systemdlete | OK, next beer on me if you can figure this out... lol | 02:43 |
---|---|---|
systemdlete | I've got a daedalus host (not VM) running xfce desktop with a couple of terminals open and one of the connman frontends open. I switched away to my other PC to do some work then switched back to the PC in question, using a USB switchbox. | 02:45 |
systemdlete | Now the PC won't accept mouse or keyboard input... but it is NOT frozen! | 02:46 |
systemdlete | at least not completely | 02:46 |
onefang | Think it drank too many beers. | 02:46 |
systemdlete | I have the date and time formatted to include seconds, and those seconds are definitely ticking away, so it can't be that the entire system is frozen (though maybe portions are) | 02:46 |
systemdlete | onefang, you could be right. I was unable to watch what it was doing while I was working on the other PC! | 02:47 |
systemdlete | I tried ctl-alt-bs to see if I could force the desktop to crash down to the console, but that fails. | 02:48 |
systemdlete | It seems to be ignoring all input | 02:48 |
systemdlete | I replaced the main board on this PC in the past week. I had to do this because the nvidia card I had for it wasn't working well using the nouveau drivers, and I am trying to avoid non-free software (main exception being vbox, but I really don't have many problems with it) | 02:50 |
systemdlete | so I did not install the proprietary nvidia wares | 02:50 |
systemdlete | Instead, I installed a mb with video onboard. | 02:51 |
systemdlete | I suppose I could browse the bios options and see if I can twiddle with its video a bit | 02:51 |
systemdlete | I'm also wondering if, being it IS a test box for me, maybe it needs a complete reset. But daedalus was just installed there only a short time ago (maybe a couple of months?) | 02:54 |
systemdlete | gnarface (if you are around), you might like this one: The onboard video has 4 options in the bios, consisting of 4 different combinations of GFX0, GPP, IGFX, and PCI | 02:59 |
systemdlete | I'm going to start trying various of these options and see if this lock up problem goes away | 03:00 |
u-amarsh04 | Having two pc's on the LAN and being able to have terminal sessions open between them helps with video lock-up diagnostics | 03:16 |
onefang | May have been the source of the problem, one PC was jealous of the other getting quality time with systemdlete, | 03:24 |
gnarface | systemdlete: that's interesting, never seen those 3 of those 4 options... i think ctrl+alt+backspace is disabled by default in software though now days, you have to re-enable it in a config, did you do that already? | 04:14 |
rrq | systemdlete: when switching keyboard+mouse away from that host, does that host get any hotplug events? likewise when switch back to it, does it get hotplug events? sounds like keyboard+mouse gets switched into new input streams that don't get registered by Xorg | 04:36 |
gnarface | hmm, yea that's come up before with some cheap usb pseudo-win-kvm devices | 04:41 |
gnarface | also a possibility | 04:41 |
systemdlete | rrq, gnarface: Yes, depending on whether I have a desktop up or not, I can see console messages coming from the switchbox. But 99.999999% of the time it works fine. | 06:33 |
systemdlete | onefang: Believe me, in the past 2 weeks, that test box has gotten far more of my time than I would care to give. | 06:34 |
systemdlete | gnarface, and yeah, I think you are right about the ctl-alt-* family of commands having been disabled in recent years. | 06:34 |
systemdlete | I'll look at that, thanks. | 06:35 |
systemdlete | Got another issue, maybe you guys can help me with? | 06:35 |
systemdlete | A couple of years back, I bought one of these: https://www.newegg.com/p/0J2-00G2-00019?Item=9SIACJF6J85056 | 06:35 |
systemdlete | It did not play well with Linux back then, but folks here told me that solid usb 3 support was coming. | 06:36 |
systemdlete | So, today I tried to see if those usb 3 ports are there, but sadly, linux does not seem to recognize them. | 06:36 |
systemdlete | lsusb does not list them. And I don't see anything from lspci either | 06:37 |
systemdlete | I tried modprobe xhci-pci and it did load, but still no sign of those ports. | 06:37 |
systemdlete | (yes, I plugged in both the USB block on the mb and attached power to the sata...) | 06:37 |
systemdlete | I just double-checked my connections by opening up the box and inspecting it. All looks ok, I... THINK | 06:38 |
rrq | and the host port plugged into is usb 3.0 ? | 06:38 |
systemdlete | (yes, I plugged in both the USB block on the mb and attached power to the sata...) | 06:38 |
systemdlete | USB BLOCK= that black 20 pin thingy whatever | 06:39 |
systemdlete | sorry if that was not clear | 06:39 |
systemdlete | apparently, it sips power off the sata chain from the PS | 06:39 |
systemdlete | and the block connects the USB ports to the mainboard's internal idea of USB3 | 06:40 |
gnarface | hmm, USB3 support did show up but i think later than you'd expect... IIRC the stock beowulf kernel still didn't have it | 06:40 |
systemdlete | right, gnarface exactly! | 06:40 |
systemdlete | that was about the time of beowulf. | 06:40 |
gnarface | it seems to work in daedalus though | 06:41 |
systemdlete | I think it might have been you who told me that USB 3 was coming | 06:41 |
systemdlete | gnarface, with the sort of unit I linked to? | 06:41 |
systemdlete | Actually, you know what. | 06:41 |
gnarface | that, i don't know. i only have one USB3 host device here to test with | 06:41 |
systemdlete | This box is still running chimaera, not daedalus | 06:41 |
gnarface | i can' | 06:42 |
systemdlete | there are 2 back panel 3.0 ports | 06:42 |
gnarface | i can't be sure the chimaera stock kernel had it either, you might want to try the chimaera-backports kernel | 06:42 |
systemdlete | and those do not work either. lsusb doesn't mention any usb 3.0 hubs or ports | 06:42 |
gnarface | all i know is the ones on my motherboard here are working in daedalus, and they were working in ceres before that, i never tested chimaera on this one actually though | 06:42 |
systemdlete | gnarface, thanks. I'll look, but you know what, I probably should just upgrade the box to daedalus anyway | 06:42 |
systemdlete | ok, very good then. I want to upgrade my host to daedalus anyway, so I will be oh-so-way-kool like the linux hipster guys | 06:43 |
systemdlete | chimaera is soooo yesterday man | 06:44 |
systemdlete | (never mind me, just a mood, that's all) | 06:44 |
systemdlete | I'm punchy due to several nights lost sleep | 06:44 |
systemdlete | trying to sort out various problems here with various systems | 06:44 |
systemdlete | gnarface, one thing. | 06:45 |
gnarface | if daedalus doesn't make the usb3 ports show up, maybe look into module or kernel parameters | 06:45 |
gnarface | i never had to resort to them for this but there may be some relevant ones to certain hardware | 06:45 |
systemdlete | when I tried the unit originally (maybe on beowulf, idr now), it ended up causing some really erratic behavior in the system. So I disconnected the unit for the longest time. | 06:45 |
systemdlete | I checked to see if I might be missing some sort of usb 3 firmware, but I did not see anything that seemed relevant | 06:46 |
systemdlete | also, your suggestion that I might have a power problem with some of the USB devices could be supported--I am noticing messages from the kernel about power cycling | 06:47 |
systemdlete | I don't know if that is normal, or if that is an actual fault | 06:48 |
systemdlete | this is re: the enumeration issue | 06:49 |
gnarface | i can't be sure, but some of the things you were complaining about a day or two ago looked distinctly like one of those power management bugs i have on another board where the kernel developers and the BIOS developers both point the blame at the other one and the only recourse is to just disable the feature | 06:53 |
gnarface | but... that said, it also looks a lot like something that just appeared in the dmesg of this machine just a moment ago, which i suspect has an actually failing power supply | 06:54 |
gnarface | unfortunately there doesn't seem to be a good way to distinguish between those two problems without waiting for something to actually burn out as evidence | 06:55 |
systemdlete | well, I could replace the PS on this box, as I have an extra new one | 06:56 |
systemdlete | (power management! yeah!) | 06:56 |
gnarface | is it old? i was doing the math on this one and i think it's over 10 years now, so it's due | 06:56 |
systemdlete | eh... let's see | 06:56 |
systemdlete | The last time I rebuilt this box was about (I think) 5 or 6 years ago | 06:57 |
systemdlete | I don't recall replacing the PS since then | 06:57 |
systemdlete | And | 06:57 |
gnarface | also, if it's good but just rated too low for the combination of things that you've plugged into it, that also would be indistinguishable from an actual hardware failure case or a bios bug related to the power management... | 06:57 |
systemdlete | I happen to have in my possession, amongst the many hardware diagnostic tools I have acquired in the past year or so, a PS tester! | 06:57 |
gnarface | (at least just with the crude understanding of the data i have - an actual kernel developer might be able to narrow it down better) | 06:57 |
systemdlete | I've thought of that | 06:57 |
gnarface | yea, try the PS tester if you've got one | 06:58 |
systemdlete | well there are sites (like power utility companies) you can enter info about your PC or PCs and they tell you about how much power they would consume | 06:58 |
systemdlete | not sure how accurate they are though | 06:58 |
gnarface | if this one is really over 10 years old i think that'll be the longest one has ever lasted for me | 06:58 |
systemdlete | some web posts claim those sites are not reliable | 06:58 |
systemdlete | gnarface, maybe the longest I've ever heard of anyone's PS lasting | 06:59 |
gnarface | somewhat ironically, if you leave them on the whole time they last longer | 06:59 |
gnarface | i've got very good data on that | 06:59 |
systemdlete | I state a claim, and I'm sticking with it: EVer since I stopped shutting down my computers at night, my mainboards, power supplies, and disks are lasting incredibly long times | 07:00 |
systemdlete | YES | 07:00 |
systemdlete | exactly what you are saying | 07:00 |
systemdlete | My understanding is that the constant cooldown-warmup cycling puts wear on components. Steady temps, even if rather hot, are actually better for componenets it seems | 07:00 |
systemdlete | (sorry on spelling) | 07:01 |
systemdlete | It makes sense to me | 07:01 |
systemdlete | OTOH, maybe components, even the crap ones coming from a certain nation the west has become dangerously dependent upon, | 07:01 |
systemdlete | are starting to improve in quality | 07:01 |
systemdlete | or maybe both are factors | 07:02 |
systemdlete | either way, I'm happy | 07:02 |
systemdlete | gnarface, Yeah, I see that lsusb on daedalus definitely shows usb 3 hubs | 07:13 |
gnarface | ah, good to know | 07:15 |
gnarface | i forget exactly when usb3 support actually went into the kernel, but i remember being surprised at how late it was, but i had already been running ceres on this one for some time before that, and sid even before that, so the actual transition point is vague to me | 07:16 |
gnarface | (i backtracked to daedalus during the freeze) | 07:16 |
systemdlete | gnarface, I booted into the bios, and sure enough I needed to enable usb 3.0. So I did that and rebooted, but now only lspci sees the hub, but not lsusb | 08:08 |
systemdlete | AND | 08:09 |
systemdlete | I note that none of the mb's I have here support 4 USB3.0 headers, only 2 USB3.0. | 08:09 |
systemdlete | So I am thinking, maybe it makes more sense to get a pci card with some USB3.0 ports | 08:09 |
systemdlete | of course, all of these mb's come with 2 usb 3.0 ports on the back panel. | 08:10 |
systemdlete | I am wondering if having that front header connected might be interfering with lsusb seeing the ports | 08:11 |
systemdlete | we'll only know after I upgrade to deadalus I guess | 08:13 |
yuuri | systemdlete, gnarface: speaking of component longevity correlated to shutdowns, where do sleep states belong in there | 08:15 |
systemdlete | yuuri, that's a really good question! | 08:15 |
systemdlete | Well, I woud think that suspend to ram would impact differently than full sleep (where the machine is actually shut off, fans turn off, etc) | 08:16 |
systemdlete | But I really don't know a whole lot about sleep states--I've only STARTED using these very recently, and I hesitate a lot because I really don't want to cut the power for exactly the reasons we are talking about here. | 08:17 |
systemdlete | I'm not saying it's a bad thing to shut off the machine. | 08:17 |
systemdlete | But if gnarface's and my own experiences tell a story, it seems to work | 08:18 |
systemdlete | yuuri, what has been your own experience with this topic? Have you observed any notable savings in component wear? | 08:19 |
yuuri | not sure, I'm a simple consumer so haven't got to measure such things | 08:19 |
yuuri | and I used to turn off my PC every day when I was using Windows | 08:20 |
systemdlete | so you tend to shut down your equipment when not in use | 08:20 |
yuuri | well, used to | 08:20 |
systemdlete | right | 08:20 |
onefang | I tend to leave my main desktop running all the time. Last time I built anew one the old ten year old one died after I had left it off for a few days coz the new one was working fine. Think it's the PSU. | 08:20 |
systemdlete | could be, onefang | 08:21 |
onefang | It was like "OMG he doesn't love me any more, he loves that new bitch more I'm gonna sulk". | 08:21 |
yuuri | aside from reboots (my desktop is running Void) my uptime is only affected by blackouts/brownouts | 08:22 |
onefang | Exactly. | 08:22 |
yuuri | computers are affectionable | 08:22 |
systemdlete | do you guys plug into a UPS? | 08:23 |
yuuri | don't have one | 08:23 |
systemdlete | They are not terribly expensive for the cheapest ones. | 08:23 |
systemdlete | I don't use them so I can keep my PCs running through a blackout or brownout. | 08:23 |
systemdlete | I use the UPS to keep the power clean--that can damage equipment also when the utility power dips and spikes | 08:24 |
systemdlete | It's worth having for a few dollars | 08:24 |
systemdlete | (or whatever currency you use) | 08:24 |
onefang | I've been thinking of getting one for ages. lol | 08:24 |
yuuri | same | 08:24 |
systemdlete | do | 08:24 |
gnarface | highly recommended | 08:24 |
systemdlete | I think that may also be a contributing factor, now that I think of it. | 08:24 |
onefang | But I wouldn't want a cheap one, and I'm on a pension. | 08:24 |
gnarface | yuuri: dunno really about sleep, i would assume it puts it somewhere inbetween, but it probably also really depends on the duty cycles at that point | 08:25 |
yuuri | now speaking of wear, I saw a post on HN that expressed a wish for manufacturers to limit their lithium batteries at 80% capacity so that they don't get damage from overcharge | 08:25 |
yuuri | increasing the longevity a significant amount | 08:25 |
onefang | I do that to my phone batteries. | 08:25 |
systemdlete | So possible causes of longer component life: (1) leaving equipment on 24x7, (2) better, sturdier hardware coming out of China and elsewhere, and (3) UPS preventing spkes and drops | 08:25 |
onefang | Keep them between 40% and 80%. | 08:25 |
gnarface | i'd heard 25% and 85% | 08:26 |
gnarface | for batteries, operating temps matter though too | 08:26 |
gnarface | some devices just run hotter | 08:26 |
onefang | Likely depends on all sorts of things. | 08:26 |
onefang | And we are well off topic. lol | 08:26 |
yuuri | should I join the offtopic channel? :) | 08:27 |
systemdlete | I don't know. Discussing best practices for running Linux seems related | 08:27 |
systemdlete | but yeah, we probably should move over there | 08:28 |
* systemdlete is in OT now, hoping gnarface and onefang and others will join there to continue this | 08:41 | |
eyalroz | Hello devuaners :-) | 16:09 |
eyalroz | Is someone proficient with py2dsp ? | 16:09 |
gnarface | it's usually a better idea to just ask the actual question | 16:10 |
gnarface | then stay connected in case anyone answers later | 16:10 |
gnarface | i mean, nobody takes bait like that, certainly you know this? | 16:11 |
eyalroz | gnarface: So, py2dsp gives me: | 16:14 |
eyalroz | W: py2dsp cache:67: cannot load cache (_load_package_names:():{}): unpackb() got an | 16:14 |
eyalroz | unexpected keyword argument 'encoding' | 16:14 |
eyalroz | E: py2dsp py2dsp:172: 'build-system' | 16:14 |
eyalroz | ... when I try to build a package (named gdown). I don't understand what these errors are about and what I can do with them. | 16:14 |
gnarface | which devuan release? | 16:15 |
gnarface | eyalroz: ^ | 16:30 |
eyalroz | gnarface: excalibur | 16:30 |
eyalroz | but maybe I should try an apt-get dist-upgrade first | 16:31 |
gnarface | both py2dsp and gdown source here are not in the devuan repos? | 16:32 |
eyalroz | gdown is not in the devuan repos | 16:32 |
eyalroz | ... and it looks like py2dsp isn't, either. I was sure it was | 16:33 |
gnarface | do you think maybe... it just can't read the cache? | 16:35 |
eyalroz | gnarface: I don't know where that cache is; and why won't it be able to read its cache? I'm not stopping it :-( | 16:37 |
eyalroz | Anyway, since it's not a proper Debian package, I guess I'll ask the developer | 16:41 |
gnarface | eyalroz: it's probably just missing some other dependency | 16:42 |
eyalroz | gnarface: ... and I'll gladly oblige it, if it only told me what it was missing | 16:42 |
gnarface | unexpected keyword argument 'encoding' | 16:42 |
gnarface | that might be a clue ^ | 16:42 |
eyalroz | Oh, it looks like it _is_ in a Debian package: pypi2deb | 16:43 |
gnarface | ah | 16:43 |
gnarface | you know that in general, build dependencies are in packages named *-dev, right? | 16:44 |
eyalroz | Ok, someone beat me to it: https://bugs.debian.org/1033841 | 16:44 |
gnarface | ah | 16:44 |
Guest4445 | How move files to subdirectores back based on a duplicate parent folder with same subdirs? | 19:56 |
eyalroz | By the way, if I'm already here... has there been any change recently on the Debian front w.r.t. systemd? i.e. have they done anything to make it possible to choose to not-have it, e.g. during installation? Or are they completely static about their choice of sticking by it? | 19:58 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!