djph | Oh this afternoon is gonna be interesting ... assuming the backups test out as okay, gonna be rebuilding a storage box, and throwing Chimaera on it (what could possibly go wrong, amirite :) ) | 16:23 |
---|---|---|
gnarface | you probably won't have issues with a headless install | 16:25 |
djph | Yeah, I'm hoping that's the case | 16:26 |
djph | Have done in-place updates since Jessie (on Beowulf right now), wanted to kick the tires on Chimaera | 16:27 |
brocashelm | ceres ftw | 16:40 |
* yeti waits for vesta and stories about the battle of vesta then... | 16:50 | |
Katje | in devuan, how to I get nftables to load on boot ? | 17:33 |
* gnarface usually just calls a script from /etc/rc.local | 17:36 | |
Katje | http://dev1galaxy.org/viewtopic.php?id=2889 | 17:38 |
Katje | this was what I went with | 17:38 |
Katje | specifically | 17:38 |
Katje | # cp /usr/share/doc/nftables/examples/sysvinit/nftables.init /etc/init.d | 17:38 |
Katje | # update-rc.d nftables defaults | 17:38 |
Katje | with a chmod in there | 17:38 |
gnarface | oh they bundle an example? good to know | 17:38 |
Katje | that was my reaction | 17:39 |
gnarface | i would have suggested using an init.d script if i knew they gave you an example as a starting point | 17:39 |
gnarface | the /etc/rc.local file is just a placeholder that gets called last in the chain | 17:40 |
gnarface | the init.d script is better if you set it to start at the exact right time | 17:40 |
Katje | well it didn't start | 17:40 |
gnarface | hmm | 17:41 |
gnarface | needs symlinks | 17:41 |
gnarface | and maybe a proper LSB header (i didn't look in it) | 17:41 |
gnarface | it needs to be executable too, it's not clear to me from here if the example file already is by default (i would expect not though) | 17:42 |
Katje | /etc/init.d/nftables start works | 17:42 |
Katje | so it's something about boot that means it doesn' | 17:42 |
Katje | t | 17:42 |
gnarface | probably the symlinks then | 17:42 |
gnarface | ls -l /etc/rc?.d/*nft* | 17:43 |
Katje | lrwxrwxrwx 1 root root 18 Jun 2 17:38 /etc/rc6.d/K01nftables -> ../init.d/nftables | 17:43 |
gnarface | just that one? | 17:43 |
Katje | no, it's there for 0- 6 | 17:43 |
gnarface | K means "kill" | 17:43 |
Katje | lrwxrwxrwx 1 root root 18 Jun 2 17:38 /etc/rc2.d/K01nftables -> ../init.d/nftables | 17:44 |
gnarface | change them to start with S instead | 17:44 |
gnarface | K for kill, S for start | 17:44 |
Katje | for all of them ? | 17:44 |
gnarface | yes | 17:44 |
gnarface | well | 17:44 |
gnarface | just in the runlevels you care about | 17:44 |
gnarface | you should set them all so you don't get unexpected behavior but it probably is only ever gonna look at #2 | 17:44 |
Katje | reboot time | 17:45 |
Katje | nope, didn't do it | 17:47 |
gnarface | hmmm | 17:47 |
gnarface | what's the permissions on /etc/init.d/nftables | 17:47 |
gnarface | ? | 17:47 |
Katje | lrwxrwxrwx 1 root root 18 Jun 2 17:38 /etc/rc2.d/S01nftables -> ../init.d/nftables | 17:48 |
gnarface | no that's the symlink | 17:48 |
Katje | -rwxr-xr-x 1 root root 2582 Jun 2 17:37 ../init.d/nftables | 17:48 |
gnarface | hmm, that should be fine though | 17:48 |
gnarface | change all the symlinks to S01nftables instead of K01nftables, just in case | 17:49 |
gnarface | but maybe the example file has to be edited too | 17:49 |
Katje | I did | 17:49 |
gnarface | i can't imagine what it would be other than the LSB headers though if running "/etc/init.d/nftables start" works though | 17:49 |
Katje | what do the LSB headers look like | 17:50 |
Katje | # Default-Start: | 17:50 |
Katje | # Default-Stop: 0 1 2 3 4 5 6 | 17:50 |
gnarface | they look like double-commented comments, but they're actually parsed | 17:50 |
gnarface | yea | 17:50 |
gnarface | those | 17:50 |
Katje | wait, should I change this ? | 17:50 |
gnarface | yea you might need to change it | 17:50 |
gnarface | paste it to paste.debian.net so i can see it, i might be able to answer questions about it better then | 17:51 |
Katje | have set it to 0-6 again | 17:51 |
gnarface | they're not all the same | 17:51 |
Katje | just rebooted, once it's back | 17:51 |
gnarface | oh, if you got it working then nevermind | 17:51 |
Katje | we'll find out if it comes up | 17:51 |
gnarface | but note that there is actually dependency settings below that line, and it is possible to tell it an impossible situation that will result in it never booting | 17:51 |
Katje | yeah, I realised that... | 17:52 |
gnarface | like to depend on $time if you're not running ntpd or something like that | 17:52 |
Katje | fingers crossed | 17:52 |
gnarface | adding a line to call a script from /etc/rc.local shares none of these concerns, which is why i recommended that first | 17:53 |
Katje | didn't work | 17:53 |
gnarface | but this is the "right way" to do it | 17:53 |
gnarface | paste the LSB whole header block at paste.debian.net and i'll see if anything looks obviously whack to me | 17:53 |
Katje | https://pastebin.com/ibFnXMDz | 17:54 |
gnarface | paste.debian.net please | 17:54 |
gnarface | (no ads) | 17:54 |
gnarface | or you know what? | 17:54 |
gnarface | you can just /msg it to me | 17:54 |
Katje | http://paste.debian.net/1199770/ | 17:54 |
gnarface | ok | 17:55 |
gnarface | you're running a syslog, right? | 17:55 |
Katje | yes | 17:55 |
gnarface | like syslog-ng or rsyslogd? | 17:55 |
gnarface | ok | 17:55 |
gnarface | default-start and default-stop shouldn't be mirrored | 17:55 |
gnarface | they're mutually exclusive | 17:56 |
gnarface | try that, and if it still doesn't work, remove $remote_Fs | 17:56 |
gnarface | $remote_fs | 17:56 |
gnarface | but i think $remote_fs isn't hurting anything, just not 100% sure. | 17:56 |
Katje | mdadm uses: | 17:57 |
Katje | # Default-Start: 2 3 4 5 | 17:57 |
Katje | # Default-Stop: 0 1 6 | 17:57 |
Katje | # Short-Description: MD monitoring daemon | 17:57 |
gnarface | it shouldn't be possible to even see this file without $local_fs and it's for $network so running it without that would be nonsensical... | 17:57 |
Katje | shall I copy those numbrs ? | 17:57 |
gnarface | yea that looks good | 17:57 |
gnarface | otherwise it would have started then immediately stopped in every single runlevel... | 17:58 |
Katje | rbooting... | 17:58 |
gnarface | if it still doesn't work, check the paths to binaries in the actual script | 17:58 |
Katje | nope... | 17:59 |
gnarface | i'm running out of ideas | 18:00 |
Katje | [ -x "$BIN" ] || exit 0 | 18:01 |
gnarface | but the one last thing would be that maybe the script is inheriting something from your user's environment that makes it work, that isn't present at boot | 18:01 |
Katje | BIN=/usr/sbin/nft | 18:01 |
Katje | /usr/sbin/nft | 18:01 |
Katje | /usr/sbin/nft: no command specified | 18:01 |
Katje | echo $? | 18:01 |
Katje | 1 | 18:02 |
Katje | I wonder if it is that line | 18:02 |
Katje | but no, that can't be, as /etc/init.d/nftables start works | 18:02 |
gnarface | the script doesn't refer to any lock files or anything in /etc/default/ or anything like that does it? | 18:03 |
gnarface | it is just an example script and if it was expected to work out of the box they would have packaged it in /etc/init.d/ | 18:03 |
Katje | E486: Pattern not found: default | 18:03 |
gnarface | i can't imagine what it's doing wrong at this point | 18:04 |
gnarface | but it's probably something simple | 18:04 |
gnarface | do you have sysv-rc-conf installed? | 18:04 |
Katje | CONF=/etc/nftables.conf | 18:04 |
gnarface | i'm curious whether it thinks this is on or off | 18:04 |
Katje | ls -al /etc/nftables.conf | 18:05 |
Katje | -rwxr-xr-x 1 root root 1144 Jun 2 17:26 /etc/nftables.conf | 18:05 |
Katje | dpkg -l | grep sysv-rc-conf | 18:05 |
Katje | no | 18:05 |
Katje | should I | 18:05 |
Katje | ? | 18:05 |
gnarface | yea try it | 18:05 |
Katje | installed | 18:06 |
gnarface | i think it's just for manipulating symlinks but maybe we can use it to get a sanity check | 18:06 |
Katje | so I have X's in columns 1,2,3,4,5,6 but not S | 18:07 |
gnarface | just make them match something already installed that works right | 18:07 |
gnarface | i keep forgetting where the devaun documentation about this is, but i had always referred to https://wiki.debian.org/LSBInitScripts before | 18:08 |
Katje | and reboot... | 18:08 |
onefang | sysv-rc-conf is for manipulating the symlinks that determine which services run when, but can also start and stop them manually with the + and - keys. | 18:08 |
Katje | and no... | 18:10 |
Katje | I can start it manually. | 18:11 |
Katje | but nothing else | 18:11 |
Katje | and no mention in the logs | 18:11 |
gnarface | does your network not start automatically at boot? | 18:11 |
Katje | it does | 18:12 |
gnarface | hmmm | 18:12 |
gnarface | is the boot shell dash? | 18:13 |
Katje | where do I check that? | 18:13 |
Katje | it's a basic install of beowulf | 18:13 |
gnarface | uh | 18:13 |
gnarface | dpkg-reconfigure something... | 18:13 |
gnarface | i think | 18:13 |
gnarface | i'm ddrawing a blank | 18:13 |
Katje | dash is installed | 18:14 |
gnarface | ls -l /bin/sh | 18:14 |
gnarface | oh | 18:14 |
gnarface | i think it's just "dpkg-reconfigure dash" maybe | 18:15 |
gnarface | but it would be symlinked from /bin/sh | 18:15 |
Katje | done | 18:15 |
gnarface | oh damn | 18:16 |
gnarface | does this change have to be flushed into the initrd? | 18:16 |
Katje | oh | 18:16 |
Katje | fuck | 18:16 |
Katje | good question | 18:16 |
gnarface | well, try the dash thing first | 18:17 |
Katje | done | 18:17 |
Katje | BINGO! | 18:18 |
gnarface | ok, so there must be some bashism in that example script | 18:18 |
gnarface | and that's probably why it's in the doc examples instead of in the /etc/init.d file | 18:18 |
gnarface | *directory | 18:18 |
Katje | aah | 18:18 |
schillingklaus | to hell with those creepin bashisms | 18:19 |
Katje | quite | 18:19 |
* gnarface is opposed to dash | 18:20 | |
Katje | now, why is it not setting the gateway right... | 18:21 |
gnarface | to be clear, there may be a number of reasons why that example script isn't considered done... i'd recommend combing over it carefully, but i haven't even seen it myself | 18:23 |
Katje | right, let's go stick this in the other room and connect it to the internet! | 18:26 |
gnarface | good luck | 18:26 |
schillingklaus | bash has become insanely bloated since the last decade, so I switched to mksh | 18:27 |
Katje | it works!!! | 18:42 |
Katje | Thank you for your help gnarface | 19:09 |
gnarface | no problem | 19:09 |
DashiePie | I have a question | 20:42 |
gnarface | just ask it | 20:43 |
DashiePie | okay, so I have the Beowulf 3.1.1 iso, full 3.6 GB boot usb, and there's no EFI executable | 20:43 |
gnarface | missing package | 20:44 |
DashiePie | this is the install iso | 20:44 |
DashiePie | that I grabbed directly from files.devuan.org | 20:44 |
gnarface | uh, maybe you need to enable it in the "load additional installer components" section? | 20:44 |
DashiePie | that would be great, if the computer I was trying to install it on even loaded the installer | 20:44 |
gnarface | oh it can't even boot the image? | 20:45 |
gnarface | hmmm | 20:45 |
DashiePie | no, it can't boot the installer, it always defaults to win10 | 20:45 |
gnarface | you sure it's not a bios setting? | 20:45 |
DashiePie | this computer is new, it's an office pc that was bought not a month ago or so | 20:46 |
DashiePie | temporarily until I get my new hardware | 20:46 |
DashiePie | and I can't find shit in the BIOS | 20:46 |
DashiePie | it's UEFI all the way, and this iso just won't boot | 20:47 |
gnarface | which iso exactly? the netinstall one? | 20:47 |
DashiePie | 3.1.1, it's the 4GB one, which is rounded up in size | 20:47 |
DashiePie | when I said full install, I meant FULL install | 20:47 |
DashiePie | I'm not letting win10 touch the net in any capacity, so this computer's gonna need an offline install | 20:48 |
gnarface | there should be a hotkey you can hit at system POST time to intercept the boot drive choice | 20:48 |
gnarface | it's probably F2 or ESC but check the hardware manual | 20:49 |
DashiePie | I believe I've tried that, and if it's a menu that pops up, controlled by the keyboard to select a file to boot, there's nothing to let me boot, other than windows 10 | 20:50 |
gnarface | no usb image present? | 20:50 |
gnarface | hmmm | 20:50 |
gnarface | the burn is suspect | 20:50 |
gnarface | i'd expect some error if it was trying to boot it at all | 20:50 |
gnarface | although i'm not sure it's actually trying to boot it | 20:51 |
gnarface | if it is, my assumption is it doesn't appear bootable | 20:51 |
gnarface | leading cause of that is forgetting to run "sync" after running dd, or specifying a block size other than the default | 20:51 |
gnarface | but if you have a 3.6GB USB drive and a 3.6GiB disk image to put on it, you might be running into an even more basic problem | 20:52 |
gnarface | this system BIOS... is there anything in there about "secure boot?" | 20:53 |
gnarface | i'm wondering if you have to disable secure boot | 20:53 |
gnarface | i thought you didn't have to anymore but maybe that's not with every image | 20:53 |
gnarface | or mabye it's just hardware specific | 20:53 |
gnarface | some bioses do seem to have been created to keep you from displacing windows | 20:54 |
DashiePie | I used unetbootin, I dunno if that makes any difference | 21:14 |
DashiePie | 32 GB flash drive, no space problem | 21:14 |
DashiePie | I cleared secure boot keys, no issue | 21:14 |
DashiePie | I hate to say that's one of these bioses, but it may be | 21:15 |
gnarface | it could be unetbootin... that image should have been copied to the usb directly with no need for modification | 21:15 |
DashiePie | huh | 21:15 |
DashiePie | okay, well maybe I'll just try that, then | 21:16 |
gnarface | dd or cp should work | 21:16 |
gnarface | just make sure you don't overwrite the wrong drive | 21:16 |
DashiePie | graphical drag and drop isn't a problem, is it? | 21:16 |
gnarface | i actually don't know | 21:16 |
gnarface | it is important it gets written raw instead of copied to the filesystem on it, which i assume would be the default action for graphical drag&drop, but that's based on a number of assumptions i can't test here | 21:17 |
gnarface | my assumption is that graphical drag and drop is a problem | 21:17 |
DashiePie | okay, yeah, that seems to be what my friend is sayign as well | 21:17 |
gnarface | but regular "cp" is smart enough to raw write to /dev/ nodes by default | 21:18 |
rwp | I'll just say that normally one images the ISO image directly to the USB. It contains it's own partition table so it goes to the raw device and not to a partition of the USB. | 21:30 |
DashiePie | yeah, dd writes both bios and efi partitions, which was my problem | 21:31 |
rwp | Although we often use dd for this there is no reason that cp or cat cannot be used. Anything that does a verbatim image copy is okay. | 21:31 |
DashiePie | and probably why the installer previously booted, because I did that one from windows with rufus, I think | 21:31 |
rwp | I have always been a little suspicious of unetbootin (note I have not used it however) because the description says "creates bootable usb flash drives; it is a live usb creator" making me think that it modifies the raw image in some way. And that is not desired for the installer images. | 21:33 |
rwp | My favorite way to raw copy with a progress bar to a slow flash device uses pv for the progress bar and dd for O_DIRECT. | 21:34 |
DashiePie | not sure, but I know it didn't write an efi partition | 21:34 |
rwp | time pv image.iso | dd of=/dev/sdX obs=16M oflag=direct | 21:35 |
gnarface | no no no | 21:36 |
rwp | I know that dd now has status=progress but the progress information is not anywhere near as nice as what comes from pv. | 21:36 |
gnarface | obs=16M???? hell no | 21:36 |
rwp | yes yes yes | 21:36 |
gnarface | not to usb flash | 21:36 |
gnarface | you're just asking for corrupt writes | 21:36 |
rwp | You think it should be larger? Like 64M? That would be fine. | 21:36 |
gnarface | no it should be left at default | 21:36 |
gnarface | for maximum compatibility | 21:36 |
gnarface | there's no way it's gonna speed up a write to USB flash anyway | 21:37 |
rwp | Nothing smaller than 16M please. Because otherwise the NAND flash controller has to write and re-write repeatedly. That would be bad. Not to mention slow. | 21:37 |
rwp | Spinning disk drives traditionally had a 512 byte block size. Then came along AF advanced format drives with 4k sectors. Okay fine. | 21:37 |
gnarface | i'm telling you there's a percentage of media out there that is incompatible with this treatment, and will fail 100% of the time for an image that large | 21:37 |
gnarface | if you haven't run into it yet you've just been lucky | 21:38 |
rwp | But then NAND flash came along and they can't write small blocks like that. They can only write in big chunks. The smallest I know is 16M. | 21:38 |
gnarface | and there are no USB flash drives that will speed up from this treatment anyway | 21:38 |
gnarface | at best it will cause no speed difference | 21:38 |
gnarface | you might make it a tiny bit slower though | 21:38 |
rwp | This has nothing to do with speed. It has to do with writable block size of NAND flash. | 21:39 |
gnarface | we're not talking about NAND here | 21:39 |
gnarface | it's a USB key | 21:39 |
rwp | WAT?! USB storage uses NAND flash. | 21:39 |
gnarface | it's certainly not like the NAND in my arm handheld | 21:39 |
gnarface | it's something entirely different | 21:40 |
gnarface | much slower | 21:40 |
rwp | NAND flash is the most common type of flash used. Cheapest. Highest density. So it is mostly used everywhere. | 21:40 |
gnarface | slow enough that specifying arbitrarily large block sizes won't help | 21:40 |
omen | I'm missing one key from my keyboard and would like to make CAPSLOCK to act as this key. Any keywords to search? | 21:40 |
rwp | The alternative is NOR flash but it is more expensive and so less commonly seen. | 21:40 |
gnarface | omen: xmodmap? | 21:41 |
rwp | omen, You could use xmodmap to do this mapping. | 21:41 |
omen | ty | 21:42 |
rwp | omen, Probably create a file with "keycode 66 = " and put the keysym of the key you are missing there. Probably. | 21:42 |
rwp | Use "xev" to see what keysyms and keycodes are produced by the various keys. | 21:42 |
rwp | There is a very old program called xkeycaps that is interactive and used to graphically show and modify this information. Old. But still very useful. | 21:43 |
omen | if you happen to have key between left-shift and Z, I would be interested about it's keycode | 21:53 |
rwp | omen, That would likely be an ISO standard keyboard layout then? Unfortunately I have only ANSI layout keyboards. | 21:57 |
rwp | omen, Maybe keycode 94 maybe. | 21:59 |
rwp | 0x5E would be the hexmap for it. Maybe. Just looking at documentation... | 22:00 |
omen | I think just using "less greater less greater bar brokenbar" will work for me | 22:02 |
omen | since that is what the key does | 22:02 |
rwp | Good! I don't have an ISO keyboard to check myself. Sorry. | 22:03 |
omen | you have been great help :) | 22:04 |
rwp | So xmodmap -pke | grep brokenbar for me says "keycode 94 = less greater less greater bar brokenbar bar" so it would seem that assigning it to the keycode for the Capslock should work for you. | 22:04 |
rwp | Were you able to verify the keycode of your Caplock? | 22:04 |
omen | yeah it's 66 | 22:05 |
rwp | Therefore perhaps "xmodmap -e 'keycode 66 = less greater less greater bar brokenbar bar'" just on the command line might do it? Maybe? | 22:05 |
omen | works like a charm. thanks a ton | 22:07 |
rwp | \o/ | 22:07 |
omen | now how would I disable the original bind since it still also acts like capslock? :D | 22:08 |
rwp | Probably "clear Capslock" | 22:08 |
rwp | Sorry. Looks like it is "clear Lock". | 22:09 |
omen | perfect. thanks again | 22:10 |
rwp | omen, Do you ever work on the Linux console outside of X? If so then the commands there are dumpkeys and loadkeys. | 22:10 |
omen | not with this keyboard | 22:10 |
rwp | I do this on my machine to make Capslock into a Control key: dumpkeys --full-table | sed -n '/^keycode *29/p;/^keycode *58/p' | sed '/^keycode *29/s/Control/Escape/g;/^keycode *58/s/CtrlL_Lock/Control/g' | loadkeys -q | 22:10 |
rwp | And then with that I get my normal Control key left of the A on the Linux console too. | 22:10 |
rwp | I do it as a dump and sed because at one point I found I had different systems producing different things. So that allowed me to read-modify-write just the single key I wanted. | 22:11 |
omen | harry you are a wizard | 22:11 |
rwp | Oh, the above also changes the left Control into an Escape key. (Just like an old HP HIL keyboard that had that layout that I used to use.) | 22:12 |
rwp | And just for completeness on my system I have that dumpkeys|loadkeys in my /etc/rc.local file to happen at boot on the Linux console. | 22:13 |
omen | maybe I can one day do that also now that I have the pipe symbol at all | 22:13 |
rwp | Missing the pipe key makes working with the computer a lot harder! There is all of that cutting and pasting needed all of the time. No fun. No fun at all! | 22:14 |
rwp | If there were devuan-offtopic this would trigger a rant about C-language trigraphs which can be used for many of those special characters. | 22:14 |
rwp | Glad to hear it worked out for you. Perhaps a new keyboard is in your future? In any case I need to be afk for a bit. Back much later... | 22:15 |
omen | yeah, new keyboard is definitely on my list :) | 22:15 |
redrick | omen: This is for you! http://linuxmafia.com/pub/humour/keyboard-failure | 23:18 |
brocashelm | which packages are needed for apt-listchanges to temporarily clear the previous terminal entries and prompt you to scroll until you hit q? i was cleaning up a bunch of packages i didn't need and wasn't sure if a package that made it work normally was removed by accident | 23:21 |
brocashelm | i'm guessing it's a python package? | 23:21 |
fifihyperbola | how to register on libera?? | 23:31 |
golinux | fifihyperbola: https://libera.chat/guides | 23:33 |
fifihyperbola | tnx :) | 23:33 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!