systemdlete | remmina won't install: | 00:01 |
---|---|---|
systemdlete | The following packages have unmet dependencies: | 00:02 |
systemdlete | remmina : Depends: libssh-4 (>= 0.8.1) but 0.7.3-2+deb9u3 is to be installed | 00:02 |
systemdlete | (ascii host system) | 00:02 |
systemdlete | I tried fix-broken | 00:02 |
systemdlete | I tried fix-missing | 00:03 |
rwp | systemdlete, I know nothing but is remmina in ASCII? I might be missing seeing it there. Are you using a backport? | 00:25 |
rwp | What does "apt-cache policy remmina" say it is coming from? | 00:26 |
fsmithred | N: Unable to locate package remmina | 00:26 |
fsmithred | yeah, it's in ascii-backports | 00:29 |
fsmithred | so is libssh=0.8.1-1~bpo9+1 | 00:30 |
fsmithred | sorry... libssh-4=0.8.1-1~bpo9+1 | 00:30 |
fsmithred | systemdlete, add: -t ascii-backports | 00:31 |
fsmithred | or install <package>=<version> for those two | 00:32 |
mtnman | helo | 00:36 |
mtnman | how can i ensure that a network interface will be brought up at boot time? i have a stanza in /etc/network/interfaces for usb0. sometimes usb0 is brought up at boot, sometimes not. when usb0 is not brought up at boot, sometimes "ifup usb0" works sometimes it does not. how can cause the interface to be brought up automatically at every boot? | 00:39 |
fsmithred | auto vs. allow-hotplug? Did you try changing that? | 00:40 |
mtnman | i have it "allow" | 00:41 |
mtnman | does it need "allow-hotplug"? | 00:41 |
fsmithred | I think so | 00:42 |
mtnman | ok brb. | 00:42 |
fsmithred | that will let udev/eudev handle it | 00:42 |
fsmithred | if that doesn't work, change it to auto | 00:42 |
mtnman | sorry, i meant i have it already set to auto. | 00:43 |
fsmithred | so try allow-hotplug | 00:43 |
mtnman | ok.. | 00:43 |
fsmithred | or manual if you want to use ifup every time | 00:43 |
fsmithred | is this ascii or beowulf? | 00:43 |
mtnman | no i want it at boot because it will be running headless. | 00:44 |
mtnman | ascii | 00:44 |
mtnman | fsmithered: well i changed it to "allow-hotplug" and rebooted. usb0 is not up and "ifup usb0" fails with the following error: Cannot find device "usb0" | 00:48 |
rrq | sounds like a race between initializing the usb stack for that device and configuring it. | 00:51 |
rrq | I would suggest going back to "auto", which gves more time for the usb stacl to initialize | 00:52 |
rrq | then you might even want to add a "pre-up sleep 3" or similar to that, to extends that time window | 00:52 |
mtnman | rrq: i actually had just changed it back to "auto" and rebooted. no luck, "ifconfig -a" does not show usb0 and "ifup usb0" fails. | 00:53 |
mtnman | rrq: does "pre-up sleep 3" go in /etc/network/interfaces or in a separate file? | 00:55 |
rrq | in interfeaces, within the usb0 stanza | 00:56 |
rrq | interfaces | 00:56 |
rrq | but if it's not in /proc/net/dev it hasn't initialized well enough | 00:56 |
mtnman | ok if it goes in just after "netmask" line? | 00:57 |
rrq | yes | 00:57 |
mtnman | brb | 00:57 |
mtnman | rrq: no luck. | 00:59 |
rrq | it appears in /proc/net/dev ? | 00:59 |
mtnman | rrq: does not | 01:00 |
fsmithred | stop and start eudev | 01:00 |
rrq | ok; so the device has not initialized ... can you find it under /sys/class... somewhere? | 01:00 |
fsmithred | this sounds a little like a bug in beowulf. Not sure why it would happen here. | 01:01 |
mtnman | "find . -name usb0" yields nothing. | 01:02 |
mtnman | (from within /sys/class) | 01:03 |
mtnman | /sys/class/net shows eth0 and lo | 01:04 |
rrq | right. it hasn't initialized as net device. there should be something /sys/bus/usb/ for that device | 01:04 |
rrq | this is towards the idea of putting an explicit re-initialization attempt as a pre-up command | 01:05 |
mtnman | /sys/bus/usb/devices shows usb1 but not usb0 | 01:06 |
rrq | mmm that's probably the hub... is there an 1-* something for the device? | 01:08 |
mtnman_ | /sys/bus/usb/devices shows "1-0:1.0" and "usb1" | 01:10 |
rrq | do you know the vendor/product id ? is there something in dmesg re this? | 01:12 |
mtnman | dmesg says: "usb usb1: Manufacturer: Linux 4.9.105 musb-hcd" and "hub 1-0:1.0: USB hub found" | 01:16 |
rrq | yes that's the hub; the devices is at one of its ports .. maybe "find /sys/bus/usb/devices/usb2/ -type f" shows where it is? | 01:18 |
rrq | find /sys/bus/usb/devices/usb1/ | 01:19 |
mtnman | /sys/bus/usb/devices contains 1-0:1.0 and usb1. that is all. | 01:20 |
rrq | right .. so that device is not powered up properly .. | 01:21 |
rrq | the network device I mean | 01:21 |
rrq | that's a hardware problem | 01:22 |
mtnman | what about the network device? | 01:22 |
mtnman | oh you mean that usb0 is not powered up at boot? | 01:22 |
rrq | afaiui the network device is a usb device and it doesn't fire up properly | 01:22 |
rrq | it adds to confusion the the "usb0" label is a different namespace from "usb1" (as well as naming different things); the former is an interface name, the latter is a sys name | 01:27 |
mtnman | rrq: if you're still around, i rebooted the board with a debian image and usb0 appears at boot, so not a hw issue. | 01:32 |
rrq | mmm was that with a cold boot or the same warm reboot ? | 01:35 |
mtnman | cold | 01:35 |
rrq | ok, but beowulf still fails with cold reboot? | 01:36 |
mtnman | ascii | 01:36 |
rrq | sorry | 01:36 |
mtnman | yes cold boot of ascii image and no usb0 | 01:37 |
rrq | ok.. same kernel version ? actually what does /proc/net/dev have? | 01:38 |
mtnman | /proc/net/dev shows eth0 and lo only. | 01:40 |
mtnman | same as ifconfig -a | 01:40 |
rrq | hmm so it seems the usb net device is not powered up properly by the ascii kernel.. | 01:44 |
mtnman | rrq: yes, but that seems odd since usb0 had been functioning at one point when booting ascii. | 01:44 |
rrq | agreed | 01:45 |
mtnman | can you remind me of the module name for usb networking? | 01:46 |
mtnman | because lsmod is not showing it. | 01:47 |
rrq | probably usbnet | 01:49 |
mtnman | ok its "usbnet" (looked it up) and it was not loaded on the ascii system. i modprobed it and lsmod showed usbnet. | 01:49 |
rrq | and now you vahe usb0 ? | 01:50 |
rrq | have (need mor coffee here :)) | 01:50 |
mtnman | no. ifconfig usb0 fails. | 01:50 |
mtnman | ifconfig -a does not show usb0 | 01:50 |
mtnman | rrq: i should buy you a coffee | 01:51 |
mtnman | hmm. plugged in a pinephone to usb port and now usb is up. | 01:53 |
mtnman | well, not up, but initialized. | 01:53 |
mtnman | rrq: ok, now it is functioning again. apparently usbnet module was not loading at boot time. | 01:55 |
mtnman | add it to /etc/modules? | 01:56 |
mtnman | yes that did the trick. now usb0 comes up at boot. | 01:57 |
rrq | yes, well it needs to come into initramfs I guess | 01:58 |
rrq | that's /etc/initramfs-tools/modules | 01:58 |
mtnman | please elaborate. | 01:58 |
mtnman | hmm i added usbnet to /etc/modules. are you saying it belongs in /etc/initramfws-tools/modules instead? | 01:59 |
rrq | yes; add it there, then run update-initramfs -u -k all | 01:59 |
mtnman | well i have not ever touched that file. when i open it with vim, it looks kinda strange. looks like comment but they are not preceeded by "#" | 02:02 |
mtnman | looks like comments (plural) that is. | 02:02 |
fsmithred | yeah, those are comments, and the last two lines are examples that are commented out | 02:03 |
mtnman | nm... terminal window was obscured by two columns on the left hand side. | 02:03 |
fsmithred | works just like /etc/modules | 02:04 |
rrq | yes it has the same syntax as /etc/modules ... for modules to include with the first boot phase | 02:04 |
rrq | it'll then be loaded quite early | 02:04 |
rrq | and I suppose it won't really need the "pre-up sleep" | 02:05 |
mtnman | ok, edited /etc/initramfs-tools/modules, adding usbnet. did update-initramfs -u -k all. now rebooting. | 02:06 |
rrq | thumbs are held :) | 02:07 |
mtnman | success! | 02:07 |
* mtnman loves it when a plan comes together | 02:07 | |
mtnman | rrq: what is your location? i have never heard "thumbs held" but assume is similar to "fingers crossed" | 02:11 |
rrq | I'm a Swede living in Aus ... so "thumbs" is mostly Swedish I think :) | 02:12 |
mtnman | ok. just looked it up and it exists in english too according to urban dictionary. | 02:14 |
mtnman | rrq: Austria or Australia? | 02:15 |
rrq | Australia | 02:15 |
rrq | sunny spring day .. | 02:16 |
mtnman | nice. i have been to that hemisphere a number of times, but not that continent. | 02:16 |
mtnman | we had snow here yesterday. | 02:17 |
mtnman | thanks for the help. i owe you a coffee. | 02:18 |
mtnman | 1 | 02:21 |
rrq | mtnman: sorry I got side-tracked ... we'll share a coffe when traveling is a thing again | 03:33 |
mtnman | rrq: sounds good. do you know anything about matrix protocol? i am having a rough time connecting. | 03:39 |
rrq | nope | 03:40 |
mtnman | k thanks anyhow. bye for now... | 03:42 |
mtnman | ls | 03:59 |
mtnman | oops | 03:59 |
fluffywolf | . | 04:00 |
fluffywolf | .. | 04:00 |
fluffywolf | damnit, he quit. | 04:00 |
deva | I am trying to install libvirt on a head-less beowulf server. Apparently I need to install the package libvirt-daemon-system | 08:19 |
deva | What then baffles me is that this package is dragging in dependencies such as libav, gstreamer and a ton of other non-relevant packages | 08:20 |
deva | The full list can be seen here: https://pastebin.com/SMw2KiUD | 08:21 |
deva | Is this really as intended? Or can I do something to narrow down what should be installed? | 08:22 |
rennj | After this operation, 481 MB of additional disk space will be used. oof | 08:26 |
rennj | guess i wont complain at vmware player size | 08:27 |
rennj | kvm/qemu and libvirt is fat | 08:27 |
rennj | there is always the cli and scripts...libvirt/virtmanager is gui bonus | 08:27 |
deva | But I only need the daemon, but you are probably right though | 08:28 |
rennj | qemu-system-x86_64 -m 512 -enable-kvm -soundhw es1370 -serial `tty` -net user -net nic,model=rtl8139 -hda amiga.qcow2 -cdrom cdrom.iso -boot d | 08:28 |
rennj | oof command line | 08:28 |
rennj | qemu-img convert is also nice to know | 08:29 |
deva | This si what libvirt is running under the hood | 08:29 |
rennj | vpc vbox vmware qcoq conversion VDI's | 08:29 |
rennj | libvirglrenderer0 | 08:32 |
rennj | so its doing more | 08:32 |
rennj | all the mesa fu also | 08:32 |
rennj | you get the opengl acceleration like i have in vmware | 08:32 |
deva | I guess that makes sense then | 08:33 |
deva | I would have been nice if there was a head-less variant of it though | 08:33 |
rennj | vnc/spice or whatever | 08:33 |
rennj | qemu has that option | 08:34 |
rennj | $ qemu-system-x86_64 -vnc :0 | 08:34 |
deva | True, but I cannot prevent all the X dependencies from being installed on the server | 08:34 |
rennj | dont use libvirt/virtmanger | 08:34 |
rennj | no dependencies | 08:35 |
rennj | hehe | 08:35 |
deva | I need something to control starting/stopping of all my VMs | 08:35 |
rennj | or recompile the source | 08:35 |
rennj | fix the depends at the source | 08:35 |
deva | In ASCII it had only non-X dependecies | 08:35 |
rennj | sure you can tweak some things | 08:35 |
deva | Compiling myself is a bad idea, since I will loose the option to use distro security updates | 08:36 |
rennj | my system 360MB.iso but i use squashfs so more like 1.2GB | 08:36 |
rennj | in-memory ram os with vmwarep layer | 08:36 |
rennj | been doing poor mans esxi now for over decade i got like 8 iso's going back to like 2006 or so | 08:37 |
rennj | im happy at being under 500MB i dont think 5.4 or 5.8 kernel and vmware player 16 os going to happen | 08:38 |
rennj | lib/firmware alone is 500MB now | 08:38 |
rennj | bloat! | 08:39 |
rennj | 5.1.5 vs 5.1.11 is 20MB diff | 08:39 |
rennj | squashfs 20MB diff so more like 80MB | 08:40 |
deva | Only problem is that VMWare isn't exactly free | 08:40 |
rennj | vmware player is free for non commercial use | 08:40 |
deva | I meant free as in free software free | 08:41 |
rennj | heh i was running amd fglrx till 2019 | 08:41 |
rennj | that was closed source also | 08:41 |
rennj | besides the vmware player | 08:42 |
rennj | and my 802.11ac is outside the kernel sources | 08:42 |
deva | I usually try to use the free tools only, unless I have a really good reason not to | 08:42 |
rennj | meh..i do what i have to do..to get working hardware | 08:42 |
deva | Even if they are slightly worse than their proprietary counterparts | 08:42 |
rennj | https://pasteboard.co/JkIQb7D.png for the games! | 08:42 |
rennj | i want stuff running optimal..https://pasteboard.co/ILhcp6u.png 3GB of vram for guest vm | 08:44 |
rennj | new vmware player 16 does 8GB of vram for guest vm with 16GB of ram | 08:44 |
deva | So you use VMWare for gaming mostly? | 08:45 |
malikith | deva: To answer your question about all those dependencies that libvirt-daemon-system pulls in, there is a way to avoid some of them. You could try: apt --no-install-recommends install libvirt-daemon-system | 08:46 |
rennj | nextstep,solaris,amiga, vm's | 08:46 |
rennj | https://pasteboard.co/IwuAPBu.png 60days 4 vm's | 08:46 |
rennj | thats on my crappy 2015 hp laptop im still waiting on ryzen craptop | 08:46 |
rennj | 4 vm's in 16GB of ram, winblows and linux being the big one's compared to haiku or amiga | 08:47 |
deva | malikith, Thanks. I'll give a try | 08:47 |
rennj | 1GB of ram free | 08:47 |
deva | malikith, That seemed to do the trick. I was under the impression that the recommeneded packages wasn't installed by default? | 08:48 |
malikith | deva: Nah they are, to avoid them from the start on a fresh system I made a file at /etc/apt/apt.conf.d/01norecommends with APT::Install-Recommends "0"; inside the file. That makes it permanent | 08:51 |
deva | malikith, thanks for the tip | 08:58 |
malikith | deva: No problem. Glad I could help :-) | 09:03 |
gnarface | deva: when they're off by default everyone's like "hey why doesn't my email client make sounds anymore!?" | 09:31 |
gnarface | no matter what you do, the defaults are going to piss of someone | 09:32 |
deva | gnarface, Yeah, you're probably right | 09:34 |
deva | what confused me was the "Suggested packages" section which isn't being installed | 09:34 |
deva | So the difference between "suggested" and "recommeneded". I simply read the first few suggetsed package names and verified that I didn't see them in the install list, and (errornously) concluded that non-critical packages weren't being installed | 09:35 |
deva | I know I am probably barking up the wrong tree here; but what would be nice was if apt install would explicitly print the list of recommended packages as well so it is easier to wee which packages are coming from where in the install list | 09:37 |
gnarface | deva: yea, the choice of those two synonyms probably didn't age very well. suggested in this context just means literally "we suggest you try out these tangentially related packages too!" but recommended means "we recommend you let us stuff this down your throat now" | 09:37 |
deva | hehe | 09:37 |
gnarface | deva: oh, if you use apt-get though, it will print out suggested, recommended, upgraded, newly installed, removed, and orphaned/unneeded packages all in separate categories. that might help you too | 09:38 |
gnarface | then you can just hit "n" to cancel and alter the installation request accordingly | 09:38 |
gnarface | not sure what apt does, i don't use it bare typically | 09:39 |
gnarface | i think most of what apt does you can accomplish with a combination of apt-get and apt-cache commands instead, and they might have better verbosity | 09:39 |
* gnarface has been doing this since well before there were multiple choices | 09:40 | |
gnarface | and you're right this is the wrong tree to bark up | 09:40 |
deva | I think apt is justa front-end to apt-get and apt-cache | 09:41 |
gnarface | Debian is the right tree to bark up, but they won't take you seriously for stuff like this unless you come bearing actual patches | 09:41 |
deva | You are probably right | 09:41 |
gnarface | (or corporate funding) | 09:41 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!