av6 | hey, i've noticed that "publicly archived" link on https://www.devuan.org/os/community still points to http://reisenweber.net/irclogs/freenode/_devuan/ (i.e. freenode, not libera) | 05:01 |
---|---|---|
av6 | apparently the content is the same regardless the network name, so it's just the link that looks out-of-date | 05:02 |
adhoc | av6: might pay to poke #devuan-www for the Devuan Web presence | 05:02 |
av6 | (although it would be nice to get https there) | 05:03 |
av6 | adhoc: okay, i'll try there | 05:03 |
av6 | actually came here to say thanks for bringing /etc/init.d script back to rsyslog package, but immediately noticed that the archive link should indeed refer to libera | 05:06 |
bb|hcb | av6: the freenode redirects to libera; that is kept for not breaking bookmarks. If you check the channel topic, there is the new link | 05:13 |
av6 | i don't mind keeping the old link working | 05:22 |
av6 | but if you say the channel is on libera and link to libera, then archive link should also link to libera | 05:24 |
bb|hcb | av6: https://git.devuan.org/devuan/www.devuan.org that is kept here, why not propose a pull? | 05:27 |
rwp | That link is one of those known misleadingly labeled things. It says Freenode. But it's really Libera. | 05:33 |
golinux | av6: IIRC there were technical reasons why it was done that way. joerg can explain when he pops in. | 05:34 |
rwp | On the server side I can imagine there is a symlink making the two things actually point to the same thing. | 05:35 |
golinux | All the old freenode logs are there too | 05:37 |
bb|hcb | av6 means to change the link on the web pages to the libera one (as in topic). In the logs both links work as expected both ways | 05:41 |
av6 | yes, on www.devuan.org: https://git.devuan.org/devuan/www.devuan.org/src/branch/new-beta/source/os/community.html#L61 | 05:57 |
av6 | i might do a request there when i have time, but i assume just a quick fix from a maintainer would be a lot faster | 05:57 |
bb|hcb | Sure, but unless that is done now, it may easily be forgotten. That is why I proposed to make a PR - it act as a reminder too | 06:05 |
golinux | Good idea. | 06:08 |
golinux | But before it's changed, joerg has to sign off on it. | 06:08 |
golinux | Xenguy: ^^^ | 06:09 |
Xenguy | No idea what or where this error is. Please add some details | 06:21 |
bb|hcb | Xenguy: there is no error. Logs were kept in /freenode/ which is also avail as /libera/. The link on www points to the freenode and its better to point to the libera one (both have same content). | 06:22 |
Xenguy | Yes. Is there a change request? | 06:23 |
Xenguy | You're saying a search/replace would work, for example? | 06:23 |
bb|hcb | As joerg put that link in topic (the libera one), I'd assume it is safe to change. I do not see a PR as of now | 06:24 |
Xenguy | If the change request is clear, then it's probably easy to do | 06:25 |
Xenguy | aha, /topic | 06:25 |
Xenguy | No worries, we'll get things patched up, and hopefully ditch that 'freenode' confusion | 06:26 |
jason1234 | hello | 07:54 |
jason1234 | i am late concerning jitsi meet installation ion devuan | 07:55 |
jason1234 | https://jitsi.github.io/handbook/docs/devops-guide/devops-guide-quickstart | 07:55 |
UsL | FatPhil: I had the exact same issue, and downgrading the kernel helped. I had the two latest installed and 4.15.X worked while 4.16.X did not. | 08:19 |
joerg | :-/ | 08:44 |
joerg | signed off, fix that stuff please | 08:52 |
joerg | Xenguy: search/replace should work | 08:55 |
joerg | sed oneliner | 08:56 |
joerg | /freenode/ is a link to /libera/ | 09:02 |
FatPhil | rwp: alas this lenovo has no sysrq, and subbing in a prtsc key in its place doesn't seem to work either. In other news, the lububtu live-USB locked up hard overnight too, so maybe it's the h/w or maybe it's modern kernels.... Investigating mroe... | 13:24 |
FatPhil | UsL: I'm on 4.9.0-16, but I'll try a Beowulf upgrade today, which will put me on the same release as the g/f, whose using similar h/w (she's x250, I'm x240) and having no problems. Hopefully, that will solve more than just the one issue I'm having. I'm not ruling out thence going to Chimaera of course. | 13:30 |
FatPhil | I'm booting from a fully encrypted drive that was set up back in the old days of debian - there's no way something in that early pre-boot could be interfering with the devuan I now have, is there? | 13:34 |
FatPhil | wowsers, the ascii->beowulf upgrade is a big one. converting from debian was tiny in comparison. faster internet needed... | 13:40 |
critr | i skipped the headache and did a clean install of chimaera | 13:45 |
critr | but i keep notes of most of my little tweaks and mods, so that helps. | 13:46 |
FatPhil | I have a unified partition scheme, and don't want to lose any user data (from other ex-users) that might be hiding in forgotten corners of the filesystem | 13:47 |
FatPhil | I'm a bit surprised that I chose such a partitioning scheme, but what's done is done. I'm normally very anal about keeping different things apart. | 13:48 |
critr | you can still switch to /home on a separate partition on an installed and used system. | 13:49 |
critr | just rsync it over and fix fstab. | 13:50 |
FatPhil | that would be a good first step. I like a separate /usr/local too. However, neither solves the problem of some things in /var being the system's, and some of it being the users'. | 13:52 |
FatPhil | user stuff I've tried to keep on /usr/local/var, and then use symlinks to make it visible to things that expect stuff in /var. | 13:54 |
GyrosGeier | only /var/spool/*/* should belong to users | 13:54 |
FatPhil | What's the system doing in /var/www/html ? | 13:55 |
GyrosGeier | like /var/spool/cron/* for crontabs | 13:55 |
FatPhil | I'd forgotten about crontabs - yup, I'd have forgotten to have backed those up. | 14:07 |
FatPhil | but the fault I think was httpd's, chosing /var/www was wrongthink decades ago, and noone's corrected it. | 14:08 |
FatPhil | I can take /usr/local vs. /opt to #devuan-offtopic :D | 14:10 |
FatPhil | The influences of the dinosaurs still lives on. | 14:11 |
GyrosGeier | /var/www is system-wide | 14:27 |
GyrosGeier | user homepages live in ~user/public_html | 14:27 |
FatPhil | devuan has no business sniffing in /var/www. You're conflating the system, devuan, with the sysop, which is not devuan, and who will rightly be dabbling in /var/www | 14:28 |
GyrosGeier | indeed | 14:30 |
GyrosGeier | IIRC that's why the default page got moved out | 14:30 |
GyrosGeier | and people are now expected to change the server config if they want to serve files from /var/www | 14:31 |
djph | for ... apache? | 14:33 |
GyrosGeier | IIRC all webservers were told to stop putting the default page in /var/www | 14:35 |
GyrosGeier | because while they can technically be conffiles, that needs trickery | 14:35 |
GyrosGeier | and if they aren't conffiles, upgrades will overwrite them | 14:35 |
FatPhil | One of the problems is that even if the packages have started to not do dinosaur stuff, the old tutorials still refer to it, so people are recommended to use /var/www/ or something under that. Bad behaviour takes a long time to fix. | 14:45 |
FatPhil | At least the '.d/' one conf per site promotes a clean start with each new thing you want to host. | 14:46 |
djph | I mean, I could see the webserver hosting out of /srv more than /var ... but honestly, a default "index.html" under /var/www isn't much of a config | 14:47 |
djph | "hey it works, now go create a proper site config" | 14:47 |
FatPhil | which they copy off stack-exchange, and which refers to /var/www | 14:48 |
djph | I mean, I just stuck stuff in /var/www/mystuff/ since userdir doesn't play nice with scripts without more work (and static html under ~djph is FINE) | 14:50 |
djph | oops, where'd that capslock come from :| | 14:51 |
FatPhil | oooh, looks like some deps have been changed in beowulf, and I can finally ditch a whole bunch of stuff I don't want on my system. This is good. | 14:58 |
FatPhil | Aha! I've found a rat king! ruby was in a loop of interdependencies | 15:02 |
djph | ew | 15:11 |
FatPhil | UsL: beowulf has taken me all the way to 4.19.0-18. We'll see how it behaves... | 15:16 |
FatPhil | Well, it may have taken a long time, and a couple of apt-get -f install loops, but the upgrade seems to have worked smoothly, and every niggle was covered by the devuan.org documentation, so props for that! | 15:21 |
FatPhil | a bit worried by some kernel BUGs and WARNINGs tho' | 15:24 |
FatPhil | But alas, qemu's -host cpu still fails with the "requires KVM" error message, and again strace gives me no indication of what it was checking that failed. | 15:37 |
FatPhil | kvm and vm_intel modules both loaded, and /dev/kvm exists rw to me. | 15:38 |
FatPhil | Which means my g/f's won. She got the id-card reader that's "only supported on ubuntu" working on devuan by just bashing the ubuntu packages until they fit into her machine, and tweakign configs. I was hoping for a VM solution. | 15:39 |
GyrosGeier | CPU model passthrough might imply that nested virtualization is needed | 15:43 |
GyrosGeier | because otherwise we'd be passing through a CPU without virtualization support | 15:43 |
FatPhil | the lubuntu doesn't need virtualisation, I don't care much what it sees. I just want it to boot of my live-usb, and run X and a browser | 15:45 |
GyrosGeier | indeed | 15:49 |
GyrosGeier | but the question is whether you are asking kvm to create a CPU that is virtualization capable | 15:49 |
FatPhil | what would that "asking kvm" look like in an strace. The only instance of the string 'kvm' is at the final write to stdout before the exit. | 15:57 |
FatPhil | Having said that, I can't even get qemu running without the -cpu switch now, as I don't want VNC. | 15:58 |
FatPhil | Not sure what it was using in the past, I didn't provide a -display switch, but it "just worked" by popping up an X window. Now it doesn't do that. | 15:59 |
FatPhil | -display sdl and -display gtk fail presently, I might need to install something extra for those. The error message gives no clues. | 16:00 |
buZz | what is -display a option for? | 16:01 |
buZz | oh kvm?, no clue then :) | 16:01 |
fsmithred | FatPhil, what qemu option is supposed to let you boot a live-usb? I would love to be able to do that. | 16:05 |
FatPhil | fsmithred: just provide the whole device as the image to run, and it runs the whole grub and live boot happily | 16:09 |
FatPhil | E.g. I (used to) do qemu-system-x86_64 -m 4G /dev/sdb | 16:09 |
fsmithred | like /dev/sdX ? | 16:09 |
fsmithred | ok | 16:09 |
fsmithred | I will try | 16:09 |
FatPhil | Yup, my usb is a 'persistent live-USB' , latest lubuntu. There was a tutorial on how to prepare the key, something like 'mkusb' | 16:11 |
fsmithred | ok, so it's not just isohybrid-imaged usb? | 16:11 |
fsmithred | is it a real fat32 partition? | 16:11 |
FatPhil | IT's a complex partitioning scheme, because the 'persistent' aspect means that there are overlaid filesystems. | 16:13 |
fsmithred | how does it boot? grub or syslinux? | 16:13 |
FatPhil | -display gtk works now qemu-system-gui is installed, sdl still fails. | 16:14 |
FatPhil | lubuntu live-usbs contain their own grub | 16:14 |
FatPhil | probably this was what I used: https://www.howtogeek.com/howto/14912/create-a-persistent-bootable-ubuntu-usb-flash-drive/ | 16:15 |
FatPhil | final state of key: https://www.howtogeek.com/wp-content/uploads/2019/05/mkusb_24.png?trim=1,1&bg-color=000&pad=1,1 | 16:15 |
FatPhil | the ntfs partition is not needed, it just puts that there so that you can share files between the onboard ubuntu and a windows machine | 16:17 |
FatPhil | I'm pretty sure the complexity is because of the "persistent" aspect, which was a must-have. It might be possible to just dd and run | 16:18 |
fsmithred | ...checking qemu options, about to try it. | 16:24 |
fsmithred | qemu-system-x86_64: /dev/sdb: Could not open '/dev/sdb': Permission denied | 16:25 |
FatPhil | you're not in the right group. sudo chown it to something you're in | 16:27 |
fsmithred | works fine if I'm root. I probably should add myself to the kvm group. I'm not in it, and I'm not sure how I've been using qemu all this time without that. | 16:28 |
FatPhil | more permanently, add yourself to the disk group, or better, create a udev rule to recognise it. | 16:28 |
FatPhil | conservation of workingness is in force. qemu seems to have changed how I pass through usb devices to the guest, and the manpage gives no clues... | 16:29 |
fsmithred | brb | 16:33 |
FatPhil | the psychopathy of the qemu developers is astounding | 17:03 |
used____ | On Beowulf, I would like to add wakeup hooks to the suspend to ram wakeup script, but I can't. I tried to listen to acpid S3 wakeup events but they are not generated. Is there some magic one can use to make acpid react to events generated by entering and exiting sleep? | 17:05 |
used____ | Kernel ACPI layer reports S3 entry before and then exit at wakeup, but no event is generated for acpid? | 17:07 |
used____ | I am currently running acpid with OPTIONS="-d -l" so it all goes to syslog. | 17:10 |
FatPhil | ah, qemu no longer has the option -cpu host, and its response to you passing that option is to tell you something useless rather than "-cpu host not understood" or better "-cpu host is deprecated, use --enable-kvm [or whatever is compiled in for h/w virtualisation] instead". Or, even better, just converting that old switch into the new one. | 17:14 |
FatPhil | Anyway - now running native with kvm. Finally. | 17:14 |
used____ | With all the paravirtualization jungle it is getting hard to do a true cpuid? | 17:15 |
FatPhil | the -device usb-host is a bit of a head-scratcher - I can't get the device through using the vendor/product identifiers, which are the only things that will be constant. Using bus/addr is wrongthink. | 17:16 |
used____ | Followed a guide much? The Arch one for example? Should be good? | 17:16 |
FatPhil | there's no "true" cpu any more, you may have been migrated off to a different processor between your CPUID and relying on h/w features. | 17:17 |
used____ | https://bbs.archlinux.org/viewtopic.php?id=263568 | 17:17 |
used____ | FatPhil: "cloudy" times we are living | 17:17 |
FatPhil | I'm using https://qemu-project.gitlab.io/qemu/system/devices/usb.html , and it contains no examples, so I'm trying random shit now. I'll see what arch have to say.. | 17:18 |
used____ | https://wiki.archlinux.org/title/QEMU#Pass-through_host_USB_device | 17:18 |
FatPhil | this is another example of qemu psychopathy. In the old days (yesterday, before upgrading) I'd type "-usbdevice host,769b:1021", and that was fine. Now I do "-device usb-host,vendorid=769b,productid=1021" and ... it doesn't even parse those numbers (typos are possible, that wasn't copy paste) | 17:20 |
FatPhil | Bosh! well done arch - it seems the 0x is compulsory on the numbers | 17:21 |
FatPhil | thanks used____! | 17:21 |
used____ | Strictly, all numbers should be prefixed with the base prefix on any command line... | 17:22 |
FatPhil | I, idiotically, converted the hex numbers it refused to parse into the decimal 30363 4129 - and that still failed. | 17:22 |
FatPhil | Not strictly at all. That's what protocols are for. %04x is good enough for the linux kernel and for lsusb, it should be good enough for qemu. | 17:22 |
used____ | Heh nice try with decimals. Tried octal too? :p | 17:23 |
FatPhil | Nope, but I was also just about to suggest that chmod 420 ... is a very bad way to try and achieve 0644 | 17:24 |
FatPhil | Everyone just does 644, that's the protocol for file modes (see the LKML discussions on the wrapping of numbers with macros a decade ago if you want Linus' take on this) | 17:26 |
rwp | FatPhil, The PrtScn should work for SysRQ. To verify easiest is on text console first then get help for it. | 20:49 |
rwp | FatPhil, Control-Alt-F1 to get to the linux vt text console. Then Alt-PrtScn-h will dump some help out. If it is active. Works here. | 20:49 |
bb|hcb | On some laptops it is Alt-Fn-PrtScn-xxx... | 20:57 |
rwp | I am using a Lenovo new island style keyboard here and though it doesn't say SysRq anywhere it does have PrtScn on the lower right. | 21:16 |
* enyc meows | 23:38 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!