leafwiz | fsmithred, yeah, I changed it to #!/usr/bin/env python2, but I don't think I got all of the scripts since wicd is still a bit wonky. | 00:51 |
---|---|---|
leafwiz | Is there plans for devuan to use another networking daemon? for me its seems it default when I choose the mate environment. | 00:51 |
leafwiz | Or is there another one that you would recommend? | 00:53 |
leafwiz | for my devuan system running mate on beowulf | 00:53 |
fsmithred | leafwiz, I've been using network-manager. I don't like it, but it seems to be working. I've also tried connman. Some people like that one. | 01:01 |
fsmithred | both of those will work with vpn, which wicd does not. | 01:02 |
leafwiz | So wicd is kinda the one we kinda want to use? | 01:02 |
fsmithred | but I miss wicd | 01:02 |
leafwiz | But it should be updated to p3_ | 01:02 |
leafwiz | ? | 01:02 |
leafwiz | aka python3 | 01:02 |
fsmithred | it's the one we have been using | 01:02 |
leafwiz | mm | 01:02 |
fsmithred | some like n-m better | 01:02 |
fsmithred | I'm using it because I figure it's the most likely one to replace wicd | 01:03 |
leafwiz | Sure, if wicd was updated to python3 do you think people would like that better? | 01:03 |
fsmithred | I think we'd probably stick to that as the default | 01:04 |
fsmithred | got someone in mind to do the work? | 01:04 |
leafwiz | maybe.. :) Well, I was just thinking about updating all of the env statements in it to point to python2. That would prob be a nice small fix | 01:06 |
fsmithred | oh, that might work | 01:07 |
leafwiz | Do you know the source of the source-code devuan uses for its wicd install? | 01:08 |
fsmithred | yeah, getting the link now | 01:09 |
fsmithred | https://git.devuan.org/devuan/wicd | 01:09 |
fsmithred | hasn't been touched in a long time | 01:09 |
fsmithred | newest is probably in the unstable branch | 01:10 |
leafwiz | I see. | 01:10 |
fsmithred | 1.7.4+tb2-6+devuan1.1 is the version in beowulf and ceres | 01:11 |
leafwiz | yeah, so its the unstable branch that is the , well, newest? | 01:12 |
leafwiz | Ah.. I get it. It was the tag | 01:16 |
leafwiz | Sure okay, so that is the newest version which I should clone if I wanted to update the code | 01:20 |
leafwiz | I have been using Linux/bsd for a while, but i have never actually worked on other peoples code like this before :) | 01:20 |
leafwiz | I'll have a further look tomorrow. | 01:21 |
fsmithred | thanks | 01:58 |
rwp | I'll be sad if I have to leave wicd behind. I have been wanting to try out connman since I haven't ever tried it before. | 02:01 |
Xenguy | Yeah that's disappointing news, I use wicd as my daily driver | 02:01 |
golinux | Me too | 02:03 |
golinux | Hope someone gets it working for devuan. | 02:03 |
rwp | Since it is all python it would seem like that someone who was a python person could walk through it and convert it from 2 to 3 syntax relatively easily. | 02:05 |
rwp | I am not a python person but I keep getting told by python people how good it is... | 02:05 |
rwp | NetworkManager works acceptably well on personal laptops and things okay. Really not going to have a problem there. Since that is where it was designed to work. | 02:06 |
rwp | But my experience with it on servers is that it is *terrible* for remote systems. Too often it has left things disconnected and won't automatically reconnect. | 02:07 |
systemdlete | I think the latest firefox has a bad leak in it. | 04:09 |
systemdlete | I've been (trying to) watch Pluto TV and it eats up the memory in about a half hour or so. I've had to reboot the system about 3 times now. | 04:10 |
tuxd3v | systemdlete, probably it uses /tmp to store the video, and it will grow with time.. | 04:11 |
systemdlete | I call that leaking. Others may use other terms for it. | 04:14 |
systemdlete | tuxd3v: I think you are probably correct. | 04:14 |
systemdlete | But still, if it ends up eating all the memory, what is the difference what we call it? | 04:15 |
tuxd3v | hehe you need to be monitoring the /tmp | 04:15 |
tuxd3v | for example if you know that a episode consumes X memory | 04:15 |
systemdlete | Either firefox is run amok, or pluto's code has run amok. Either way, I don't think I have much control over it. | 04:15 |
systemdlete | (I'm actually measuring that very thing atm) | 04:15 |
systemdlete | So, essentially, I have to babysit Pluto's poor coding... right. | 04:16 |
tuxd3v | and you know the amount of ram you have in the computer, you need to restart the application so that he can cleanup the /tmp | 04:16 |
tuxd3v | where do you got that application? | 04:17 |
tuxd3v | you are watching directly from https://www.eu-landing.corp.pluto.tv/ ? | 04:17 |
systemdlete | I'm half way through an episode of All in the Family, and it has already eaten close to a gig of memory. | 04:18 |
tuxd3v | yeah specially 1080p it hits hard :( | 04:19 |
systemdlete | My TV decided to go moving partial black screen on me after just 2 years. So I am watching on my desktop. | 04:19 |
systemdlete | I've ordered a new TV, but it will be a few days... | 04:19 |
tuxd3v | what is the app you are using? | 04:19 |
systemdlete | firefox? | 04:19 |
systemdlete | is there pluto app for linux? | 04:20 |
systemdlete | I don't know which I love more, Pluto or Youtube. Neither one of them has nearly enough commercials. Neither one repeats the same ones nearly enough for MY liking. | 04:21 |
tuxd3v | I believe not :(, at least not yet | 04:21 |
tuxd3v | in the web version you need to create an account right? | 04:22 |
onefang | I tend to watch youtube (which is rare to be honest) via my VPN to a country where the main language is one I don't understand. Makes the adverts fun. | 04:23 |
onefang | Did leave me wondering why the Dutch are always trying to sell cars to children. | 04:24 |
* onefang stops being OT and wanders out for groceries. | 04:24 | |
* systemdlete hopes onefang FINDS groceries... | 04:31 | |
systemdlete | tuxd3v: I looked and noticed there were no temp files in /tmp opened by firefox | 04:32 |
systemdlete | I even killed the tab running pluto.tv | 04:32 |
numzob | ps -e something may still be running and holding that memory | 04:33 |
systemdlete | firefox has barely released memory, even a couple minutes after killing that tab. | 04:33 |
tuxd3v | systemdlete, it will probably get that ramspace somewere, a lot of applications use /tmp and store the video | 04:33 |
systemdlete | I'm watching the process in ps output, measuring usage every 30 seconds. | 04:34 |
systemdlete | I watched it run while my program was running. It continually ate memory, and even after I killed the tab running pluto, it still has not released all the memory it ate while pluto tab was open. | 04:35 |
systemdlete | The other tabs are more or less quiescent. At very least, I did not open any of them. | 04:35 |
tuxd3v | it could be a memory leak indeed | 04:35 |
systemdlete | Look, it would not be the first time firefox got a leak... | 04:35 |
systemdlete | And, depending on how pluto has "programmed" its javascript and such, who knows what impact that might have. | 04:36 |
systemdlete | Pluto could well be using one of those javascript include files with all the fancy routines that make coding js easy. | 04:37 |
systemdlete | And who knows what dwells within that... | 04:37 |
systemdlete | Some months ago, I had run into this memory management issue (leak) with ff. At that time, someone suggested I try to use ulimit to force oomkill or the like. That did not seem to work though. | 04:39 |
systemdlete | I'm guessing when one thread can not brk() any more, maybe it goes and grabs a new thread and continues eating memory there. But I'm not a threads programmer, so I really don't know. | 04:40 |
systemdlete | I'd really rather have firefox itself govern memory limits (which should be configurable, of course) and have a tab crash than having to reboot the entire system after it has strangulated it. | 04:41 |
systemdlete | s/after it/after ff/ | 04:42 |
tuxd3v | I just limit the amount of threads in preferences to 3 threads.. case closed :) | 04:51 |
rwp | systemdlete, My Firefox hovers around 4 GB of memory in steady state. I don't think I have ever seen it as small as only 1 GB. | 05:11 |
Criggie | ~3.5 hours till the flash killswitch goes off. | 08:35 |
Criggie | (for my timezone anyway) | 08:35 |
golinux | off topic (and does anyone really care?) | 08:48 |
gnarface | Criggie: there's #devaun-offtopic now | 08:56 |
Criggie | gnarface: thank you. | 09:28 |
APic | Hi. I want to upgrade my Server-VM from ascii to beowulf. aptitude shows me this: https://nopaste.linux-dev.org/?1326494 | 12:57 |
APic | What can i do please? | 12:57 |
APic | Current /etc/apt/sources.list.d/devuan.list is https://nopaste.linux-dev.org/?1326495 | 12:58 |
APic | What is maybe important that Devuan ascii was preinstalled on my VM | 13:01 |
fsmithred | APic, you're using a deprecated repository | 13:03 |
fsmithred | change packages.devuan.org to deb.devuan.org or choose a specific mirror from the list | 13:03 |
fsmithred | and you can comment out the deb-src lines if you're not downloading sources to rebuild packages | 13:04 |
APic | I do download Sources from Time to Time | 13:05 |
APic | Thanks for the quick Answer! | 13:05 |
fsmithred | edit, update and try again | 13:05 |
fsmithred | you shouold be getting libc6 2.28-10 | 13:06 |
APic | (apt update && aptitude full-upgrade) | tee apt.log → still https://nopaste.linux-dev.org/?1326498 | 13:10 |
APic | Maybe i should finish the few upgradable Packages it shows me first. apt full-upgrade gives me no Error like aptitude full-upgrade: https://nopaste.linux-dev.org/?1326499 | 13:13 |
n4dir | using apt-get i usually first upgrade, then dist-upgrade | 13:14 |
n4dir | no idea what the according switches for apt or aptitude are | 13:15 |
APic | full-upgrade _is_ dist-upgrade nowadays i think | 13:15 |
APic | Yup, dist-upgrade shows exactly the same Output as full-upgrade | 13:15 |
n4dir | i think it is the difference between apt, aptitude and apt-get. I meant first upgrade. Did you do that? | 13:15 |
n4dir | though i don't think it should result in problems. | 13:16 |
APic | lol | 13:17 |
APic | I did it, and now it created this: | 13:17 |
n4dir | also it is recommended to first upgrade the release one is coming from | 13:17 |
APic | # Devuan repositories | 13:17 |
APic | deb http://packages.devuan.org/merged unstable main | 13:17 |
APic | deb-src http://packages.devuan.org/merged unstable main | 13:17 |
APic | I think i want to change that back to beowulf again | 13:18 |
APic | unstable is ceres, no? | 13:18 |
n4dir | yes | 13:18 |
n4dir | you sure want to upgrade leaving out one release in between (even if you would want to go all the way up to unstable) | 13:18 |
n4dir | oh. you sure *don't want* | 13:19 |
APic | Yes | 13:19 |
APic | Nah, now it still shows the same broken fdisk-Deps | 13:22 |
APic | Maybe i shall try another Mirror than deb.devuan.org | 13:22 |
fsmithred | try installing libc6=2.28-10 | 13:23 |
fsmithred | is libc6-dev installed? I had trouble with that on a couple upgrades | 13:24 |
fsmithred | yes, sometimes you have to run the upgrade a few times to get everything | 13:25 |
APic | # aptitude install libc6=2.28-10 | 13:26 |
APic | Unable to find a version "2.28-10" for the package "libc6" | 13:26 |
APic | Unable to apply some actions, aborting | 13:26 |
fsmithred | maybe I need to update | 13:27 |
n4dir | Installed: 2.28-10 | 13:27 |
fsmithred | apt policy libc6 | 13:28 |
fsmithred | 2.28-10 500 | 13:28 |
fsmithred | 500 http://sledjhamr.org/devuan/merged beowulf/main amd64 Packages | 13:28 |
fsmithred | run 'apt update' | 13:28 |
fsmithred | after changing sources | 13:28 |
APic | libc6: | 13:28 |
APic | Installed: 2.24-11+deb9u4 | 13:28 |
APic | Candidate: 2.24-11+deb9u4 | 13:28 |
APic | Version table: | 13:28 |
APic | *** 2.24-11+deb9u4 100 | 13:28 |
APic | 100 /var/lib/dpkg/status | 13:29 |
fsmithred | yeah, that's the one you have installed. It's probably the ascii version. | 13:29 |
APic | Yes | 13:29 |
APic | Need to collect my Pizza now | 13:29 |
APic | cul8r | 13:29 |
n4dir | i would check anything in preferences or apt.conf in /etc/apt to be sure there is nothing standing in the way. The command "apt-get policy" should tell too, if i am correct | 13:30 |
n4dir | sure would tell about pinning? i think | 13:30 |
n4dir | as not sure what "apt-cache policy" tells me (not apt-get, sorry), i have to check the according files and folders manually anway. | 13:32 |
fsmithred | looks like he's running without dbus. That might cause trouble, too. But I'd expect it would say more about that. | 13:32 |
fsmithred | you can just do 'apt policy' now | 13:33 |
n4dir | i don't apt | 13:33 |
fsmithred | lol | 13:33 |
fsmithred | apt search gives sucky output | 13:33 |
n4dir | i saw a nice output, but it is more easy for me to stick to muscle memory. And in general i see no advantage in apt (yet) | 13:33 |
fsmithred | apt show and apt policy look the same as the other commands | 13:34 |
fsmithred | less to type | 13:34 |
n4dir | Well. Compared to what i typed only in this little chat, i can live with it. | 13:34 |
n4dir | i'd say if you don't like typing, use a different OS | 13:34 |
n4dir | or get used to command expansion/completion | 13:35 |
n4dir | but to each his own, sure | 13:35 |
DocScrutinizer05 | happy new year! | 13:39 |
n4dir | sounds like neither US, nor south America, nor Europe :-) But happy new year | 13:40 |
fsmithred | I think he's near you. | 13:43 |
fsmithred | happy new year (early) DocScrutinizer05 | 13:43 |
n4dir | 2 PM, some european regions might be 3 PM. Some 1. weird stuff, time | 13:43 |
n4dir | US is 6 or 7 AM, i assume | 13:44 |
fsmithred | yeah, a little before 8am now | 13:44 |
onefang | It'll be my turn to "happy new year" everyone in a bit over an hour. | 13:52 |
fsmithred | yeah, I think Doc has his AM and PM confused | 13:54 |
fsmithred | or he's in Japan today | 13:54 |
APic | Re | 14:00 |
APic | # apt-cache policy | 14:00 |
APic | Package files: | 14:00 |
APic | 100 /var/lib/dpkg/status | 14:00 |
APic | release a=now | 14:00 |
APic | 500 http://deb.devuan.org/devuan beowulf/main amd64 Packages | 14:00 |
APic | release v=3.0,o=Devuan,a=beowulf,n=beowulf,l=Devuan,c=main,b=amd64 | 14:00 |
APic | origin deb.devuan.org | 14:00 |
APic | Pinned packages: | 14:00 |
APic | Ok, sorry, will use a Pastebin in Future again | 14:01 |
fsmithred | did you run apt update after editing sources.list? | 14:01 |
fsmithred | you should be able to see the newer libc6 | 14:01 |
APic | Yes, i did. But now i changed it to http://sledjhamr.org/devuan/merged and now it shows: 573 packages can be upgraded. Run 'apt list --upgradable' to see them. | 14:03 |
APic | So deb.devuan.org/devuan seems b0rk? | 14:03 |
fsmithred | whatever mirror you're hitting with deb.devuan.org may be down | 14:03 |
APic | apt full-upgrade seems to work fine now: 573 upgraded, 155 newly installed, 6 to remove and 0 not upgraded. | 14:03 |
fsmithred | cool | 14:03 |
APic | ☺ | 14:03 |
APic | Thanks to You all | 14:03 |
fsmithred | yw | 14:04 |
n4dir | slow hand clapping from here. As in: respect | 14:04 |
fsmithred | here is the mirror checker: https://sledjhamr.org/apt-panopticon/results/Report-web.html | 14:08 |
fsmithred | and here is the mirror list, in case you want to pick one close to you: https://pkgmaster.devuan.org/mirror_list.txt | 14:09 |
APic | ic | 14:09 |
APic | Thanks | 14:09 |
fsmithred | yw | 14:09 |
onefang | Take note of the "EXPERIMENTAL CODE - double check all results you see here, and read the logs if it's important." warning at the top of https://sledjhamr.org/apt-panopticon/results/Report-web.html, there's a few bugs in it I need to fix. It's mostly accurate though. | 14:27 |
onefang | Hippy GNU Year! | 15:02 |
danyspin97 | can I report a bug regarding chimaere here or should I use #devuan-dev? | 15:02 |
danyspin97 | onefang: happy GNU Year :) | 15:02 |
fsmithred | danyspin97, you can start here | 15:18 |
danyspin97 | fsmithred: I see, thanks! | 15:19 |
fsmithred | don't paste a bunch of lines here | 15:20 |
danyspin97 | while building a chimaera version, we encountered a problem with btrfs. Some package that depends on btrfs-tools hasn't been updated | 15:20 |
danyspin97 | because btrfs-tools was a virtual provided by btrfs-progs in beowulf | 15:21 |
fsmithred | yeah, -tools doesn't exist | 15:21 |
fsmithred | what package wants btrfs-tools? | 15:22 |
danyspin97 | one of the following https://pastebin.com/HwshdKV9 | 15:24 |
danyspin97 | though I don't know which one for sure | 15:24 |
danyspin97 | Has this been already reported? | 15:24 |
fsmithred | ok, I'll look | 15:24 |
danyspin97 | thank you | 15:24 |
danyspin97 | this error has been generated by a CI, so I don't have shell access | 15:24 |
fsmithred | np | 15:25 |
fsmithred | once I remember how to get one package on a line, I'll figure out which one needs btrfs-tools | 15:26 |
danyspin97 | you need to split them in bash? | 15:26 |
fsmithred | yeah | 15:26 |
danyspin97 | for word in "pkg1 pkg2" | 15:27 |
fsmithred | I'll try that | 15:27 |
fsmithred | ok, nothing in that list depends on any package that has btrfs in the name | 15:30 |
fsmithred | danyspin97, was this live-sdk-beowulf #1 build? Or do you have a link so I can look at the console output? | 15:32 |
danyspin97 | fsmithred: yup, live-sdk-beowulf | 15:32 |
Guest32 | Hi All, sorry if this has been noted already but I've just tried to access `https://files.devuan.org/devuan_beowulf/Release_notes.txt` and got a cert expiry error. | 16:52 |
Guest32 | Looks like the Let's Encrypt cert expired a couple of days ago. | 16:52 |
Guest32 | Only realised because I've (finally) got around to upgrading from Ascii to Beowulf and thought that I ought to read the release notes... ;-) | 16:53 |
fsmithred | thanks, I passed the word along | 17:01 |
Guest32 | After a painful 2020, I would like to take this opportunity to thank everyone who has worked on Devuan in any way, and to wish you all a much better 2021! | 17:01 |
Guest32 | Thanks fsmithred! | 17:02 |
fsmithred | :) | 17:02 |
fsmithred | I think everyone is looking forward to the new year | 17:02 |
n4dir | if 10% of the news is right, than you sure don't hope for 2021. | 17:05 |
nemo | fsmithred: guess a cron job is having trouble, normally that sort of thing doesn't happen | 17:25 |
nemo | I do appreciate LE emailing me. avoided that situation on my teensy domain just last month | 17:25 |
APic | Happy YOLD 3187 to all of You ☺ | 17:29 |
systemdlete | rwp, tuxd3v: My memory snooping script shows that firefox grew in memory usage overnight -- as I slept -- so there is definitely something going on here. I will look into limiting the number of threads and memory, thanks for the suggestion. | 17:47 |
nemo | systemdlete: yeah, my SO had that issue when tabs were open on facebook or gmail | 18:09 |
nemo | systemdlete: something in the JS leaking memory | 18:10 |
nemo | systemdlete: about:memory for more info there, about:performance for a bit higher level... | 18:10 |
nemo | can also be due to addons. adblock with huge blocklists plus a dynamic site seems problematic. | 18:11 |
DashiePie | I was planning on upgrading to Chimaera today, and then I saw that Guest32 brought up that error, and wondered if that would affect my upgrade | 18:15 |
nemo | DashiePie: certificate is updated now | 18:17 |
nemo | 2021-03-31 expiration | 18:17 |
DashiePie | welp, that solves that, I guess | 18:17 |
nemo | plus that wasn't the update server anyway | 18:17 |
DashiePie | about the only thing I could tell you is that error codes are easy to google | 18:21 |
DashiePie | that's the extent of my knowledge on stuff like that | 18:21 |
gnarface | you really should read that file then | 18:23 |
DashiePie | the release notes? | 18:23 |
gnarface | yes | 18:24 |
nemo | gnarface: hm. he's already on beowulf though, there's no chimaera release notes yet, are there? | 18:25 |
DashiePie | how would the Beowulf release notes help my knowledge of server backends or my upgrade to Chimaera? | 18:25 |
DashiePie | also, what nemo said | 18:25 |
DashiePie | I'm already on Beowulf, cause that was the latest release when I installed it like, last month | 18:25 |
nemo | https://www.devuan.org/os/documentation/dev1fanboy/en/upgrade-to-beowulf might possibly be helpful mapped to chimaera, even though a few of the steps are obviously not necessary. | 18:27 |
nemo | (the jessie/ascii specific stuff) | 18:28 |
nemo | " | 18:29 |
nemo | This site is free of cookies and javascript" ♥ | 18:29 |
gnarface | oh, i just misread as him upgrading to Beowulf | 18:33 |
DashiePie | Guest was the one who did that, not me :P | 18:36 |
gnarface | yea i just got the relevance transposed because i wasn't paying close enough attention | 18:39 |
gnarface | i'm not sure that there's not already chimera release notes though | 18:40 |
fsmithred | no chimaera release notes yet | 18:40 |
nemo | fsmithred: so... chimaera will be released at same time as next debian stable? 'cause I was poking around and seems that is due out in 1 year? | 18:44 |
nemo | fsmithred: or will it be an incremental update to current stable? | 18:44 |
fsmithred | lol | 18:44 |
nemo | not sure how these things work | 18:44 |
fsmithred | when have we ever been in sync with debian releases? | 18:44 |
fsmithred | if we get it out less than a year after debian, we'll be doing better than before | 18:44 |
fsmithred | and actually there's a good chance of that because a lot of forked packages have already been forked | 18:45 |
nemo | fsmithred: ok. so... I guess I worded that badly | 18:45 |
nemo | fsmithred: "the absolute earliest chimaera could be stabilised is after next debian stable in at least a year?" | 18:45 |
fsmithred | yeah, I supposed it's possible we could release after final freeze and before debian release, and I keep hoping for that, but I doubt it will happen. | 18:46 |
n4dir | i don't even know when the next debian might happen. Next year? | 18:47 |
nemo | I thought their wiki said 2021-12 | 18:47 |
fsmithred | final freeze will be some time after March | 18:47 |
nemo | oh | 18:47 |
n4dir | ah. well. good enouogh | 18:47 |
nemo | oh.. just freeze | 18:47 |
fsmithred | beginning freeze might be in a couple weeks | 18:47 |
nemo | damn... we need to get another release out. *sigh* | 18:48 |
nemo | if only we weren't so laaaazy | 18:48 |
fsmithred | we need to get beowulf point-release out | 18:48 |
nemo | the backport system is a nice idea, but it does so suck for unimportant software like games where users still expect to have the latest hotness and have no idea how to do a backport install | 18:48 |
n4dir | something like a dual boot? | 18:49 |
nemo | wish there was another overlay like stable-noncritical-fun | 18:49 |
n4dir | There is that old list of Markus or Michal or such Kraft (or such), when you want and when you don't want debian | 18:49 |
nemo | every time we do a release where we miss the debian release we end up spending the next year or two supporting debian/ubuntu/mint users | 18:50 |
nemo | I guess if we can get it in before final freeze our odds are good | 18:50 |
fsmithred | DashiePie, you might want to look at this post. I made notes of my upgrade to chimaera back in july. Don't follow these instruction exactly because you'll probably run into different stuff. https://dev1galaxy.org/viewtopic.php?pid=26252#p26252 | 18:51 |
bmelo | will devuan really switch to openrc in next release? | 18:55 |
fsmithred | bmelo, who told you that? It's not true. | 18:57 |
fsmithred | the default is sysvinit and there are no plans to change that (yet) | 18:57 |
fsmithred | the installer gives you a choice of sysvinit, openrc or runit | 18:57 |
DashiePie | if there's no security or updates, then does that mean the single line does all of that? | 19:00 |
fsmithred | yeah | 19:01 |
DashiePie | ah, okay | 19:01 |
fsmithred | if you want fixes faster, add a line for ceres but also pin it to a lower priority | 19:01 |
fsmithred | then you can selectively install stuff before it moves down into chimaera | 19:01 |
fsmithred | if you're motivated to pay attention to which packages are at what stage | 19:02 |
DashiePie | I think what I need from Chimaera is just some updates so amdgpu stops crashing... if the problem is with amdgpu, and not my actual gpu | 19:02 |
fsmithred | oh | 19:02 |
fsmithred | first thing to try would probably be newer kernel from beowulf-backports | 19:03 |
fsmithred | and firmware-amd-graphics from the same | 19:03 |
gnarface | and maybe mesa too if it's in there | 19:10 |
nemo | DashiePie: yeah, the random crashing is worrying for sure... hope you're not getting overheating or bad vram | 19:11 |
nemo | DashiePie: but... I had a ton of probs w/ amd gpu on beowulf that went away in chimaera - mostly due to being able to ditch that stupid "pro" thing I was force-fitting into debian ☺ | 19:11 |
DashiePie | I hope so too, I want to play a bit of WoW again (1.12, compiled my own cmangos) | 19:11 |
nemo | but also 'cause their stupid "pro" thing was just buggy | 19:11 |
nemo | sent them a bunch of stack traces from games | 19:12 |
nemo | I think I'll only ever try that again if I wanna try OpenCL, and I certainly don't have time for that right now | 19:12 |
nemo | s/debian/devuan/ | 19:13 |
DocScrutinizer05 | >><fsmithred> I think he's near you.<< isn't the location of destination more relevant than that of source? I wish "happy b-day" when YOU have, not when I ;-D | 19:47 |
fsmithred | so you were speaking to the new zealanders? | 19:48 |
DocScrutinizer05 | to those, and "on pull when due" for everybody else | 19:48 |
n4dir | git pull? or is there a cron job? | 19:52 |
systemdlete | tried ulimit with memory set to 100 (which is k, per man page), then launched firefox, but ff grabs memory as usual, or at least I did not notice any constraint. The memory for just one thread was well over 500k | 20:38 |
systemdlete | is ulimit broken in recent linux kernels? Or am I doing something wrong. | 20:39 |
systemdlete | (and, yes, I started ff from the SAME shell as ulimit) | 20:39 |
systemdlete | So I went and tried cgroups. But that opens an entirely new can of worms. I had to install several packages for cgroups, then rebooted. | 20:40 |
systemdlete | so now the cgmanager is running, but if I try to create a cgroup, I am told that cgroups is not mounted. And on and on... | 20:40 |
nemo | systemdlete: I've successfully set ulimits for firefox, but it tends to crash things | 20:41 |
systemdlete | I'm ok with ff crashing. | 20:41 |
nemo | systemdlete: don't forget that the tabs are run in separate processes | 20:41 |
nemo | by default anyway | 20:41 |
nemo | no idea if that option can be disabled anymore | 20:41 |
systemdlete | I know. | 20:41 |
systemdlete | It doesn't matter for me. As long as each tab is limited, I'd be OK. | 20:42 |
systemdlete | Does devuan have a sort of "default" cgmanager/cgroup configuration? There is no /etc/cgmanager.conf or anything like that. | 20:42 |
systemdlete | And the instructions for cgroups seems to be written in Redhatese, not linuxese. | 20:43 |
systemdlete | (not too surprisingly I think) | 20:43 |
systemdlete | nemo: You say you've gotten ulimit to work? Or was that ulimits? | 20:44 |
systemdlete | I was using ulimit in bash. I can try ulimits; maybe bash's ulimit is buggy, idk. | 20:45 |
nemo | systemdlete: a few years ago. haven't had issues with firefox memory usage since (probably due to umatrix/noscript, but also maybe due to not using social media and gmail's crappy new interface) | 20:47 |
nemo | systemdlete: have you tried in a clean profile without addons? | 20:47 |
nemo | systemdlete: at the time I think I just set ulimits on any processes that matched firefox or content process names, and it just worked | 20:48 |
nemo | apart from crashing | 20:48 |
nemo | at the time it was necessary 'cause firefox grew to like 100% of system memory, so I restricted it to a few gigabytes. | 20:48 |
nemo | prefer crashing before locking up desktop | 20:48 |
nemo | I don't know of any way firefox could possibly dodge a system ulimit... | 20:48 |
nemo | so long as you keep doing it to new content processes that are spawned | 20:49 |
nemo | but debugging it might be worthwhile | 20:49 |
nemo | I'm a fan of about:memory | 20:49 |
nemo | was added as part of the memshrink project long ago | 20:49 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!