02:55 |
systemdlete | I think thuban was the 6 core versions of athlon II | 02:56 |
systemdlete | iirc | 02:56 |
systemdlete | yeah, I was wondering about C1E actually. | 02:56 |
gnarface | yea, i upgraded that board to a newer 4-core athlon64-x-II thinking maybe the old CPU was subtly but rarely damaged from power surges, but that didn't change shit, so confirmed for being a problem with the motherboard BIOS | 02:56 |
systemdlete | So you think that by disabling it, maybe this issue with mdadm (see above) might dissapear itself? | 02:56 |
systemdlete | which mb, if I may ask? Did you end up replacing it? | 02:57 |
gnarface | yea, i think so. but for good measure i'd also disable temperature sensor polling from your logging system if you are doing that oo | 02:57 |
gnarface | too* | 02:57 |
gnarface | and obviously the cpufrequency governor should be set to something static | 02:57 |
gnarface | don't ask for trouble | 02:57 |
systemdlete | logwatch shows me some temp data. I don't think I'm runnig a temp monitor | 02:57 |
systemdlete | don't overclock, etc? | 02:58 |
gnarface | yea, don't overclock hehe | 02:58 |
systemdlete | lol | 02:58 |
systemdlete | time to order dinner... | 02:59 |
gnarface | board is: ASUS M3N-HT Deluxe/HDMI (the cheaper model, the one one without the add-on heatpipe) | 02:59 |
gnarface | and no, never replaced it | 02:59 |
gnarface | it's there right now | 02:59 |
systemdlete | I'll try bonnie++ and look at disabling the power saver | 02:59 |
gnarface | purely academically, but the soundcard on that board, which i don't use, is also one of those cursed components that shares a driver with too much other hardware (snd_hda_intel) and it's been a pain in the ass for me on literally every motherboard it was pressed into | 03:00 |
gnarface | not in a way that destabilizes the system though, luckily | 03:01 |
gnarface | but, like the Marvell crappy sata/pata controller, it also generally had problems of the form "driver fails to correctly auto-detect the hardware, launches it in a partially functional but not stable mode" | 03:05 |
gnarface | so that may be a common thread here you might want to pull | 03:05 |
gnarface | especially if it turns out that board you're using has a Marvell disk controller | 03:05 |
gnarface | (because flaws in that type of auto-detection have been historically sporadic across kernel versions for me) | 03:06 |
systemdlete | So on my board, the lshw identifies the device as: product: SB7x0/SB8x0/SB9x0 IDE Controller [1002:439C] | 03:35 |
systemdlete | vendor: Advanced Micro Devices, Inc. [AMD/ATI] [1002] | 03:35 |
gnarface | that doesn't tell me anything useful, but google might | 03:36 |
gnarface | looks almost identical to the SATA controller here i'm not having trouble with, but that proves nothing | 03:37 |
gnarface | oh | 03:37 |
gnarface | hmmm | 03:37 |
gnarface | though that does bring up another idea about BIOS settings... | 03:38 |
Xenguy | You crazy people, trying to run linux without the essential systemd | 03:38 |
Xenguy | It boggles my mind | 03:38 |
Xenguy | The imagination knows no terror | 03:38 |
Xenguy | Greater than | 03:38 |
Xenguy | The Mirror | 03:38 |
gnarface | systemdlete: if that controller has a setting that's different for "IDE" mode and "AHCI" mode, the old advice to run it in "IDE" mode by default has been updated for modern kernels to "AHCI" mode. some BIOSes confusingly conflate RAID and AHCI modes, some don't. hard to know without consulting the manual. | 03:39 |
Xenguy | Pardon me for interrupting, wrong channel | 03:39 |
gnarface | systemdlete: 00:11.0 SATA controller: (for reference, not worth much, but this is the one i'm using here: Advanced Micro Devices, Inc. [AMD/ATI] SB7x0/SB8x0/SB9x0 SATA Controller [AHCI mode] (rev 40)) | 03:40 |
systemdlete | Yeah, I forgot about the AHCI/IDE biz. | 03:40 |
gnarface | it can be annoying because not all motherboards handle AHCI and IDE modes equally stably, but at least most the time my experience has been that the kernel version was the biggest deciding factor in which was more stable | 03:41 |
gnarface | AHCI mode might also be faster but usually in my experience not by much on these old boards, because we were already so close to the max throughput on the drives for that bus | 03:42 |
systemdlete | looks like it is setting to AHCI mode. | 03:44 |
gnarface | alright. i wonder why mine shows that in the lspci identity and yours doesn't though | 03:44 |
gnarface | maybe not important | 03:44 |
systemdlete | product: SB7x0/SB8x0/SB9x0 IDE Controller [1002:439C] | 03:46 |
systemdlete | vendor: Advanced Micro Devices, Inc. [AMD/ATI] [1002] | 03:46 |
systemdlete | 00:11.0 SATA controller: Advanced Micro Devices, Inc. [AMD/ATI] SB7x0/SB8x0/SB9x0 SATA Controller [IDE mode] | 03:46 |
systemdlete | strange. | 03:46 |
systemdlete | the kern.log shows it as being AHCI mode | 03:46 |
gnarface | oh, yes, strange indeed. so it does say it, but it's clearly claiming IDE mode. yea that's weird. | 03:46 |
systemdlete | but lspci shows it as AHCI | 03:46 |
gnarface | lspci shows it as IDE you mean | 03:47 |
gnarface | i don't know what's up with that, we're getting beyond my depth here | 03:47 |
systemdlete | kernel: [ 0.432594] pci 0000:00:11.0: set SATA to AHCI mode | 03:47 |
systemdlete | and my depth as well | 03:47 |
gnarface | anything in the log about it setting it back later in the boot up maybe? | 03:47 |
systemdlete | I did not see that, but it is possible I guess. Does it do that? | 03:48 |
gnarface | dunno. it's news to me it can change it to AHCI mode | 03:48 |
systemdlete | even so... with all of this said, gnarface | 03:49 |
systemdlete | it should work either way | 03:49 |
systemdlete | this is an ASUS mb, btw | 03:49 |
systemdlete | and, really, I have had few issues with it. | 03:49 |
gnarface | yea, it should work either way. sometimes it's not fair though. | 03:49 |
gnarface | i like ASUS | 03:49 |
gnarface | it's got the newest BIOS, right? | 03:50 |
systemdlete | the main one seems to be that the USB 3 slots cause some harm to the system | 03:50 |
gnarface | harm? like heating? | 03:50 |
gnarface | or like stability problems? | 03:50 |
systemdlete | (oh yeah. Last update was a few years back, and there havent been new ones since) | 03:50 |
systemdlete | stability of the USB | 03:50 |
gnarface | hmm, that's less weird. | 03:50 |
gnarface | usb3 is still not really done | 03:50 |
systemdlete | I pulled the 2 usb3 devices from the 2 usb3 ports and put them in usb2 ports | 03:51 |
systemdlete | the devices work ok, because neither of them are actually requiring usb3 speeds. One is a wireless device I am using as an WAP and the other is an ethernet dongle | 03:52 |
systemdlete | my comcast here is <100Mbps; I think you need usb3 for 1000Mbps | 03:52 |
systemdlete | although now -- get this -- comcast is forcing me to switch plans. I'll be paying $10 less and I will get up to 200mbps and some tv! | 03:53 |
systemdlete | ???? wtf? | 03:53 |
gnarface | then they sneak into your house at night and steal a kidney | 03:54 |
gnarface | not really a support issue, but i would advise you do anything possible to use any other ISP | 03:54 |
gnarface | i know the savings can be hard to shrug off though | 03:54 |
systemdlete | comcast, charter, century cox, and the rest of the c-words should be dissolved and replaced with net-neutral publicly-owned and operated Internet. But we are getting O.T. here | 03:54 |
gnarface | so, one thing comes to mind about the USB thing but it's vague | 03:54 |
gnarface | there were once two usb drivers, and they supported separate hardware | 03:55 |
systemdlete | Well, AT&T offers actual HSI these days (I think they run fiber in the streets) so it is no longer dialup or DSL | 03:55 |
Xenguy | re: C's: Amen | 03:55 |
systemdlete | heheh, Xenguy | 03:55 |
gnarface | ... and then at some point one of those usb drivers took over for the other i think, in some cases? | 03:55 |
Xenguy | Represent! | 03:55 |
DonkeyHotei | at&t in most places is still VDSL | 03:55 |
gnarface | anyway, it was for speed, but mabye that's to do with your usb3 issue, systemdlete... maybe there's a driver choice you can change back to the legacy one. not sure about that, but worth looking into. | 03:56 |
systemdlete | maybe. When I had the usb3 ports in use, my kern.log was filling up with warnings about something or other. | 03:57 |
gnarface | VSDL is fibre to the phone-pole, though from what i can tell they only let you use it for the download direction | 03:57 |
systemdlete | those are gone now, but I still get some noise from that corner. | 03:57 |
gnarface | systemdlete: something about ohci, ehci... uhci... xhci??? i dunno | 03:57 |
gnarface | systemdlete: i remember reading it as errata at the time but i got lucky and never ran into any troubles due to the change | 03:58 |
systemdlete | reset high-speed USB device number 12 using ehci-pci | 03:58 |
systemdlete | or similar | 03:59 |
DonkeyHotei | ehci = usb2 | 03:59 |
gnarface | systemdlete: yea, that suggests either a driver problem or a physical link problem | 03:59 |
gnarface | systemdlete: i mean, one or two are ok, but if it's spamming them repeatedly it's a problem | 03:59 |
systemdlete | *was* spamming, yes | 04:00 |
gnarface | systemdlete: though i think it could also be an issue with just a particular device doing that | 04:00 |
gnarface | systemdlete: so it might not be the USB controller driver going awry, but just the driver for that one device (had problems like that with some webcams) | 04:00 |
systemdlete | I could try plugging one of the two back in to a usb3 port at a time | 04:00 |
mason | Hrm, without PulseAudio, Firefox can't talk to BlueJeans. Wacky. | 04:01 |
mason | Other than that things seem to be working. | 04:02 |
mason | I might want to switch from the nVidia DKMS from backports to the upstream package, but we'll see. | 04:02 |
mason | But if I don't encounter any showstoppers, this'll be my workstation running Devuan. \o/ | 04:03 |
DonkeyHotei | if by upstream you mean debian, ok. if by upstream you mean nvidia's own package, don't do that | 04:04 |
mason | nVidia's own. Should be relatively safe, but if you've had bad experiences, I'm listening. | 04:04 |
DonkeyHotei | my experiences may be out of date, but they were badly packaged, overwriting GL libraries irretrievably | 04:05 |
mason | There, BlueJeans works with ALSA in Google Chrome and Chromium both. | 04:05 |
mason | Oh, that'd kind of suck. My inspiration here is that the DKMS build failed to do something unspecified with nvidia-persistenced. | 04:06 |
mason | It *works* but something was unhappy. | 04:06 |
mason | Then again, I guess if it breaks I'll just fix it. | 04:07 |
mason | But, alright, I'll stay away from the upstream. | 04:07 |
DonkeyHotei | maybe you could try rebuilding the ubuntu package from the source package? idk | 04:08 |
DonkeyHotei | only other real option | 04:08 |
mason | Anyway, everything else is working. Email, GSSAPI, VPN, A/V stuff. Haven't encountered anything bad yet. I didn't expect to, but I'm still happy. Also, root on ZFS working nicely. | 04:09 |
mason | I'll look at the Debian DKMS and see what it is about nvidia-persistenced. Maybe it insists on systemd or something. | 04:09 |
mason | Oh, I guess it's effectively cosmetic and I was silly for not using --no-install-recommends per http://dev1galaxy.org/viewtopic.php?id=3372&p=4 | 04:11 |
mason | Fixed. | 04:12 |
gnarface | mason: nvidia's shell-script installer runs roughshod over the debian package tree heirarchy, causing permanent unrecoverable corruptions that linger silently until they destabilize your system or derail a future upgrade | 04:14 |
mason | \o/ | 04:14 |
gnarface | mason: it's really good at being the latest driver version, but if you need a newer version you should always try the package in backports first | 04:15 |
mason | I've heard from a number of people who are very happy with it under Fedora, but it's probably assuming Fedora for its layout. | 04:15 |
gnarface | mason: basically if you never go back and you were going to do a full reinstall at the next release anyway, then it's usually fine | 04:16 |
gnarface | mason: if you want to successfully upgrade your same install across 6 releases though, stay far away from it | 04:16 |
mason | Noted, and I'm happy to benefit from the collective suffering of those who have gone before me. :) | 04:16 |
gnarface | mason: as for the bluejeans/pulseaudio thing, i'm guessing bluejeans must have something to do with bluetooth? bluez-alsa was removed from debian so now you need pulseaudio for bluetooth-alsa support, or you find someone's archive of bluez-alsa on github and just build it yourself | 04:17 |
gnarface | *bluetooth-audio support, i mean | 04:18 |
mason | gnarface: BlueJeans is a commercial videoconferencing platform. Work uses it, and with PulseAudio it works on Firefox, but I'd rather use Chromium than have to install PulseAudio. | 04:18 |
gnarface | mason: (in the latter case i think you also need to find a quick&dirty pairing app somewhere) | 04:18 |
mason | I try to avoid BlueTooth like the plague. | 04:18 |
mason | And wireless things generally. | 04:18 |
gnarface | mason: oh, if it's not bluetooth related i dunno | 04:18 |
gnarface | mason: you could try apulse but i haven't had luck recording with apulse | 04:19 |
mason | I'm fine with Chromium honestly. | 04:19 |
Artemis3 | mason, wireless is evil. Imagine, having to keep track even more battery based devices, its a nightmare i hate those things | 04:19 |
mason | Yar. | 04:19 |
debdog | I've been using ATI/AMD's and Nvidia's installers for many years with barely a problem. the trick is to uninstall it before a major distro update. | 04:20 |
gnarface | debdog: that nvidia's installer has a working uninstall feature may be newer than my last experience with it | 04:21 |
* gnarface hasn't used it since before devuan | 04:22 | |
mason | Mm, Steam works fine with ALSA. I'd have lived without it, but I'm happy not to have to. | 04:22 |
gnarface | mason: oh, i actually found a few tricks to work around the one thing steam requires pulseaudio for | 04:22 |
mason | What's that? | 04:23 |
gnarface | mason: you'll only notice it missing for VOIP in certain games, or if you're doing in-home streaming | 04:23 |
mason | Oh, I never use voice in games. But what's your fix? | 04:23 |
gnarface | mason: the VOIP in games can be fixed with a custom ~/.asoundrc, and audio in in-home streaming can be detoured to another device with alsa Loopback device trickery and netcat :) | 04:24 |
gnarface | mason: (probably could be just routed the same way to the steam link itself but only if you mod the linux install on the steamlink) | 04:25 |
gnarface | mason: (i'm just using a nearby raspberry pi connected to the same stereo) | 04:25 |
gnarface | mason: (you'd think delay is an issue but not if you run netcat in udp mode) | 04:25 |
mason | hrm | 04:26 |
gnarface | mason: oh, and there's a couple games that outright crash at launch without pulseaudio, but for those you can just make them fail over to alsa properly by... get this... symlinking libpulse-simple0 (or whatever it's called) to /dev/null from their install directories | 04:27 |
mason | The only multiplayer I do is with people at most a room away, and I never stream, so I randomly avoid all those things. :) | 04:27 |
gnarface | mason: anyway, you're not expected to retain all this info but if you run into trouble ask me again | 04:27 |
mason | That's what logs are for. :) | 04:27 |
mason | I don't have to retain everything, I just have to remember how to use grep. :P | 04:28 |
gnarface | mason: i also have some udev rule fixes for using the steam controller without systemd | 04:28 |
mason | Ah, I don't have one. I'm a dinosaur. | 04:28 |
mason | Dying Light seems to work nicely. | 04:28 |
gnarface | all i heard at launch was how it didn't, but lately all i hear is how fun it is, so they must have fixed the issues | 04:28 |
mason | I don't remember when I started playing, but we love it here. I play it multiplayer with my wife and kids. | 04:29 |
gnarface | sounds awesome | 04:30 |
mason | Recommended. It's beautifully solid. Only issue is that you have to do a dance for multiplayer. | 04:30 |
gnarface | dying light is a fps zombie hunter right? | 04:31 |
mason | With parkour, yes. | 04:31 |
gnarface | i highly recommend a controller with motion-capture aim for such games, of which the steam controller is one example, but i think steam actually supports the nintendo switch ones too now, lol | 04:31 |
gnarface | so, you know, whatever works | 04:31 |
mason | Honestly I just use my trackball. Have for years. | 04:32 |
mason | <- dinosaur | 04:32 |
gnarface | but i never thought i'd find a better controller for FPS games than a keyboard & high-res mouse combo, honestly | 04:32 |
gnarface | i was sure of it | 04:32 |
gnarface | so sure | 04:32 |
gnarface | but my scores tell me i was wrong | 04:32 |
mason | Maybe I'll try one someday. I'm not opposed. I suspect I Just don't game enough to have wanted more. | 04:33 |
mason | hah | 04:33 |
gnarface | dude my scores told OTHER PEOPLE i was wrong, and they notified me of it, before it stopped feeling awkward even | 04:33 |
gnarface | but if you think about it, it makes sense. it changes point+click to something more directly representing actually pointing at the thing you're clicking on | 04:34 |
gnarface | your brain literally has to do less work | 04:34 |
gnarface | so it goes faster whether it feels like it or not | 04:34 |
mason | Alright. I'll see about getting a controller. | 04:34 |
gnarface | if you go for a steam controller, wait for the sale | 04:35 |
mason | I saw one and passed. I think I might be happier with a traditional one. | 04:35 |
gnarface | gamestop usually has used controllers for cheap though if you're not afraid of germs | 04:35 |
mason | heh | 04:35 |
mason | nowadays... | 04:35 |
gnarface | oh | 04:36 |
gnarface | if you get something from amazon though you can return it for no reason | 04:36 |
gnarface | so you could just try it out and you're only down shipping if you don't like it | 04:36 |
mason | We don't use Amazon. Nor Wal-Mart. We try to avoid corps wherever possible. | 04:36 |
gnarface | oh, well that limits things | 04:37 |
mason | Not as much as you might think! Took a while to adjust, but we're long since there. | 04:37 |
gnarface | amazon's new policy reduces a lot of the risk, but i don't deny they're evil and your reasoning is probably sound | 04:37 |
mason | Yar, it's mostly their treatment of their workers, but this might be better in #devuan-offtopic. | 04:38 |
gnarface | yes, agreed, sorry | 04:38 |
mason | Which, you should be in there! Come along! | 04:38 |
Artemis3 | i wish amazon had a policy about inactive accounts | 04:42 |
Artemis3 | i can pray someday they'll just lose the database and the backups | 04:43 |
golinux | Not really worth dropping in here anymore. So sad . . . | 04:48 |
gnarface | golinux: sorry. it only veered off-topic about 20 minutes back | 04:49 |
systemdlete | More joy: "nf_conntrack: default automatic helper assignment has been turned off for security reasons and CT-based firewall rule not found. Use the iptables CT target to attach helpers instead." This seems to be at least 3-4 years old judging by hits by DDG. Harmless or a serious problem? | 05:17 |
systemdlete | I'm running ufw/gufw here. | 05:17 |
systemdlete | again, ascii 4.9.0-13 kernel; same physical host system. | 05:18 |
mason | systemdlete: I wouldn't have thought ASCII would ship nftables, but I might be confused. | 05:18 |
systemdlete | this is after the system being up for a couple of days now. First time I think I have seen it. | 05:19 |
mason | Oh, it's older than I realized. | 05:19 |
systemdlete | mason: Wasn't iptables implemented as a front-end for netfilter so legacy software would run | 05:19 |
systemdlete | ? | 05:19 |
mason | systemdlete: Yeah, but that's newer. Unless, again, I'm confused. | 05:19 |
mason | Yeah, this is what I'm thinking of: https://wiki.debian.org/nftables | 05:20 |
mason | Not default until Buster/Beowulf, but I'm not sure when it was available, so maybe ASCII has it in some form. I don't use ufw, so I can't speak to that - bare iptables here, using the converter you noted. | 05:21 |
mason | Anyway, g'night. | 05:21 |
systemdlete | night and thank you | 05:22 |
systemdlete | gnarface: For bonnie++, how do I specify the target filesystem or disk drive? This is not clear from the man pages. | 05:34 |
systemdlete | I'm guessing it is -d dir; this page indicates so: https://www.jamescoyle.net/how-to/599-benchmark-disk-io-with-dd-and-bonnie | 05:38 |
systemdlete | but even there, he says: "-d – is used to specify the file system directory to use to benchmark." To me, that is not the same as saying "a directory on the filesystem to be benchmarked" | 05:39 |
systemdlete | Ever the word nazi I am, but honestly, I wish devs would learn how to clearly document things. | 05:39 |
user____ | Hi. Yesterday I may have been pranked or pwn3d on beowulf fresh install. I run xfce4 and at shutdown time when I clicked the shutdown button in the ui, after the screen turned black, a UAC dialog appeared which I had not seen before in my life, asking for user password for privileged operation. And it said owned by org.systemd.something. Of course I entered nothing at all there. | 08:01 |
user____ | It is possible that it was an unusual system feature but also that it is real phishing. I am suing the state here for a lot of money and winning. I am a prevalidated target for various "interactions" by select state organizations for that reason. | 08:02 |
user____ | Has anyone got a screenshot link of a systemd related UAC "enter password for privileged operations" screen handy? | 08:02 |
user____ | The system did not shut down as I clicked cancel (was that even safe?), and the confirmation dialog did not appear again on subsequent tries | 08:03 |
user____ | Since I upgraded the ff browser to 78.0.1 yesterday and it self upgraded to 78.0.2 I thought that may have set up something. | 08:04 |
user____ | Phishing wise. | 08:04 |
user____ | maybe-related, also systemd rebels, possibly future devuan recruits... https://medium.com/@gdm85/xfce4-restart-shutdown-without-systemd-polkit-consolekit-you-name-kit-8e52ab608ddf | 08:09 |
* stovepipe chuckles | 08:19 | |
user____ | One has got to admit that seeing a systemd UAC logout / privileged password request dialog on a beowulf system is SLIGHLTY unusual. | 08:24 |
user____ | Another: pressing my power button on beowulf shortly enters suspend instead of initiating shutdown. | 08:30 |
user____ | This is fast and caught me out once already, pulled power with system in suspend without realizing. | 08:30 |
user____ | It does not seem to be editable in the usual location in inittab | 08:30 |
user____ | ca:12345:ctrlaltdel:/sbin/shutdown -t1 -a -r now -- this is the inittab line | 08:31 |
user____ | oh that is not the power button action sorry | 08:31 |
user____ | xfce4 power manager in menu is set for menu display "ask" on power button press. That does not happen. | 08:32 |
user____ | sleep and hibernate buttons are not present on this machine and both are set to nothing and grayed out in power manager. | 08:33 |
user____ | stovepipe: I am not in usa as you can see. Why chuckling? The new recruit? | 08:33 |
user____ | This is the screen I got yesterday ONCE. Why? Was it phishing? https://unix.stackexchange.com/questions/545328/sudo-equivalent-in-systemd | 08:39 |
user____ | I see my xfce4 session starts the policykit authentication agent. Is that needed in an xfce4 session at all? | 08:43 |
user____ | another: https://unix.stackexchange.com/questions/224487/uninstall-geoclue-from-debian | 08:46 |
user____ | geoclue demo is running in xfce4 session, just stopped. This is a fresh beowulf install. | 08:46 |
user____ | I am still miffed that the systemd uac screen appeared on beowulf yesterday, once, on shutdown attempt. What the h* | 08:49 |
user____ | Bookmarked that last link, incredible. You need to tell the command to not ask for a root password while running as root? It's like asking your left hand for permission to use the left hand unless you tell it not to ask for permission. These people are way off the rails, systemd people. | 08:58 |
user____ | I've made a lot of noise in the channel, will shut up now. | 09:01 |
r3boot | LOL, I've seen a lot of crap wrt systemd, but a phishing attack by some state actor is next level, nice one user____ :) | 09:12 |
user____ | Well it is not so out of the ordinary here. The local debian mirrors are known to be under "unreliable" control for one thing. | 09:15 |
user____ | Check by /whois where I am | 09:15 |
user____ | Acquiring root and account passwords is part of the "welcome package" these people extend to anyone. It may be automatic. | 09:16 |
onefang | What did I just miss? | 09:16 |
user____ | yesterday I also had yahoo email login problems. I suspect the same reason. | 09:16 |
user____ | onefang: my oem question was how come I saw a systemd privileged operation password required dialog on beowulf yesterday, exactly once. | 09:17 |
user____ | at shutdown time. | 09:17 |
user____ | the rest was speculation as to why or how. | 09:18 |
onefang | The bit about debian mirrors is what caught my interest, since I'm one of the Devuan mirror herders. We mostly use Debian package mirrors for those packages we don't fork. | 09:18 |
user____ | Hm. | 09:19 |
r3boot | Actually, you saw a reference to a namespace inside of dbus, which just happened to have the org.systemd prefix, nothing special | 09:19 |
user____ | The gui screen is a reference to what? | 09:19 |
user____ | r3boot: I saw the GUI password entry for privileged op dialog. | 09:19 |
r3boot | https://techbase.kde.org/Development/Tutorials/D-Bus/Introduction#Namespaces_and_Addresses | 09:20 |
r3boot | user____: yes, thats normal, because you need elevated privileges to perform a shutdown. Hence, something/someone will need to ask you for your password (unless you have some method of circumventing it). That is what the popup is for | 09:22 |
user____ | that is brilliant, r3boot. But the system has no systemd on it that I know of. | 09:22 |
r3boot | this has nothing to do with systemd | 09:22 |
user____ | And no you do not need "elevated privileges" for shutdown because that is sorted by xfce4 setup permissions which allow this, allowed it before yesterday, allowed it before that, allow it now. Just not once yesterday. | 09:23 |
user____ | your link is related to kde. xfce4 is not related to kde. | 09:24 |
r3boot | right | 09:24 |
onefang | Systemd itself isn't on Devuan, coz we try to strip it out. However, there are a few compatibility things we needed to add to work around various systemd infections. So there's the odd referenece to systemd still there. | 09:24 |
r3boot | okay, back to useful things :) | 09:24 |
user____ | onefang: just for laughs, trying to conjure the same dialog using a cli command to see if it is "present" (and dangerous) | 09:24 |
onefang | I just use the shutdown command on those rare times I need to. | 09:26 |
user____ | Again: the xfce4 shutdown button works fine... it did not work ONCE, apparently. And I want to know why and how. | 09:26 |
user____ | policykit pkttyagent and pkexec exist. Maybe these can do this crap? | 09:27 |
user____ | so confirmed: on beowulf with policykit installed, the following command conjures a scary systemd (see Details:) apparently-originated UAC password entry screen: pkexec /bin/bash -c date | 09:29 |
user____ | I suppose one of the packages I upgraded yesterday killed pkttyagent and then the gui password dialog appeared. With pkttyagent running it does not appear as expected. | 09:29 |
user____ | correction: details does not show systemd owned. it shows freedeskop/policykit owned. yesterday's was systemd owned. | 09:30 |
user____ | I now think it was real phishing. | 09:31 |
user____ | Has anyone ever seen such a dialog on beowulf? At any time? | 09:31 |
user____ | I now consider the machine to have been kitted. | 09:31 |
user____ | Bummer. | 09:31 |
user____ | Suggest a normal command on beowulf capable to generate the above dialog with Detail: including org.systemd.* | 09:35 |
user____ | Channel: ^ | 09:35 |
user____ | ? | 09:46 |
r3boot | just do a reinstall with something you do trust. IF the box is kitted, you can no longer trust it | 09:51 |
r3boot | or even better, get a new machine, since this one likely has implants | 09:52 |
r3boot | (that, or submit a bugreport and/or a patch which fixes it) | 09:54 |
stovepipe | chinese virus | 10:20 |
gnarface | user____: it wouldn't surprise me if that were a normal dbus/gconf namespace thingy | 10:24 |
gnarface | user____: inconclusive | 10:24 |
gnarface | user____: the thing about the power button not working right might just be a issue of package alternates... | 10:25 |
gnarface | user____: are you using lightdm? | 10:25 |
user____ | xfce4 | 10:28 |
user____ | gnarface: okay, provide a cli command which produces that password entry screen with Detail including org.systemd.* | 10:28 |
gnarface | user____: i'm not sure if that answered my question. the graphical login manager is lightdm? | 10:29 |
user____ | slim | 10:29 |
gnarface | oh | 10:29 |
gnarface | hmm, well a similar thing happens with lightdm but it goes away if you install lightdm-gtk-greeter instead of whatever gets installed by default | 10:30 |
gnarface | the thing about the power button not working i mean | 10:30 |
user____ | but does the DETAIL field show org.systemd.* or not? | 10:30 |
gnarface | not sure about the org.systemd.* thing | 10:30 |
user____ | Oh the power button. | 10:30 |
user____ | I tinkered with it a little we'll see later if it behaves now. | 10:31 |
r3boot | it would help if you tell us what '*' is, since then we can identify the exact thing on dbus which is throwing this | 10:31 |
user____ | I don't remember. I was shutting down and on the black screen with all ui elements already gone this popped up. I was in a hurry and incredulous so I looked at Detail and then clicked Cancel which got me back to the logon screen, logged in, then again shutdown, then it worked without nag screen | 10:32 |
user____ | Should have taken a pic with the phone. Screenshot in the usual way was not possible by then, all ui was gone. | 10:32 |
r3boot | Also, slim does not do any form of polkit/consolekit out of the box, you have to configure that manually. | 10:32 |
user____ | I am aware. Thus the question. Who or what caused that crap?! Once! | 10:33 |
r3boot | give us the full path to the namespace, and we can tell you | 10:34 |
user____ | if it happens again yes. | 10:34 |
user____ | *when | 10:34 |
user____ | I installed a lot of packages yesterday, something in one of them may have lingered. Or not. | 10:35 |
r3boot | could be. We are deep in crystal ball territory now :) | 10:35 |
user____ | Does one want the policykit auth agent running in an xfce4 session? | 10:35 |
user____ | r3boot: in my version of the story the crystal ball is Sauron's eye watching me. | 10:35 |
onefang | lol | 10:36 |
* user____ moons the crystal ball and heads for the kitchen's pots and pans - noon here soon | 10:36 | |
r3boot | user____: you need /some/ framework which allows you to elevate your privileges to shutdown. This can be su, sudo, gtksudo, polkit, consolekit, or some other contraption. | 10:36 |
r3boot | w | 10:36 |
user____ | I'll try shutdown with polkitd stopped. But I want to see org.systemd.* in Detail. Not *polkit*... | 10:37 |
r3boot | we only care about '*' | 10:38 |
stovepipe | "I am a prevalidated target for various "interactions" by select state organizations" | 10:38 |
stovepipe | i would like to know more about this | 10:38 |
r3boot | mja, maybe take that to #devuan-offtopic | 10:39 |
stovepipe | RIP IRC | 10:40 |
user____ | ripoff irc "freedom" | 10:54 |
user____ | stovepipe: google "state surveillance romania" -- done with it here. | 10:55 |
user____ | specifically: (all the laws mentioned therein are valid since then and there are a few worse things meanwhile, like blocking internet sites at national router/isp level with no recourse and so on) https://giswatch.org/en/country-report/communications-surveillance/romania | 10:57 |
* user____ is now really done | 10:57 | |
systemdlete | good news and bad news. Those "hardware" errors (or whatever they SEEMed to be) coming from the hard drives are now all gone. Drives are all fine, old and new ones. In fact, many of my older drives I thought had died (well, they SHOULD be dead by now) probably work. | 11:23 |
Artemis3 | i wonder what chipset is this | 11:23 |
Artemis3 | hope not p67h67, nah too old :) | 11:24 |
systemdlete | The problem was the Kingwin disk bay unit. I noticed that the individual power LED for some of the drives were no longer lighting. So, on a hunch, I took the drives out and hooked them up directly to power and sata connectors on the mainboard. | 11:24 |
Artemis3 | well done | 11:25 |
user____ | Kingwin must be a pun on Kingston :) The Lolex of Rolex! | 11:25 |
systemdlete | I had aproblem with another Kingwin unit a few years back, but they repaired it for nothing and it kept working. But it was noisy, due to the fact they had not laid out the backplane so you could put in a large fan; it had 2 40mm fans I think. This one has one large 120mm fan, which is quieter, but honestly, with the unit out completely, it has never been quieter in my apartment. | 11:26 |
stovepipe | user____: so romania is evil surveillance now? are they watching you on IRC? | 11:26 |
systemdlete | yeah, I've been pulling my hair out for months now, first with one drive that went bad (for real) but then another drive, different brand and model, also seemed to fail only on one PC but not the other (which has no Kingwin or other intervening issue). | 11:27 |
Artemis3 | big brother always watches you, everywhere. | 11:27 |
stovepipe | if state organizations are watching you, they certainly would be here | 11:27 |
user____ | unlikely but there are rumors of "blanket deployed" root kits / phishing kits especially targeting new installations of "dangerous" systems (i.e. linux). | 11:27 |
Artemis3 | they don't need to, but they sometimes do | 11:27 |
user____ | I am curious if the stupid error can be reproduced. I will try hard and report. I *hope* I am paranoid. But I have reasons... | 11:28 |
stovepipe | dangerous systems are status quo and without eyes | 11:28 |
systemdlete | I'm wondering if revisiting the older, 3-bay unit is worthwhile. I've looked at others, like Rosewill, but I'm not impressed with their units. | 11:28 |
Artemis3 | such an evil move that state malware, not that im surprised or anything | 11:28 |
stovepipe | all eyes are on devuan | 11:28 |
systemdlete | I like having the separate power and activity lights on each bay so I can tell if there is a problem. | 11:29 |
user____ | all eyes are on people installing vpn's which I did yesterday. | 11:29 |
Artemis3 | what their malware exploits systemd? | 11:29 |
systemdlete | OTOH, I've set up a watchdog on my kern.log now and it sends me telegrams to my phone if there is trouble. | 11:29 |
systemdlete | so maybe all the blinky-blinkies are not so necessary now. | 11:29 |
Artemis3 | who needs bays when you can always screw/unscrew the things all the time :) | 11:30 |
systemdlete | To #debianfork with that. Otherwise, "mom" will come out here and yell at you. | 11:30 |
systemdlete | Artemis3: Yeah, but that's not the only thing: The one big fan helps keep the disks cool. | 11:30 |
Artemis3 | its its just that, put the fan in the case? | 11:31 |
Artemis3 | if | 11:31 |
user____ | Invest in an even bigger fan, run it as slow as possible. Should be hush level. 9dB? | 11:31 |
user____ | Yes inside the case. | 11:31 |
Artemis3 | 140mm | 11:31 |
user____ | I think that is big enough :) | 11:31 |
systemdlete | as long as there is adequate air flow, ok | 11:31 |
Artemis3 | what was the silend brand, artic? forgot | 11:31 |
systemdlete | I've already got about 4 fans in the case now. | 11:32 |
user____ | If you really worry about noise get a conduction cooled case from Lian Li etc. This is really ##hardware talk now, should maybe not be here on #devuan. | 11:32 |
systemdlete | Some of the new cases I've been looking at have sideways slots for all the disk drives. | 11:32 |
systemdlete | well, this began due to a hardware issue. We were uncertain if it was the kernel version or some flakey mb chips, or what | 11:33 |
Artemis3 | yeah i have seen those, so easy to handle | 11:33 |
Artemis3 | but problem solved, well done | 11:33 |
systemdlete | you can do everything without any tools but your hands and fingers. These new cases are so f-ing cool I want one just because they are so f-ing cool. | 11:34 |
Artemis3 | rgb leds and stuff? | 11:34 |
systemdlete | some of them, yeah. | 11:34 |
systemdlete | if that's your thing. | 11:34 |
systemdlete | they have cases with the usb and jacks on the top, so they can help collect dust more efficiently. | 11:35 |
user____ | Don't need tools for cases since 2005 or so. | 11:35 |
user____ | systemdlete: and coffee | 11:35 |
systemdlete | you haven't seen these. | 11:35 |
user____ | But conduction cooling is the best. No need for fans. | 11:35 |
user____ | systemdlete: I have seen a lot of things, even poured my share of liquids into some by mistake over the years. | 11:35 |
systemdlete | I mean, the guys who review the cases (unboxing vids) take the whole case apart and then reconfigure it completely right in front of your eyes. THATS COOL | 11:36 |
systemdlete | Artemis3: Thank you for the congrats. I've been thinking this for some time now, but I love the ability to pop drives in and out (even if you have to shut down the machine first to do it). It's so convenient, and the cooling like I said. | 11:37 |
Artemis3 | well if your motherboard has the internal esata ports you don't even need to turn it off (unless it the /) | 11:38 |
systemdlete | It's spooky quiet here now. I miss my Kingwin... | 11:38 |
user____ | lol | 11:38 |
systemdlete | not esata; just plain sata I think. | 11:39 |
user____ | https://www.google.com/search?q=conduction+cooled+pc+case | 11:39 |
systemdlete | that better not be a link to some site in romania... :p | 11:40 |
* user____ buys systemdlete 10lbs of carrots, for the eyes. google links. | 11:40 | |
systemdlete | not gogle.com or gooogle.com ? | 11:40 |
Artemis3 | haha they can probably mess people's browsing at the isp or country level | 11:40 |
gnarface | google does | 11:41 |
Artemis3 | i know my country has done it too... | 11:41 |
user____ | they are interested in hoovering up information that can be used against people later, not in immediate attacks. | 11:41 |
gnarface | they also serve different content for different languages in the same region, too | 11:41 |
gnarface | but this is definitely offtopic | 11:41 |
Artemis3 | which is why now i use dnscrypt-proxy, no other dns cuts it | 11:41 |
user____ | indeed. Clean out your cookies frequently. Google cookies. | 11:41 |
systemdlete | Or how about a <h link="www.someplaceinromania.ro">www.google.com</h> | 11:41 |
systemdlete | Ever fell for something like that? | 11:42 |
user____ | have a lollipop and go back to kindergarten | 11:42 |
user____ | this is irc, antisocial media of ascii emoticons. | 11:42 |
Artemis3 | it also caches dns so is great for your lan | 11:42 |
systemdlete | who you talking to user____? | 11:42 |
user____ | just to myself. | 11:42 |
systemdlete | oh | 11:42 |
systemdlete | you mean, to the hackers? | 11:43 |
Artemis3 | leaving words for the logging | 11:43 |
systemdlete | ah. | 11:43 |
systemdlete | Artemis3: What about dnsmasq? | 11:43 |
systemdlete | it caches, and it has dhcp server built in. | 11:43 |
Artemis3 | thats just cache, this one does the secure encrypted dns thing | 11:44 |
systemdlete | dnssec? | 11:44 |
Artemis3 | yeah that and the other two, check it out | 11:44 |
systemdlete | ok, I will. thanks for the tip | 11:44 |
Artemis3 | should be in the repo | 11:45 |
Artemis3 | "A flexible DNS proxy, with support for modern encrypted DNS protocols such as DNSCrypt v2, DNS-over-HTTPS and Anonymized DNSCrypt." | 11:46 |
xinomilo | unbound can do caching/DoT/DoH too. or use dnscrypt-proxy.. | 11:46 |
systemdlete | so gnarface: Have you caught up in the channel here? See what I did to solve my drive problems? | 11:47 |
Artemis3 | Here is project page: https://github.com/DNSCrypt/dnscrypt-proxy/ | 11:47 |
gnarface | it was just hardware failure right? | 11:47 |
gnarface | the drive bays? | 11:48 |
gnarface | or rather, the drive bay controller failed? | 11:48 |
Artemis3 | i'd suspect either cables or contacts | 11:48 |
systemdlete | I think the drive bays don't really have a controller per se | 11:48 |
gnarface | they might have some sort of processing unit, i wouldn't be sure the fancier ones don't | 11:48 |
systemdlete | they have a small backplane for each of the 4 drives with a sata power and sata data connector. | 11:49 |
gnarface | it might not be all resistors and wires | 11:49 |
systemdlete | so you push the drive in, it catches the connectors in the back and then you have the cables on the outside to the board. | 11:49 |
Artemis3 | a fancier one would probably let you do things like bod or raid | 11:49 |
systemdlete | There's 2 molex plugs (why not have sata plugs, but oh well) and the 4 sata connectors. | 11:50 |
systemdlete | I was doing RAID with it, actually. But I was using mdadm for RAID | 11:50 |
Wonka | wireshark-commons has grown a dependency on libsystemd0 :/ | 11:50 |
Artemis3 | lol | 11:50 |
Wonka | wireshark-common, of course | 11:50 |
gnarface | well if it's just contact corrosion, that should be fixable with some polish | 11:50 |
systemdlete | wireshark-common-systemd? | 11:51 |
systemdlete | I don't know gnarface. The LED lights were wonky also... | 11:51 |
systemdlete | and the strangest thing is that the 2nd bay was working as far as the data to and from the hard drive, no errors, but the power LED was not working, though the activity LED was working on all 4 bays. | 11:52 |
user____ | Wonka: wow. What for? | 11:52 |
user____ | network manager related tentacle? | 11:52 |
Wonka | user____: don't know, I only tried to "apt upgrade" | 11:52 |
systemdlete | That was the point I said NUF. Let's see what happens if we completely eliminate the Kingwin unit altogether. | 11:52 |
Artemis3 | when you need your net sniffer to be launched from init service because reasons? | 11:53 |
systemdlete | Wonka: You are sure you are not hitting a romanian repo? | 11:53 |
user____ | haha | 11:53 |
user____ | a nice ventilated case :) if it gets too hot put it in the *bathtub* https://www.overclockers.co.uk/lian-li-pc-y6a-odyssey-yacht-mini-itx-case-silver-ca-74i-ll.html | 11:54 |
Wonka | systemdlete: quite sure it's deb.devuan.org - no idea if there's romanian servers in the rotation ;) | 11:54 |
user____ | I think the romanian debian archives I referred to were cleaned meanwhile. | 11:54 |
systemdlete | romanian? Maybe romulan? | 11:54 |
Artemis3 | packages are signed and checked by apt anyway | 11:54 |
systemdlete | I know, was just... having some dumb fun with the topic... | 11:55 |
Artemis3 | yeah but its a valid concern | 11:55 |
systemdlete | that's too bad. That it is a valid concern. | 11:55 |
onefang | No known Romanian servers in the deb.devuan.org round robin. I am talking to some Romulans though. | 11:55 |
systemdlete | It should not be. | 11:55 |
user____ | not in devuan round robin in debian round robin | 11:56 |
Artemis3 | thats remans not romans (those are from a bit earlier age) | 11:56 |
Artemis3 | actually that country did took the name from ancient rome, but whatever | 11:57 |
systemdlete | onefang: Tell the Romulans they can have their leader, Trump, back. | 11:58 |
systemdlete | We'll throw in Biden too. | 11:58 |
systemdlete | Anyway, too many hours at this. Time to hit the hay | 11:59 |
Artemis3 | user____, nice case lol | 12:01 |
Artemis3 | ps: youtuber tried it in a pond, doesn't float xD | 12:16 |
user____ | lolsubmarine yacht. Perfect for diving in a cooling isolating oil filled pond. | 12:20 |
user____ | what is the "planet number" in devuan releases? | 12:22 |
user____ | Ah ok I got it | 12:23 |
user____ | Artemis3: you will like this | 12:26 |
user____ | https://www.pugetsystems.com/mineral-oil-pc.php?__cf_chl_captcha_tk__=d53d317b6eeb73a205bcc7edd6bf8208d7261bcf-1598437547-0-AepUVggQJgZbRVUeByZZ0sOuWXTWQPXuPd1TkLSFu1nchQILJjKDeUG_3nsqpYNcBghhy5IkXQufVwOoXcrXJx9lEAlAPlhSqKxJH1wcSz4ns1pOjE3l7qxlk3zsKI52-5BrYokO6p_xb3qnNzDdNOgO0Ei78pMo4mu1LOxQxt51EapA4jmiI8m-J5huh8Qq4IZOEXVfKZk2R6yq41TxfhTEkgT3Dxg2cRl4ie0btUJ-Ji2SZK-aKAJ9k1y098nADjn2X1SbCTeAYOoHCCe4wTc0moQK | 12:26 |
user____ | upKEpRGpLeK5TzeB_l0fK6WpHeuYXohA6ccXpJarsVokPvWQvFWfjtm4ZfPiCXB2toBPvw_jm0KMXsH55V7c_7ejzKvUYSblcbZP9tP8NBwweXQQvASflRFd4aJVdFS4wN9otZLtuixaoDOK2wADIiI85bKc6ivHpElWNOR_QmaOZFe5F9qfrV_JyALrMdXMm1Gsv_AEUxUibWBbcUZTNuZRoq5JLi0QidFU7g | 12:26 |
user____ | aargh | 12:26 |
user____ | https://www.pugetsystems.com just this | 12:26 |
r3boot | ghe .. reminds me of the cray2 https://en.wikipedia.org/wiki/Cray-2 | 12:37 |
user____ | Plumbing one's "data center" into the hot water house heating system is a valid move. No fans needed then. | 12:50 |
user____ | rebooted with all polkitd and agent things killed, no root password request screen seen | 12:57 |
user____ | will be back later maybe someone knows something I should check in continuation? | 12:57 |
user____ | xfce4 wm re-arranges desktop icons on reboot. I want them to stay where I put them. This is not good. Any way to avoid the auto arrange? | 12:58 |
user____ | systemd password request mystery cleared #1 | 13:52 |
user____ | problem: systemd contamination of /etc/acpi/powerbtn-acpi-support.sh root.root 755 | 13:52 |
user____ | edit file to be just: '#!/bin/bash\nlogger -p syslog.crit "shutdown by power button press commanded"\nsync\n/sbin/shutdown -h -P now "Power button pressed"' | 13:53 |
r3boot | nice, now try to execute that as a normal user and see what happens :) | 13:54 |
user____ | The immense wisdom of systemd prozelytes determined that the power button should be treated like the logout software button on the ui. However, in real life a) the power button may be in another country (telecommuting) or b) the power button needs to be used because the ui is frozen | 13:54 |
user____ | r3boot: I did. Go away. It works fine. The power button pressing person is not a "normal user". | 13:54 |
user____ | I am so sick and fed up cleaning up after baby drool (systemd) all over things which worked for 20 years in my hands and for 40 since inception. | 13:55 |
user____ | /rant | 13:55 |
user____ | r3boot: for normal users, one creates a group called perhaps r3booters and puts the users allowed to run the command in the group, then permits sudo that command with no password for the new group. | 13:56 |
user____ | similarly one can use an automounter script or use a group of mount enabled users instead of idiotic uac on a single user machine | 14:05 |
onefang | The Romulans will be most upset. | 14:13 |
user____ | Need to send some klingons to mop up that sector soon. | 14:18 |
user____ | now thunar wnats me to id whenever I try to mount one of the local drives. drive management by thunar is on. | 14:18 |
user____ | I need to find another script to neuter. Later. | 14:19 |
* user____ looks into editing polkit policies a "little" with a chainsaw | 14:19 | |
Wonka | I was not entirely correct on wireshark-common having grown a dependency on libsystemd0. That was there before and could be satisfied by libelogind0. The version in testing now depends on libsystemd >= 246, which cannot be satisfied by libelogind=243.8-1, which is the newest one in testing... | 17:09 |
mason | Wonka: A useful metric to consider - if a tool can build on FreeBSD or Slackware, a systemd dependency is just in the packaging, not the code. | 17:12 |
Wonka | yep... well, I only need a new enough libelogind0 in testing. | 17:13 |
mason | Better still would be it not requiring the renamed systemd-logind library, but yes, that's correct. | 17:13 |
Wonka | (it depends on libsystemd0 >= 246, I missed the 0 above) | 17:14 |
Wonka | would we be able to correct the packaging so it loses that dependency= | 17:15 |
Wonka | ? | 17:15 |
mason | Wonka: Maybe, but it'd be necessary to see why it was linked and if any new functionality is actually used. | 17:17 |
mason | Change logs might say. | 17:17 |
user_____ | did you see my solution for power button not shutting down system due to systemd contamination from a few hours ago? | 22:24 |
user_____ | Here is the problem and the solution again: | 22:24 |
user_____ | on power button press systemd and polkit crap "checks" that no-one is logged on and if logged on offers a gui password entry to permit the shutdown. this is a serious idiocy, the hw can be in another room or country, and the power button has priority over any logged on users for sure, especially if the gui is stuck. | 22:25 |
user_____ | solution: | 22:25 |
user_____ | #!/bin/sh\n# this is /etc/acpi/powerbtn-acpi-support.sh root.root 755\nlogger -p syslog.crit "shutdown by power button press commanded"\nsync\n/sbin/shutdown -h -P now "Power button pressed" | 22:26 |
user_____ | the previous systemd and polkit contaminated emo script was renamed to $_.bak | 22:26 |
user_____ | I wonder if brain enhancing vitamins are very expensive sometimes, where systemd developers live and work. How can one disable/override a hw shutdown/reset button with a gui?! | 22:27 |
user_____ | When the previous script was in, oem in beowulf, I had to hard reset the pc when the ui got stuck with trashing disks (firefox). Risking to lose data due to unclean disk umounts. Of course pressing the power button shortly did nothing at all, since it tried to put something up in the already unresponsive gui. Sick. | 22:29 |
specing | you can blame systemd for lots of things, but you can't blame it on Linux's OOM handling being piss poor | 22:38 |
Xelraa | user what does devuan has to do with god damn systemd... | 22:38 |
specing | s/it on/it for/ | 22:39 |
luser978 | oom? | 22:40 |
* luser978 is user_+ | 22:40 | |
specing | out of memory | 22:42 |
luser978 | polkit .policy files changed while i installed packages. default policies in the live system worked fine. | 22:42 |
specing | one non-root process eating all memory shouldn't make the machine freeze | 22:42 |
specing | and yet it does | 22:42 |
luser978 | specing: is the hdd light on | 22:43 |
luser978 | you can set ulimits to prevent greed | 22:43 |
luser978 | i do that and nice +5 on firefox and it is tame. | 22:44 |
Owner | if machine freezes then it could be the hdd failing as it writes to swap | 23:02 |
Owner | check dmesg for ugly messages | 23:02 |
tom_ | hello | 23:03 |
tom_ | Having some issues with imagemagick on beowulf | 23:04 |
tom_ | not sure if i'm doing something wrong | 23:04 |
tom_ | or if it's a bug | 23:04 |
tom_ | I type convert -size 290x70 xc:skyblue -bordercolor black -border 5 -fill black -stroke black -strokewidth 1 -font Anonymous-Pro -pointsize 30 -draw "translate -115,-3 skewX 54 gravity center text 0,0 'x'" test.png | 23:04 |
tom_ | this draws an x in the middle of the image. | 23:04 |
tom_ | I expect it to draw it on the left, or at the -115,-3 offset | 23:04 |
tom_ | the X coordinate appears broken when handling negative numbers | 23:05 |
luser977 | put -size after the 1st filename | 23:06 |
luser977 | oh you are generating | 23:09 |
luser977 | there's an imagemagick channel on freenode i think. | 23:10 |
luser977 | yes ##imagemagick tom_ | 23:11 |
tom_ | I don't think it's active | 23:12 |
luser977 | just been there, lots of nicks. ask and wait imo | 23:12 |
luser977 | wow we have bats buzzing the window. scary. | 23:32 |
* luser977 goes to zzz | 23:32 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!