libera/#devuan/ Friday, 2020-05-08

Human_G33kthe website certificate is outdated or its just me ?04:14
tuxd3vexpires on : 31 July 202004:16
tuxd3valmost there but not so :)04:16
Human_G33kfor me it says outaded 01/03/2020.04:18
tuxd3vhumm, you are talking about what domain?04:20
Human_G33kyou change the server ip ?04:21
Human_G33ki use a custom host file fore my network tests (when no dns for ie)04:22
Human_G33kwith the good ip it s look good04:22
Human_G33kbut website can also access with the old one ip04:22
tuxd3vI am not sure but here I looked into
Human_G33kme too now04:23
Human_G33kold one was
tuxd3vcheck your date :)04:24
Human_G33kmy date is good as i say i have a static entry in my hosts file04:24
tuxd3vit could be that its outdated in some server, I believe its round drobin :)04:24
tuxd3vdrobin -> robin04:25
ulleti think i'm too retarded to build devuan for VIM3 tuxd3v :/11:28
ulletbut i might be able to live convert debian to devuan again11:28
ulletthat worked on jetson nano11:29
ulletwhere did .dtbs come from anyway.   who keeps inventing this new stuff we don't need.11:30
ShorTiethey are part of the kernel12:58
ulletsince when12:59
ShorTieSorry, have no idea on exact date12:59
ShorTiei know of them in 4.9.y, for arm atleast13:00
ullet~ 200913:01
ShorTieyou just need a kernel13:17
ShorTiea compiled kernel would be the nicest13:17
ShorTiei use the raspberry pi arm64 compiled kernel with Devuan13:18
ulletnot saying it's a bad idea. just a new thing that means i can't upgrade kernel13:18
ShorTiewhy you say that ??13:19
ShorTieit is an extra step when compiling a kernel to make them13:20
ShorTiethey do not just automagically appear13:21
ulletwell i have no .dtb for jetson nano13:23
ulletnot for newer kernel13:23
ulleti am stuck with nvidia's13:23
ulletand i have no .dtb for khadas vim313:24
ShorTieguess nvidia does use/need them, but arm does13:32
ShorTiei'd guess you just need a working kernel && u-boot, the rest is just a debootstrap13:33
ShorTiedebootstrap doesn't do the kernel13:34
ShorTieso you can add any you want13:34
ShorTieif you can grab someones deb, it makes it easy13:35
ShorTielooks like a nifty board, you got 1 yet ??13:42
ulletyes ShorTie14:54
onefangSomething is wrong.  Seriously wrong.  My computers have been up for over two months.19:16
onefangI suspect Debian isn't providing kernel updates for Stretch anymore, except for the backported 4.19.  So my ASCII installs are not getting rebooted when there is a new kernel.19:17
onefangAnyone tried the backported linux-image-amd64 4.19+105+deb10u3~bpo9+1 kernel on ASCII?  Any problems?19:18
gnarfaceit might be time to consider beowulf if you need continued kernel updates.  i don't think that stretch is getting less attention is a surprise to anyone.  i haven't tried building it myself though.19:22
gnarfaceit might work but there's some question as to whether it's worth the risk19:23
gnarfaceno wait, scratch that19:23
gnarfaceit will probably work19:23
gnarfacei just remembered i did build 5.2 or 5.3 for ascii on arm6419:23
gnarfacebut there's still a question as to whether it's worth it19:23
gnarfacei guess the key info is i didn't build it on ascii.  i built it on ceres and then tested it on ascii, beowulf and ceres.19:28
gnarfaceso that should work but you're still off the rails at that point19:29
gnarface(for arm64 there's no other choice though)19:29
gnarface(note arm64 as in aarch64, not amd64, to be clear)19:30
onefangSince this involves some servers, one of which is a Devuan mirror, as well as my main desktop, I tend to be conservative and stick with what's very stable.  I wasn't planning on upgrading to Beowulf until later this year, after lots of testing.  Think I'll stick with that plan.19:46
onefangJust the lack of updated kernels got noticable.19:47
gnarfaceyea, it's oldstable though so that's fairly expected19:50
gnarfaceand it should be considered a risk, but if it's just a mirror then maybe your attack surface is low enough for it not to matter19:51
onefangNot just a mirror.  I run two virtual worlds on it, and a secure chat server for local activists.19:52
gnarfaceer hmm.  i'm changing my advice then.  build the new kernels.19:53
gnarfacehopefully at least your hardware is uniform enough that you don't have to build a different kernel for every server19:54
onefangAlas a mixture of Intel and AMD.19:55
gnarfacethough that alone might not be enough for you to need different kernels19:55
gnarfacei don't know what you have over there exactly19:55
gnarfaceon the plus side you can probably make the kernels significantly smaller in the process19:57
onefangI don't really want the extra maintenance burden of compiling my own kernels.  I sometimes do that for a living, but then people are paying me to.  B-)19:58
gnarfacewell, this too can be automated...20:03
gnarfacei don't have any better advice sorry.  life is full of touch choices.  in theory someone else might be providing 3rd party kernels that would fit the bill but i really wouldn't recommend going that way either.20:07
gnarface*tough choices20:07
gnarfaceonefang: for whatever it's worth i have a beowulf amd64 system here and it has been up for weeks with no problems so far.  seems completely stable as long as i don't try to sleep it.20:12
gnarfacethe "beta-ness" seems mostly to do with the upgrade process, but i didn't experience anything worse than a typical ceres upgrade20:13
onefangYep, lots of people with stable Beowulfs running, which is why it'll likely actually get fully released any day / week now.  Only a few issues to sort out I believe.20:13
ulleti find sid/beowulf caused me less problems than older ones20:15
_abc_slightly defocus but... a smart guy's comments on systemd in several articles, search systemd in page
_abc_I wonder if Chris Siebenmann is aware of devuan.org20:31
_abc_To me he sounds like someone we would like to have on board here. At least I would.20:32
tuxd3vdoes anyone notice the level of compression of xz in beowulf, or its just me?21:42
tuxd3vmaybe improvments compared with the version in ascii21:43
gnarfacetuxd3v: didn't notice but also wasn't paying attention... you sure it's not just a change to the default?22:10
gnarfacetuxd3v: (default compression level mentioned in the man page for xz is -6 but i note that tar can inherit a different value for that through an environment variable)22:12
tuxd3vgnarface, it could be, yes 6 is a value greatly used for compression but I was astinished by a xz compression done on beowulf..usually they tended to be around ~400MB is ascii, and I got a ~250MB compressed file.22:20
tuxd3vastinished -> astonished22:20

Generated by 2.17.0 by Marius Gedminas - find it at!