YKaelig | Hello | 00:15 |
---|---|---|
gnarface | hello YKaelig, if you have any questions just ask and be patient. this channel can be a bit slow but people do read their scrollbacks | 00:15 |
YKaelig | Please do not tell me that devuan Daedalus 5.0 does not support nvidia drivers 470.xxx version | 00:16 |
gnarface | it supports whatever version debian stable supports | 00:16 |
gnarface | plus whatever version debian backports supports | 00:16 |
YKaelig | Ok, so it doesn't | 00:16 |
YKaelig | ha! | 00:16 |
gnarface | which one was 470 from? | 00:16 |
YKaelig | didn't see anythign about that | 00:16 |
gnarface | do a search at pkginfo.devuan.org | 00:16 |
YKaelig | In Bulleyes 470.xxx was available | 00:17 |
YKaelig | nothing about nvidia from pkginfo.devuan.org | 00:18 |
gnarface | https://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=nvidia-driver&x=submit | 00:18 |
gnarface | looks like it's in chimaera and beowulf-backports | 00:18 |
YKaelig | my bad | 00:18 |
gnarface | so that's oldstable | 00:18 |
YKaelig | Not sure to understand so I can backport it in Daedalus or not ? | 00:20 |
gnarface | technically it wouldn't be a backport it would just be a port | 00:20 |
gnarface | and i don't have any idea if it would work or not | 00:20 |
gnarface | chances are you'd probably have to run the old kernel too | 00:20 |
gnarface | nvidia's behavior is really tragic | 00:21 |
gnarface | if you have the time and a solid backup though, you can try it with daedalus, it's worth a try | 00:21 |
YKaelig | Maybe there somewhere a tutorial to build it with the last kernel ? | 00:21 |
gnarface | i assume 525.xxx won't work with your hardware? | 00:21 |
YKaelig | GTX660 too old for nvidia but was a great card in the old days when ATI/AMD was a really horrible thing | 00:22 |
gnarface | yes i understand | 00:22 |
gnarface | ok, you can try installing chimaera then upgrade everything but the drivers | 00:22 |
YKaelig | and bulding only the driver with the last kernel available for Daedalus is not soemthing easier ? | 00:23 |
gnarface | it might work with the new kernel because of dkms i just don't know if it actually will, but you can always just try running the old kernel too | 00:23 |
gnarface | since they're non-free drivers i'm not sure just rebuilding them into a newer kernel is actually even an option, but i thought dkms was supposed to obviate that task | 00:24 |
debdog | I was thinking about installing chimaera then updating to daedalus _without_ updating the kernel or the 470 driver version? | 00:45 |
debdog | OR installing the driver from nvidia.com if that still works | 00:46 |
debdog | otherwise chimaera is the end of the line for my laptop | 00:47 |
gnarface | it might work. it might work with the new kernel too, you might just have to run dkms again manually, not sure. | 00:48 |
gnarface | worth a try i think, if you have a backup | 00:48 |
* debdog has quite a lot of backups for this laptop atm | 00:48 | |
gnarface | i would recommend to first just try upgrading with the nvidia packages pinned or non-free disabled or something like that | 00:53 |
YKaelig | Ok. Enough for today, Thx for your help. CU later have a good day/night | 01:02 |
systemdlete | "init 1" doesn't work, it seems. Isn't that supposed to work with sysvinit? | 02:49 |
systemdlete | (I always thought it did) | 02:49 |
systemdlete | or maybe I am missing a package or two for this. | 02:49 |
gnarface | try it with telinit? | 02:50 |
systemdlete | hmmm. telinit... I vaguely recall that. | 02:50 |
systemdlete | ok... I missed that. telinit is linked init (both in /sbin) | 02:52 |
gnarface | oh it's just a symlink, sorry | 02:52 |
gnarface | well, the man page suggests they behave differently despite that they should both work for what you're doing | 02:53 |
systemdlete | no, that's ok. What this tells me is that, if I installed my system correctly, it should be there. | 02:53 |
systemdlete | right. | 02:53 |
systemdlete | I'm still perfecting my custom install process. So maybe I just needed to set the PATH correctly, or explicitly run /sbin/init | 02:54 |
systemdlete | The default PATH when you land in the root prompt on the minimal install ISO boot may not have all of that set. | 02:54 |
gnarface | oh, yea you still need to edit /etc/login.defs or /sbin and /usr/sbin aren't in root's path | 02:56 |
gnarface | stupid debian vandalism, but i don't think that was new with this release | 02:56 |
systemdlete | well, actualy... | 02:56 |
systemdlete | /sbin might not be in the chroot environment I create | 02:57 |
systemdlete | or | 02:57 |
systemdlete | somehow that environment is not finding /sbin/init | 02:58 |
systemdlete | or | 02:58 |
gnarface | hmm, due to the usrmerge crap maybe it's all merged into /usr/sbin now? | 02:58 |
systemdlete | I just did not do that part of the install right. Though I doubt this lasst one, because I can boot the system from the drive after shutdown | 02:58 |
systemdlete | no | 02:58 |
systemdlete | I don't choose merged | 02:58 |
systemdlete | (or I don't think so anyway) | 02:58 |
gnarface | that can be prevented from fresh installs? | 02:59 |
gnarface | or you're making your own installer here? | 02:59 |
systemdlete | sort of | 02:59 |
systemdlete | I'm using the debootstrap | 02:59 |
gnarface | aah | 02:59 |
systemdlete | and I explicitly create /usr file system | 02:59 |
systemdlete | I'll try my installation process again, and see what is going on. | 03:00 |
systemdlete | It's almost 100% likely this is one more dumb thing I forget to do | 03:00 |
systemdlete | (I've been running into a ton of those... but it IS educational to me) | 03:00 |
systemdlete | the trouble is... the whole process takes about 5 hours | 03:01 |
systemdlete | in a VM | 03:01 |
systemdlete | with 4 CPUs | 03:01 |
gnarface | hmm | 03:01 |
systemdlete | and 2G memory | 03:01 |
systemdlete | I am finding that debootstrap takes roughly 40 minutes, very consistently. | 03:02 |
systemdlete | But the rest of the package install takes another 4 hours. | 03:02 |
gnarface | and that's with some sort of local repo or caching proxy in place even? | 03:03 |
systemdlete | I am setting the PKGSYSTEM_ENABLE_FSYNC=0 to avoid those long delays due to mime update's delay | 03:03 |
systemdlete | gnarface, no. | 03:03 |
gnarface | ah | 03:04 |
systemdlete | that has occurred to me though | 03:04 |
systemdlete | I should problably download, what, the full install ISO and mount it and use that as a local repo? | 03:04 |
gnarface | it helps a lot in a case like this where you end up having to run it again right away | 03:04 |
systemdlete | yeah. I've been at this for almost 2 weeks now. And I am making progress. Good progress. | 03:04 |
systemdlete | but the time it takes... whewwwweee | 03:05 |
gnarface | i think you can use the iso as a repo, yea, but i'd just run apt-cacher-ng | 03:05 |
systemdlete | hmmm. | 03:05 |
gnarface | i think any proxy would work actually | 03:06 |
gnarface | that way the packages say up to date too | 03:06 |
gnarface | *stay up to date | 03:06 |
systemdlete | avahi is needed for apt-cacher-ng I think | 03:06 |
gnarface | it wasn't before... | 03:07 |
systemdlete | I am looking at https://wiki.debian.org/AptCacherNg. Well, skimming it. | 03:08 |
gnarface | hmm, no... | 03:08 |
gnarface | Depends: libbz2-1.0, libc-ares2 (>= 1.16.0), libc6 (>= 2.34), libevent-2.1-7 (>= 2.1.8-stable), libevent-pthreads-2.1-7 (>= 2.1.8-stable), libfuse2 (>= 2.6), libgcc-s1 (>= 3.0), liblzma5 (>= 5.1.1alpha+20120614), libssl3 (>= 3.0.0), libstdc++6 (>= 12), libsystemd0 (>= 210), libwrap0 (>= 7.6-4~), zlib1g (>= 1:1.2.0), debconf (>= 0.5) | debconf-2.0, adduser, lsb-base (>= 3.0-6) | 03:08 |
gnarface | doesn't look like it to me | 03:08 |
systemdlete | So then I need to somehow pass my local repo to apt | 03:09 |
systemdlete | I guess I'd put my local repo in the /etc/apt/sources.list file in the install environment. | 03:11 |
systemdlete | Which I can easily do with my scripts I am already using. | 03:11 |
systemdlete | If I use the auto-apt-proxy, then I don't need avahi | 03:12 |
systemdlete | but would I need a proxy at all if it is all local to begin with? | 03:12 |
gnarface | i think there is two spots in the apt config you can put it | 03:12 |
gnarface | one line, "Acquire::http { Proxy ... " into a file in /etc/apt/apt.conf.d or something else the installer does instead | 03:13 |
gnarface | i forget the other one | 03:13 |
systemdlete | that web page states the same | 03:13 |
gnarface | and no, you would just use the proxy instead | 03:13 |
systemdlete | but my question is, why do I need a proxy at all if the repo is right here? | 03:14 |
systemdlete | I mean, if I am running a repo server, why the need for a proxy? | 03:14 |
gnarface | well if you're using the install disk as a repo it'll get stale eventually, and result in installs that still need to re-download a bunch of packages to be updated | 03:15 |
systemdlete | eh | 03:15 |
gnarface | if you're running an actual mirror then the proxy is useless | 03:15 |
systemdlete | not what I am asking, sorry | 03:15 |
systemdlete | so I am not running a mirror per se then | 03:15 |
systemdlete | I see | 03:15 |
systemdlete | ok, so that is the reason for a proxy. It acts as a frontend so it looks like a mirror to the installation process? | 03:16 |
gnarface | yes, it's transparent to the use of apt | 03:16 |
systemdlete | but is that really a "proxy?" I think of proxies a bit differently. | 03:17 |
systemdlete | I would call it middleware, or fudgeware, or interfacing-ware. | 03:17 |
gnarface | well, i thought so too, but i discovered that it still works if you plug the hostname and port in for the optional apt proxy field during the installation | 03:17 |
gnarface | it seems to be enough like a normal proxy that it still works like one | 03:18 |
gnarface | or at least apt doesn't make a distinction | 03:18 |
systemdlete | I meant the terminology. | 03:18 |
gnarface | i mean, apt makes a distinction technically, it puts it in a different config file, but it doesn't matter | 03:18 |
systemdlete | A proxy usually refers to something or someone that acts in place of something or someone else. | 03:18 |
gnarface | i've heard of people using squid for this too | 03:18 |
systemdlete | the same web page refers to using squid also, yes | 03:19 |
systemdlete | but I'd like to keep this simple for now. I'll be the only user. | 03:19 |
systemdlete | so no need for credentialling etc | 03:19 |
systemdlete | so... cho 'Acquire::http { Proxy "http://proxy:3142"; }' | sudo tee -a /etc/apt/apt.conf.d/proxy | 03:20 |
systemdlete | should do the job? | 03:20 |
systemdlete | (that's from the web page) | 03:20 |
systemdlete | not sure why I would need the "tee" | 03:21 |
systemdlete | I log everything the install is doing as it is. | 03:21 |
systemdlete | the important part of that line is the instruction being added to the conf.d directory | 03:21 |
gnarface | yea i think that should work but the debian convention is to name the files with numeric prefixes so you control the loading order | 03:21 |
systemdlete | Well, I'm generating the sources.list file also, so I can do that at the same time. And then I do apt update to ensure I am gong to the right places | 03:23 |
systemdlete | gnarface, let's say I specify "http://proxy:3142" as my proxy server. In the sources.list file, I'd think I would substitute the URL with this, right? | 03:27 |
systemdlete | or would it be http://proxy:3142/merged ? | 03:27 |
systemdlete | (just the URL part) | 03:27 |
gnarface | no, apt makes this transparent, your sources.list contents don't change | 03:27 |
gnarface | you're overcomplicating it | 03:27 |
systemdlete | so it overrides what is in sources.list? | 03:28 |
systemdlete | (sorry, just trying to get a hold of all of this) | 03:28 |
systemdlete | I mean, apt will then ignore what is in the sources.list file completely? | 03:28 |
gnarface | yea | 03:28 |
systemdlete | ok. | 03:28 |
gnarface | well, wait, i dunno for sure if it's completely... | 03:29 |
gnarface | no it probably still uses your sources.list file actually | 03:29 |
gnarface | it's just gonna make sure the traffic goes through the proxy | 03:29 |
systemdlete | I'd want to remove that conf.d file after the install. Unless I want to maintain my own local repo ongoing, which I don't think I need to do | 03:29 |
systemdlete | So it will ignore the URL specified in the sources.list and use the proxy instead. | 03:29 |
systemdlete | and it will still honor the suites etc in the sources.list file? | 03:30 |
gnarface | yea, it passes information from your sources.list to the proxy, which still fetches it from the servers you request, so long as they're whitelisted in the apt-cacher-ng config | 03:31 |
systemdlete | "the servers I request"--> the proxy? | 03:31 |
gnarface | the servers you put in your sources.list | 03:31 |
systemdlete | oh | 03:31 |
systemdlete | oh I see | 03:32 |
systemdlete | It will only substitute the URL for those servers that are not in the whitelist | 03:32 |
systemdlete | When I install the proxy on my network somewhere, does it start downloading the repos? | 03:33 |
gnarface | ? i think it just doesn't get used for a server not in the whitelist | 03:33 |
rrq | it's enough running it with exported end http_proxy=http://proxy:3142 | 03:33 |
rrq | don't need the conf file | 03:33 |
rrq | ... exported env ... | 03:34 |
gnarface | no, it won't download anything until it receives a request from one of your apt clients. that means it won't speed anything up the first time either. | 03:34 |
systemdlete | ok | 03:34 |
systemdlete | that's fine. Just wondering what to expect. thx | 03:34 |
gnarface | but if you're doing lots of test runs in a row... this could save you a lot of time | 03:34 |
systemdlete | yes! | 03:34 |
systemdlete | what I am looking for. | 03:34 |
gnarface | also, if you're doing multiple parallel test runs, this can scale well on a lan up to like 1000 of them | 03:34 |
systemdlete | maybe 2 in parallel, one VM and one host | 03:35 |
systemdlete | funny, that was what I was just about to test in fact. | 03:35 |
systemdlete | you must have looked in my read-ahead buffer | 03:35 |
systemdlete | I'm sure what you say is true though. The amount of traffic I have put through these past 2 weeks is probalby enough to put me over my ISP's download limit before they start charging me even more for their crummy service | 03:36 |
systemdlete | Just curious, where does the proxy stash its stash? Is it in the same archive as whatever system is hosting it, or does it establish an enirely new cache (I'm thinking disk space) | 03:37 |
systemdlete | by "archive" I mean /var/cache/apt/archive | 03:38 |
rrq | apt-cacher-ng uses /var/cache apt-caher-ng/ by default; configurable | 03:39 |
systemdlete | aha! | 03:40 |
systemdlete | no, that's good enough. I can just extend that file system if I need space. | 03:40 |
systemdlete | thanks rrq | 03:40 |
systemdlete | and gnarface | 03:40 |
systemdlete | well, I can't wait to try this... | 03:41 |
gnarface | no problem | 03:43 |
Anto | Hi, I'm new to IRC, are people under 18 allowed on Libera? I just turned 15 | 05:03 |
mason | Anto: Not seeing anything about age in the policies: https://libera.chat/policies/ | 05:12 |
Anto | Oh, so that's where I should have started, haha, thanks then | 05:15 |
golinux | Sometimes that info it hard to find | 06:18 |
systemdlete | gnarface: I like this web UI for apt-cacher-ng--I can see the amount of work it has done | 06:31 |
systemdlete | This is the first run, so there won't be benefit yet. | 06:31 |
systemdlete | I only discovered it by browsing to the URL I gave it and noticed a button to take me to the mgmt interface. | 06:31 |
systemdlete | They've got all kinds of maintenance tools too. | 06:32 |
systemdlete | thanks for this suggestion--it was waaaaay toooo eeeeasy | 06:32 |
systemdlete | I'm looking forward to the 2nd run... | 06:32 |
YKaelig | Hello | 12:12 |
djph | hi YKaelig | 12:13 |
YKaelig | From the devuan netinstall.iso where can I find the list of packages that are hidden behind the "Devuan desktop environment" and "standard system utilities" naming in the [ Software selection ] | 12:14 |
djph | Hm ... I forget offhand how to see what the installer bundles under those categories. Digging for info ;) | 12:24 |
YKaelig | Cool :) | 12:24 |
YKaelig | Well, It's a long time I didn't used sysv. Openssh-server has been installed but it is not running. How am I suppose to start the server ? | 12:25 |
YKaelig | that weird, it shouldn't be /etc/init.d/sshd ? because I have /etc/init.dssh instead | 12:27 |
YKaelig | I mean /etc/init.d/ssh instead of /etc/init.d/sshd | 12:27 |
YKaelig | ok, it's fine | 12:27 |
YKaelig | lol | 12:27 |
fsmithred | apt show task-xfce-desktop | 12:28 |
fsmithred | look at the deps | 12:28 |
djph | bah, beat me to it :P | 12:28 |
fsmithred | ssh not sshd | 12:28 |
djph | I *just* found it | 12:28 |
YKaelig | no, still can not connect to the remote | 12:29 |
djph | well, is the remote host running a firewall that's allowing port 22? is it actually listening on port 22? | 12:29 |
fsmithred | what message did you get? | 12:29 |
YKaelig | its fine, I'm too old, I need glasses | 12:30 |
YKaelig | :D | 12:30 |
fsmithred | the server needs to be running on the remote | 12:30 |
fsmithred | same here or I would have seen that you said ssh before I did | 12:30 |
YKaelig | ok, so I'm going to follow a tutorial I found to quickly create a private backport of a package, in my case it's already about nvidida. The explanation is pretty light but well I will see | 12:33 |
YKaelig | well, indeed by default the 470.xxx nvidia version is not available, so | 12:39 |
YKaelig | this is the version I need nvidia-kernel-dkms_470.199.02-1_amd64.deb | 12:41 |
YKaelig | looking in the archive I can not find for which debian it has been built. | 12:42 |
fsmithred | apt policy nvidia-kernel-dkms | 12:47 |
fsmithred | I'm in chimaera so I'm seeing a couple versions here. Try that command in daedalus. | 12:47 |
fsmithred | or are you running something else? | 12:48 |
YKaelig | running devuan daedalus | 12:48 |
fsmithred | good choice | 12:48 |
fsmithred | :) | 12:48 |
YKaelig | I don't know ^^ | 12:49 |
DelTomix | then good guess :) | 12:49 |
fsmithred | you're not the only one having trouble with nvidia right now. | 12:49 |
YKaelig | nice, so let's go together :D | 12:50 |
fsmithred | did that command give you any output/ | 12:51 |
fsmithred | ? | 12:51 |
YKaelig | I don't know how to import the source file. Manually ? by adding the chimera package_src in the source list ? That weird | 12:51 |
YKaelig | fsmithred If you mean the apt policy, yes. | 12:52 |
fsmithred | to get a source package from chimaera, yes you can add a chimaera src line | 12:52 |
fsmithred | did it find the package? | 12:52 |
fsmithred | do you have contrib and non-free enabled in sources.list? | 12:53 |
YKaelig | yes | 12:53 |
YKaelig | yes, but the version of the package is not the good one for my graphics card | 12:54 |
YKaelig | that why I'm trying to import the old one available on chimaera in daedalus | 12:55 |
fsmithred | here are the versions in chimaera: | 12:55 |
onefang | Also non-free-firmware, or whatever it's called this year. | 12:55 |
fsmithred | 470.182.03-1 500 | 12:55 |
fsmithred | 500 http://deb.devuan.org/merged chimaera/non-free amd64 Packages | 12:55 |
fsmithred | 470.103.01-1~bpo11+1 100 | 12:55 |
fsmithred | 100 http://deb.devuan.org/merged chimaera-backports/non-free amd64 Packages | 12:55 |
fsmithred | where did you see 470.199? | 12:56 |
YKaelig | There is no nvidia-kernel-dkms_470.199.02-1_amd64.deb ? | 12:57 |
fsmithred | I don't see it | 12:57 |
fsmithred | found it on pkginfo.devuan.org | 12:58 |
fsmithred | 470.199.02-1 | 12:58 |
fsmithred | http://deb.devuan.org/merged | 12:58 |
fsmithred | chimaera-proposed-updates/non-free | 12:58 |
fsmithred | amd64 | 12:58 |
YKaelig | I found this package from a mirror in the control file I see : Section: non-free/kernel but maybe it's only availbale on debian | 12:58 |
fsmithred | so add a source line for chimaera-proposed-updates | 12:58 |
fsmithred | apt -t chimaera-proposed-updates source nvidia-kernel-dkms | 12:59 |
fsmithred | if that doesn't download the source, then just pull it from the server. I can help you find it. | 12:59 |
YKaelig | ok, thx. Let's see what can I do right now with all of your help | 13:00 |
onefang | Oh, you are on Chimaera, not Daeialus. Ignore the non-free-firmware thing then. | 13:00 |
YKaelig | no I"m on Daedalus | 13:01 |
fsmithred | he wants to port a chimaera version to daedalus | 13:02 |
fsmithred | it might be easier to use the nvidia binary instead | 13:03 |
onefang | OK the point I was making is maybe this nvidia driver is in non-free-firmware? Which is only available for Daedalus, not Chimaera. | 13:04 |
fsmithred | looks like the nvidia stuff is still in non-free | 13:04 |
Anto | Wasn't there a new Nvidia driver made by the same people of Nouveau, but this time using good code, that could actually perform on par with the Nvidia drivers? | 13:05 |
YKaelig | Maybe, yes. I have read that the installer can use system dkms to build the module. But I want to learn about the official devuan method of package backporting / building | 13:06 |
YKaelig | how can I search only for a source package. I mean when I run apt search it's only looking at binaries, but how can I search for sources ? | 13:12 |
Anto | I've never packaged anything, besides the OpenARENA Quake3e flatpak and an Appimage, ArcticFox(Firefox backport for old PCs, and PowerPC based systems) | 13:12 |
fsmithred | what do you need to search for? | 13:14 |
fsmithred | apt source <package> will download it | 13:15 |
fsmithred | apt show package | 13:15 |
djph | Can't recall -- are the source repos enabled by default? | 13:15 |
fsmithred | to get dependencies and name of source package | 13:15 |
fsmithred | I think the installer isos leaves src lines enabled. i disabled them in the live isos. | 13:16 |
djph | yeah, I haven't used the installers since uh .. beowulf(?) ... when did I install this latest laptop again | 13:17 |
YKaelig | I mean I would like to find in my terminal what I can find in the Sources.xz file which is on my case : Package: nvidia-graphics-drivers | 13:23 |
YKaelig | but when I run apt search nvidia-graphics-drivers there is nothing, maybe I need to add an option to apt ? | 13:23 |
YKaelig | I don't think I made a typo or mistake in the source.list | 13:24 |
YKaelig | deb-src http://deb.devuan.org/merged chimaera-proposed-updates main contrib non-free | 13:24 |
fsmithred | I have chimaera-proposed-updates enabled and I'm not seeing the version that's there | 13:31 |
fsmithred | ...according to pkginfo | 13:32 |
YKaelig | well, not really what I needed but at least I can see that the sources are available and can be reached apt-get source -t chimaera-proposed-updates -s nvidia-graphics-drivers | 13:32 |
YKaelig | Selected version '470.199.02-1' (chimaera-proposed-updates) for nvidia-graphics-drivers | 13:32 |
onefang | I used chimaera-backports to get stuff for my recentish AMD graphics card. | 13:32 |
fsmithred | I see the other version now that I fixed my sources.list | 13:36 |
YKaelig | Re | 17:56 |
YKaelig | So. Which method do you recommend for a local repository ? | 17:57 |
YKaelig | In my opinion "aptly" seems to be a good candidate. What do you think about that ? | 17:58 |
gnarface | YKaelig: last time i had to, i used reprepro | 18:01 |
gnarface | it wasn't hard to setup | 18:01 |
YKaelig | https://wiki.debian.org/DebianRepository/Setup | 18:01 |
gnarface | you managed to build those nvidia 470 drivers against the daedalus 6.1 kernel? | 18:01 |
YKaelig | gnarface well seems that the opinion from the writer is different " not for beginners" ^^ | 18:02 |
YKaelig | gnarface yes | 18:02 |
gnarface | cool, that's kinda a big deal | 18:03 |
YKaelig | well, Not really but it is still not installed so maybe I will have bad surprise | 18:04 |
gnarface | i'm pretty sure since you brought it up at least 2 more people mentioned they were in the same position with a 600-series laptop GPU. if you get this working, the work will benefit others. then shortly after that NVidia will probably assassinate you. | 18:04 |
YKaelig | I would like to use a local repository for that | 18:04 |
gnarface | well, last i used reprepro was debian wheezy, so maybe there's something easier out there now | 18:05 |
YKaelig | Well, I took notes so it is not a problem to share tehm | 18:05 |
gnarface | debdog, onefang, i think you guys were interested in this? ^^^ | 18:06 |
gnarface | at the time, reprepro just appeared to be the most direct way to get a private repo that was not a mirror, just a repo for some out-of-distro package customizations | 18:07 |
gnarface | using it as a full or partial mirror might take extra steps | 18:08 |
gnarface | but a lot of the docs out there about setting up repos were specifically focused on mirror repos | 18:09 |
debdog | YKaelig: I am very interested in your notes! | 18:16 |
YKaelig | ok, well I'm going to use dpkg-scanpackages which is the quickest way to build a local repository. Need to reboot on Devuan. See you in a few moments | 18:19 |
Guest5 | I am trying devuan DAEDALUS LIVE (devuan_daedalus_5.0.0_amd64_desktop-live.iso) on a lenovo x200 laptop with 8gb RAM where I had tried devuan CHIMAERA LIVE many, many times. | 19:13 |
Guest5 | I am getting the feeling that this new DAEDALUS LIVE goes slower than chimaera (I know very well how chimaera works after having using it many times). When I open a text document in mousepad I can clearly see the difference when scrolling up and down, it sometimes go sloppy in Daedalus, and this did not happen in previous devuan versions. Has there | 19:13 |
Guest5 | been any major change in debian that has led to this behavior? Why has this happened with Chimaera but not with previous versions? Is this going to be fixed in futre updates? | 19:13 |
Guest5 | Sorry, at the end of the text I made a type, I meant: Why has this happened with DAEDALUS but not with previous versions? | 19:14 |
plasma41 | Guest5: I | 19:17 |
plasma41 | Guest5: Do you experience a similar slow down comparing Debian Bullseye Live to Bookworm Live? | 19:18 |
gnarface | already gone | 19:19 |
plasma41 | oh well | 19:19 |
fsmithred | Every release is slower than the previous one. | 19:31 |
fsmithred | *uses more resources* | 19:31 |
gnarface | i wouldn't expect it to cause a noticeable regression in window scrolling when i/o wasn't saturated though unless there was a video driver regression though | 19:32 |
gnarface | at 8GB it shouldn't be ram constrained | 19:33 |
fsmithred | true | 19:33 |
gnarface | and the video driver for that hardware should be faster... | 19:33 |
gnarface | so it does seem like a valid complaint, but more information will be needed for diagnosis. too bad they didn't stick around | 19:34 |
fsmithred | yeah, I'm wondering if it was a live-usb or something else | 19:34 |
fsmithred | I'm about to boot on a thinkpad with 8G ram | 19:35 |
gnarface | the first thing i thought was maybe we have another laptop with the freshly deprecated 600-series nvidia gpu, but my searches don't suggest this model came in any configuration other than intel-only video | 19:37 |
gnarface | (if the user's install had quietly failed over to nouveau, this is exactly the type of complaint i'd have expected) | 19:38 |
gnarface | but intel's latest work on mesa and their kernel driver should actually make their embedded graphics of the same generation faster than those 600 series nvidia gpus for this type of workload | 19:40 |
gnarface | so, definitely not the type of complaint i'd expect | 19:40 |
gnarface | "aak! free -mt reports over 500MB used after boot without a graphical environment when it use to be just 145!" < that's the type of complaint i'd actually expect | 19:41 |
fsmithred | scrolling is ok on the installed daedalus | 19:41 |
gnarface | hmm, well, the user didn't say where they got the image but they used an accurate file name | 19:42 |
fsmithred | I don't have that on a usb to test right now | 19:45 |
gnarface | i don't think i have anything comparable to test with here | 19:51 |
gnarface | it looks a couple generations newer than the most recent (2GB) x86 laptop i have here with intel graphics | 19:52 |
fsmithred | ok, I put the desktop-live on usb and I guess the two-finger scrolling on the trackpad is a little jumpy. | 20:06 |
fsmithred | seems a little more sensitive than I would like | 20:06 |
fsmithred | but it's usable | 20:07 |
gnarface | you think that's what the user was mistaking as a framerate drop? | 20:07 |
fsmithred | touchpad? | 20:07 |
gnarface | yea | 20:07 |
gnarface | "scrolling up and down, it sometimes go sloppy in Daedalus" | 20:07 |
gnarface | first time i read this i misread "sloppy" as "choppy" but now i realize it's more ambiguous than i though | 20:08 |
gnarface | than i thought* | 20:08 |
fsmithred | yeah, it might be consideres sloppy | 20:08 |
fsmithred | the sloppiness is actually in the fingers. | 20:08 |
fsmithred | ok, I can adjust the speed of the cursor (acceleration) but there's no sensitivity adjustment for the two-finger scrolling. | 20:10 |
fsmithred | I notice this problem if I boot a live iso in qemu. It's almost unusable. | 20:11 |
fsmithred | and not restricted to daedalus. | 20:11 |
YKaelig | Well. Some issues look like I missed to build others packages https://dpaste.com/G8LUGAEFX but looks like it is going in the right direction | 20:23 |
gnarface | yea, you probably need to build a good 20 or so packages to make this work. | 20:23 |
gnarface | a bunch of nvidia packages and possibly one or more xorg packages | 20:24 |
YKaelig | I'm little confuse about that https://packages.debian.org/sid/all/firmware-nvidia-gsp-470.199.02 | 20:34 |
YKaelig | How am I going to add a virtual package in my work ? | 20:35 |
gnarface | YKaelig: not sure, and not sure you need to if you have all the packages that were in the virtual package anyway, but also this isn't one of the nvidia packages you need for your hardware anyway | 20:41 |
gnarface | -tesla- is something else | 20:41 |
gnarface | ignore the nvidia packages with "tesla" and "legacy" in their names while you're doing this | 20:42 |
gnarface | tesla hardware is for AI or mining rigs or something like that | 20:43 |
gnarface | here seems to have been some false positive happening where it mis-detects the freshly deprecated 600 cards as tesla hardware. i have my suspicions about why but it could just be a normal mistake... | 20:44 |
gnarface | there seems* | 20:45 |
systemdlete | gnarface, if you are around... I'm trying to use the apt-cacher-ng and so far, it looks like it is "working" in the sense that the installer can reach the proxy, and I can see that there is at least some activity on the cache server (per that webui page). But the problem is that it is only caching for the chimaera release, not daedalus. | 21:00 |
gnarface | systemdlete: you gotta edit some of the config files | 21:01 |
systemdlete | So I am thinking of adding daedalus to the apt config and assigning priorities so that the local (server) will get its chimaera packages, but the proxy can serve a variety of distros | 21:01 |
systemdlete | I've read about priorities inthe preferences page, but I'm still not sure how to go about this | 21:02 |
gnarface | hmm, i think you might still be overcomplicating this in your head... | 21:02 |
systemdlete | well, I need to serve up daedalus in my proxy, not chimaera. | 21:02 |
systemdlete | believe me, I am looking for simplicity, not complexity. | 21:03 |
systemdlete | Not sure why you are saying that. | 21:03 |
systemdlete | I just want to get this to work, if it will. | 21:03 |
gnarface | well, mabye the newer apt-cacher-ng works differently, but i don't have problems with mine caching for every release in parallel | 21:03 |
systemdlete | If I go into the cache area for apt-cacher-ng, I only see chimaera | 21:05 |
gnarface | my cache directory has files for jessie, ascii, chimaera, daedalus, beowulf and ceres, some of those in both amd64 and arm64 flavors | 21:05 |
systemdlete | ok... | 21:05 |
systemdlete | how does the proxy know where to get daedalus from? | 21:05 |
gnarface | from the http request i assume | 21:05 |
systemdlete | doesn't that need to be in /etc/apt/apt.conf.d | 21:05 |
gnarface | let's do some review | 21:05 |
systemdlete | ok | 21:05 |
systemdlete | (thanks) | 21:06 |
gnarface | do you have /etc/apt-cacher-ng/acng.conf ? | 21:06 |
systemdlete | yep | 21:06 |
gnarface | Remap-devrep: file:devuan_mirrors /merged ; file:backends_devuan # Devuan Archives | 21:06 |
gnarface | in mine, this is the only line that devuan appears on | 21:06 |
gnarface | do you have this line? | 21:06 |
gnarface | i had to add it myself, i'm pretty sure | 21:07 |
systemdlete | hold on; switching machines... | 21:07 |
gnarface | i also had to add one for winehq | 21:07 |
gnarface | there's a bunch of stock ones but nothing i'm actually using so i just ignored them | 21:07 |
gnarface | Remap-winehq: file:winehq_mirrors /wine-builds/debian ; file:backends_winehq # WineHQ Archives | 21:07 |
gnarface | (there's the winehq one for reference, not very important) | 21:08 |
systemdlete | no remap line like yours | 21:08 |
gnarface | ok, so probably what is happening is the cache directories are clobbering each other | 21:09 |
systemdlete | what I have is: | 21:09 |
systemdlete | Remap-debrep: file:deb_mirror*.gz /debian ; file:backends_debian # Debian Archives | 21:09 |
gnarface | yea that won't help for devuan | 21:10 |
systemdlete | Mine says debrep, not devrep, so that might be a clue? | 21:10 |
gnarface | i'm not sure that part actually matters | 21:11 |
systemdlete | so do I replace or add your line? | 21:11 |
gnarface | one of them is labeled "Remap-fedora:" | 21:11 |
gnarface | just add my line, ignore the others, they're harmless | 21:11 |
systemdlete | ok, hold on | 21:11 |
gnarface | then, assuming it hasn't changed too much in the new version, add an EMPTY FILE at /etc/apt-cacher-ng/backends_devuan | 21:13 |
systemdlete | So I added the line, tried to restart the cacher, got a warning | 21:13 |
gnarface | we're not done yet | 21:13 |
gnarface | we gotta touch like 3 files here | 21:13 |
systemdlete | ok | 21:14 |
gnarface | add the EMPTY FILE at /etc/apt-cacher-ng/backends_devuan for starters | 21:14 |
gnarface | i also have a symlink from /etc/apt-cacher-ng/backends_devuan.default pointing to /var/lib/apt-cacher-ng/backends_devuan.default | 21:14 |
systemdlete | ok, got it | 21:14 |
gnarface | the contents of my /var/lib/apt-cacher-ng/backends_devuan.default file are as follows: https://paste.debian.net/1291287/ | 21:15 |
gnarface | i also have a file at /usr/lib/apt-cacher-ng/devuan_mirrors whose only contents are one line containing: http://deb.devuan.org | 21:16 |
gnarface | my build isn't properly patched for devuan in any sense, i just copied from the stock examples included | 21:16 |
gnarface | but i noticed it worked for winehq too and will probably work for anything you set up this way | 21:17 |
gnarface | i think this is basically just a song and dance to tell it how to organize the cache directory, and in lieu of these config files the default behavior is to just clobber the top level of the cache directory with whatever the last thing you used was | 21:19 |
YKaelig | Mmm, looks like it is working. But can I get a full description of packages when I run apt install, not just the name but the name+version | 21:19 |
gnarface | YKaelig: apt doesn't list them by default? what if you use apt-get instead? | 21:20 |
YKaelig | no | 21:20 |
YKaelig | wait | 21:20 |
gnarface | maybe see if apt just has a --verbose flag or something like that | 21:20 |
YKaelig | same | 21:20 |
gnarface | hmm, odd | 21:20 |
systemdlete | so restart now? | 21:22 |
gnarface | systemdlete: you got all those changes? honestly i'm not sure if it even needs to be restarted but it couldn't hurt. | 21:22 |
gnarface | YKaelig: what about apt list or apt show? | 21:22 |
systemdlete | same error as before | 21:25 |
gnarface | hmm, what's the error? | 21:25 |
systemdlete | I tried symlinking devuan-mirrors from the /var/lib/apt-cacher-ng directory | 21:26 |
gnarface | keep in mind my version is ancient so some of the config syntax may have changed, you may need to adapt my setup to the new context | 21:26 |
systemdlete | the error, again, is: Restarting apt-cacher-ng: apt-cacher-ngWARNING: No configuration was read from file:devuan_mirrors | 21:26 |
systemdlete | WARNING: No configuration was read from file:backends_devuan | 21:26 |
systemdlete | sorry | 21:27 |
systemdlete | Error opening file /etc/apt-cacher-ng/devuan_mirrors (Permission denied), terminating. | 21:27 |
gnarface | well make sure the permissions are right | 21:27 |
systemdlete | (that's my symlink to the /var/lib directory) | 21:27 |
systemdlete | this is root... | 21:27 |
systemdlete | or is apt running this? | 21:27 |
gnarface | its own user runs it, yes, not root | 21:28 |
systemdlete | actually, I just added read-world permissions to the new files I created in /var/lib/apt-cacher-ng | 21:29 |
systemdlete | it starts fine now, thanks | 21:29 |
YKaelig | -V to get the package version | 21:29 |
systemdlete | so Now I should write this up before I forget (which I will) | 21:30 |
gnarface | systemdlete: just remember your /etc/apt-cacher-ng/security.conf has the admin password in it, so make sure that one is not globally readable. it should only be readable by root and the apt-cacher-ng user. the rest are non-sensitive. | 21:30 |
YKaelig | well, thx to this option still some mix with nvidia version 525.xxx for some softwares. I don't know how is that even possible | 21:31 |
YKaelig | for example I can install nvidia-modprobe (535.54.03-1~deb12u1) | 21:31 |
gnarface | YKaelig: what a mess! you can probably just block them out with apt pinning | 21:31 |
gnarface | YKaelig: their package dependencies have always seemed like a mess to me, in general, so this isn't surprising | 21:32 |
YKaelig | time to build a new package manager :D | 21:33 |
gnarface | hah | 21:33 |
systemdlete | gnarface, I haven't touched security.conf, but I'll add that to the notes | 21:33 |
gnarface | systemdlete: oh, you might have to configure an admin user in acng.conf first, i forget | 21:33 |
systemdlete | It seems to be working without it... | 21:34 |
systemdlete | this is only on and for my own LAN | 21:34 |
gnarface | as i recall, it just gives you a couple extra powers on the admin interface | 21:34 |
YKaelig | I have some tabs precisely about the linglong package manager from deepin. I'm really curious how it work | 21:35 |
YKaelig | It can also install window softwares | 21:35 |
gnarface | systemdlete: it should be fine then, but i don't think i would trust it exposed to the public internet either way | 21:36 |
systemdlete | right. But for my purposes... | 21:36 |
systemdlete | gnarface, if I EVER were to expose any part of my network to the internet, i would certainly take EXTRA care. Don't you worry about THAT! | 21:37 |
gnarface | the primary benefit benefit of this one over regular apt-cacher was listed as being easier to setup | 21:37 |
gnarface | after having tried and failed to set up apt-cacher i was like "okay!' | 21:38 |
gnarface | heh | 21:38 |
systemdlete | I think it would be easier, but those devuan-specific files would need to be included. Then it would be a snap. | 21:38 |
gnarface | so, that's a very old conversation, and there's probably still actually even a patch for it on the bug tracker... but someone would have to volunteer to maintain the package and also the discussion of "do we really want one more forked package for this?" would have to happen... | 21:41 |
gnarface | so it seems like a really easy thing to get done but in practice not so easy | 21:41 |
gnarface | and it's not like a difficult thing to actually fix manually | 21:42 |
gnarface | doesn't require rebuilding the package or anything | 21:42 |
gnarface | (not my choice, just repeating what i've been told) | 21:43 |
Guest51 | "fsmithred yeah, I'm wondering if it was a live-usb or something else 19:34 | 21:45 |
Guest51 | fsmithred I'm about to boot on a thinkpad with 8G ram" | 21:45 |
Guest51 | No, all this was using "boot=live toram" so not a problem of slow usb I/O | 21:45 |
gnarface | Guest51: read further, he noticed a touchpad issue, that could be described like that... can you confirm? | 21:46 |
fsmithred | Guest51, can you describe the symptoms? | 21:46 |
YKaelig | well, it's going to be a little more work than expected, nvidia-modprobre 470.199.02 is not available as a debian package source so I will have to create it. Awesome ! :D | 21:46 |
Guest51 | gnarface, it happens with USB mouse too, not just with the touchpad. | 21:47 |
fsmithred | what is "it"? | 21:47 |
gnarface | YKaelig: does the previous version have it? | 21:48 |
fsmithred | did I miss anything? | 21:48 |
gnarface | no | 21:48 |
YKaelig | yes. That my idea, download an previous version a adapt it for the version I need. But I need to learn how I'm suppose to do that in a right way | 21:49 |
YKaelig | a/and | 21:49 |
gnarface | YKaelig: i would strongly just consider switching everything to the previous version to make it work | 21:49 |
YKaelig | There is certainly a tutorial somewhere | 21:49 |
gnarface | https://www.debian.org/doc/manuals/maint-guide/update.en.html ? | 21:50 |
YKaelig | Don't understand what do you mean. The version 470.199.02 is available here https://salsa.debian.org/nvidia-team/nvidia-modprobe/-/tags/470.199.02 just need to create a dsc file and some others things | 21:50 |
gnarface | oh, i assumed you wouldn't be able to get it | 21:51 |
YKaelig | thx gnarface | 21:51 |
YKaelig | no problemo ^^ | 21:51 |
Guest51 | fsmithred. I wish I could be more specific, but in general I feel that the computer goes slower when compared to chimaera. Maybe I should test it more extensively for a few days and see what happens in more detail. But I think it is the operating system, and not the programs that make it slow. For example, I can boot the x200 with chimaera and | 21:51 |
Guest51 | install the newest firefox version and use the computer as firefox browser mostly and it goes slower than doing the same with Daedalus, so I think it's the OS that is slower. The question is if it is normal to have more loss of speed from Devuan 4 to Devuan 5 than there was from devuan 3 to devuan 4. | 21:51 |
YKaelig | It's just a matter of time | 21:52 |
YKaelig | BUT for today, IT"S ENOUGH :D | 21:52 |
YKaelig | So, have a nice day/night cu later | 21:52 |
Guest51 | fsmithred By the way, what happened to "Monospace regular" font, why does it look so odd in Daedalus (again it looks normal in Chimaera). | 21:52 |
gnarface | peace | 21:52 |
fsmithred | Guest51, I have no idea about the font. | 21:52 |
gnarface | Guest51: they probably changed the default font family | 21:53 |
gnarface | it used to be dejavu, and then i think bitstream before that? | 21:53 |
gnarface | not sure what it is now | 21:53 |
systemdlete | gnarface, you said: "i also have a file at /usr/lib/apt-cacher-ng/devuan_mirrors whose only contents are one line containing: http://deb.devuan.org" You meant /var/lib, not /usr/lib, right? | 21:57 |
systemdlete | because I found it had to be in /var/lib. Not sure about /usr/lib one. | 21:58 |
gnarface | Guest51: is it possible you're just noticing a difference in your default cpu frequency governor behavior? they changed that too i think... | 21:58 |
systemdlete | anyway, this works, thanks! | 21:59 |
gnarface | systemdlete: yes, confusing, i know, but there are indeed files at both /usr/lib/apt-cacher-ng and /var/lib/apt-cacher-ng, that was not a typo | 21:59 |
systemdlete | here is my "recipe" notes: https://pastebin.com/kE9dauyd | 21:59 |
systemdlete | I never created the /usr/lib file. | 21:59 |
systemdlete | but it was necessary to create the /var/lib file, and symlink it to the /etc/apt-cacher-ng directory | 22:00 |
gnarface | do you have other files in /usr/lib/apt-cacher-ng that came with the package? | 22:02 |
systemdlete | yes | 22:02 |
gnarface | so this is just a guess, but if it can do without that file it might be important to constrain it to particular domains | 22:02 |
systemdlete | I mean, I can go ahead and create it. But I'm not sure if it is used. | 22:02 |
systemdlete | isn't that what the /var/lib file is doing? | 22:03 |
systemdlete | idk | 22:03 |
gnarface | my concern would just be that if it can do without it but uses it optionally that it's some security measure | 22:03 |
Guest51 | gnarface I have no idea. I think CPU temp tends to be higher than in Chimaera, but I would need to perform more tests to confirm it as these days the weather is very hot around here and that could be it. | 22:03 |
Guest51 | I would like to know what tests I can perform in both operating systems (Chimaera and Daedalus) in the same computer to compare their performances and see where the lag may be coming from. | 22:03 |
Guest51 | When you boot a live-usb with the option "boot=toram" the system sometimes hangs up if you use too many resources. This has always happened with all devuan versions but I think it tends to happen more often in Daedalus. I would need to confirm again, but the general feeling I have is that something is taking up more space and/or taking more | 22:03 |
Guest51 | resources. The question is what and where does it can from? | 22:03 |
systemdlete | the daemon would not start without the file in /var/lib/. But it starts ok without the one in /usr/lib. | 22:04 |
fsmithred | Guest51, if you have a multiboot live-usb, you need to use toram=filesystem.squashfs | 22:04 |
systemdlete | Let me see if this actually cacches daedalus files now. | 22:04 |
fsmithred | if you prepare the usb with dd or cat, then it acts like CD/DVD | 22:04 |
Guest51 | fsmithred multiboot? | 22:05 |
fsmithred | yeah, multiple live systems on a single usb stick. | 22:05 |
fsmithred | you choose the one you want at boot. | 22:05 |
fsmithred | anything other than isohybrid imaged directly to the whole device will try to read the whole stick into memory | 22:06 |
Guest51 | fsmithred oh, no, no. I use a different pendrive for each devuan iso. | 22:06 |
fsmithred | and you prepare it with dd or cat? | 22:06 |
Guest51 | fsmithred and nothing is written by the OS to the pendrive | 22:06 |
Guest51 | fsmithred dd has always worked perfect for me | 22:06 |
fsmithred | ok, cool | 22:07 |
fsmithred | then just plain toram will work | 22:07 |
Guest51 | fsmithred yes, toram works well. | 22:07 |
fsmithred | I'm looking at the package list in mousepad, booted toram on a T420. All looks normal to me. | 22:08 |
catdotjs | hello everyone. is there any tutorial specifically for migrating from deb bookworm to Devuan Daedalus? I found one for bullseye but I am scared that it might not work with bookworm | 22:09 |
Guest51 | fsmithred, but the T420 being quite faster than the X200 is probably not working so much at full capacity to really notice if v5 is slower than v4. When the computer is fast enough that is not noticeable. | 22:09 |
systemdlete | gnarface, still not cacheing daedalus pkgs | 22:10 |
gnarface | catdotjs: it's probably the same or not much different | 22:10 |
catdotjs | alright then | 22:10 |
catdotjs | thanks | 22:10 |
systemdlete | I'll try adding that file to /usr/lib | 22:10 |
Guest51 | fsmithred and that is our main concern, that this every new version being slower and slower gets to a point where we can no longer use our old computers. | 22:10 |
fsmithred | Guest51, I can sympathize. I just recently replaced devuan-jessie on ASUS EEE netbook. | 22:11 |
fsmithred | I could not use daedalus xfce on it. Too slow. | 22:11 |
fsmithred | I ended up installing a partial lxqt desktop that uses only about 200MB for the desktop | 22:12 |
Guest51 | fsmithred, how many gigabytes of ram did yo have? | 22:12 |
fsmithred | 2 | 22:12 |
gnarface | systemdlete: i'm thinking more about it now, and i'm thinking maybe it had something to do with the initial request url being a dns round-robin | 22:12 |
fsmithred | I think it's got a dual-core atom cpu | 22:12 |
Guest51 | fsmithred I think T420 is i5. X200 is core2duo. These thinkpads have got intel cpus | 22:13 |
fsmithred | yes, I have i5 | 22:14 |
Guest51 | fsmithred it is odd that chimaera was so slow that you could not use it and it works fine in an x200. Maybe you did not have 8gb ram or you were using it from a pendrive and not with the toram option. | 22:14 |
fsmithred | if you want to try the lxqt system that I put on the netbook, I'll give you the link for the iso. | 22:14 |
fsmithred | chimaera was not slow. daedalus with xfce on the netbook was slow | 22:15 |
fsmithred | with 2G ram. | 22:15 |
Guest51 | fsmithred What information would that give us? To know if it is slow because of the xwindows desktop? | 22:15 |
Guest51 | fsmithred ok, 2gb ram is not enough. | 22:16 |
fsmithred | that would let you know if you wanted to switch from xfce to lxqt | 22:16 |
systemdlete | gnarface, but even with that file in /usr/lib, the file in /etc/at | 22:16 |
fsmithred | or a window manager | 22:16 |
systemdlete | in /etc/apt-cacher-ng is linked to one of the two. | 22:16 |
systemdlete | symlinked | 22:16 |
systemdlete | it is going to read the same info. | 22:17 |
fsmithred | maybe some services could be turned off | 22:17 |
fsmithred | does 'top' show you anything that's using up the cpu? | 22:17 |
Guest51 | fsmithred I don't know, I think there is a xfwm4 that was not present in chimaera but it doesnt seem to be taking a lot of resources | 22:20 |
gnarface | systemdlete: well, maybe this line is just a reference to both files... maybe you don't actually need to use both, but i thought i remembered needing to for some reason: "Remap-devrep: file:devuan_mirrors /merged ; file:backends_devuan # Devuan Archives" | 22:22 |
gnarface | systemdlete: but like i said, i was just following the examples in place too | 22:22 |
gnarface | i might be forgetting something | 22:22 |
Guest51 | fsmithred Give me some homework. Tell me what tests can I perform in the next few days or weeks both in daedalus and chimaera to compare one with the other and see some useful information to know why one is slower than the other. | 22:24 |
gnarface | Guest51: "cat /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor" on both versions, see if they're different, i'd check that first. then maybe i'd examine both top and iotop (and maybe jnettop too if the network was in use) | 22:26 |
systemdlete | well, anyway it is working now. I was looking in the wrong place. | 22:27 |
gnarface | systemdlete: oh, good. | 22:28 |
systemdlete | it is under "devrep" not "deb.devuan.org" | 22:28 |
systemdlete | which makes sense, that's the name we gave it | 22:29 |
gnarface | ah! yea, i was considering asking you about the structure of your cache directly next | 22:29 |
gnarface | directory* | 22:29 |
* systemdlete sees we are all clumsy fingers today... :) | 22:29 | |
systemdlete | thank you SOOOOOO much, again! | 22:30 |
gnarface | no problem | 22:30 |
systemdlete | and the webui tool is showing hits now | 22:32 |
systemdlete | now for the BIG test... deboostrap | 22:32 |
systemdlete | debootstrap | 22:32 |
* systemdlete FINGERS! | 22:32 | |
systemdlete | I need to pass http_proxy= to debootstrap? | 22:49 |
systemdlete | https://askubuntu.com/questions/958795/how-to-use-the-mirror-option-of-apt-cacher-ng-for-debootstrap | 22:49 |
systemdlete | gnarface, looks like I need to set http_proxy= for debootstrap, then it starts caching | 23:17 |
systemdlete | even though I have added the proxy config to /etc/apt/conf.d some reason debootstrap still requires this. | 23:19 |
systemdlete | In a console, I can successfully get hits on the apt-cacher, just not in debootstrap without the http_proxy= variable | 23:19 |
systemdlete | I shall add this to my "recipe"... | 23:20 |
gnarface | systemdlete: are you specifying the mirror url on the debootstrap command-line too? | 23:33 |
gnarface | that might be bypassing the apt config | 23:33 |
gnarface | aiui it would use the proxy by default if you are not specifying the mirror url along with suite and target | 23:34 |
* onefang skips the lengthy walls of text that happened while I slept, coz I gotta go out soon. Anything I need to know about? Yes I read the thing gnarface pinged me about. | 23:35 | |
systemdlete | no, the last arg I pass is the target directory; no mention of a URL there. | 23:35 |
systemdlete | onefang, some of that was imporant to some of us... | 23:35 |
gnarface | systemdlete: hmm, then it defaults to using the contents of your sources.list, and i assumed the proxy as well. i may never have tested that part specifically... | 23:35 |
systemdlete | and many posts say don't use ACNG as a mirror; others say it's ok. But setting the variable right on the command to debootstrap works fine... for me, anyways. | 23:36 |
systemdlete | I took copious notes this time. | 23:36 |
systemdlete | So hopefully I won't have to pester you or others about this thing again. Other things, perhaps | 23:37 |
systemdlete | hmmmmm... a wiki page??? | 23:37 |
gnarface | systemdlete: well, so there might be other security issues but the one thing i know of is that you can use it as a arbitrary http proxy in general, just remember that if you don't configure the stuff we did for deb.devuan.org, it'll just dump your cache into the top-level of the cache directory by domain name, and in theory this could both cause breakages and be a direct attack vector | 23:38 |
gnarface | systemdlete: you're asking me if there's a wiki page to put this info on? or you're asking me for a recommendation on wiki software? as far as i know there's only the dev1galaxy forum | 23:40 |
gnarface | i don't use mine for a web cache, but i've tried it and it functions | 23:42 |
gnarface | but this specific configuration requirement makes it obvious that there's theoretical risks involved | 23:42 |
gnarface | for real you'd probably want to use squid | 23:43 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!