| eno | hola | 00:55 |
|---|---|---|
| neachdainn | I'm creating a disk image for an Ubuntu 22.04 machine and I'm trying to use chroot to prep the image to autoconnect to a Wi-Fi network after the image is flashed. I've figured out how to do that prep using `nmcli` from a live machine but that doesn't work in the chroot because NetworkManager isn't running. Is there either a way around that or documentation on what configuration files need to exist | 01:15 |
| neachdainn | for the autoconnect to work? | 01:15 |
| neachdainn | E.g., is copying `/etc/NetworkManager/system-connections/my-wifi.nmconnection` sufficient? | 01:18 |
| erizo74 | wwww | 01:50 |
| === ajfriesen1645 is now known as ajfriesen164 | ||
| === linux is now known as Guest3085 | ||
| Metro42 | i was looking for a n64 emu for my samsung galaxy. | 05:08 |
| Metro42 | anybody know any good nmapping sites? | 06:01 |
| === rvalue- is now known as rvalue | ||
| redro | hey guys | 07:31 |
| redro | hey | 07:32 |
| yazmyrat | hi | 08:13 |
| do_you_know_it_b | ok we got it | 08:51 |
| do_you_know_it_b | :>> | 08:51 |
| ajpo | Hi, | 08:58 |
| do_you_know_it_b | hello i love you | 09:00 |
| ajpott40 | Hi, me and several colleagues have problems with wi-fi disconnecting every 15 seconds or so on Ubuntu 24.04. We did not have this problem on Ubuntu 20.04. We disabled powersaving in NetworkManager and tried various different kernel versions (including mainline). This happens on multiple different computer models from fresh install (albeit they have | 09:11 |
| ajpott40 | the same wi-fi chip). I also tried disabling Ipv6, changing Wi-Fi band and channel. Should we file a bug somewhere? | 09:11 |
| mgedmin | I would absolutely file a bug | 10:17 |
| === blueyello is now known as blueyellow | ||
| omero | slightly offtopic question: does anyone knows about a pastebin-like that could work like termbin.com via commandline (with netcat/socat) | 11:15 |
| omero | seems termbin.com went offline :| | 11:15 |
| mgedmin | sudo apt install pastebinit; echo stuff | pastebinit | 11:16 |
| mgedmin | pastebinit supports several sites, e.g. bpa.st | 11:16 |
| omero | thank you mgedmin | 11:18 |
| mgedmin | oh hey I see this channel's /topic recommends bpa.st as the pastebin of choice | 11:18 |
| omero | will check that one | 11:18 |
| omero | yes, noticed .. but thought it would only work via a browser | 11:18 |
| oerheks | termbin.com still down"? | 11:19 |
| omero | in past always used something like 'echo test | socat - TCP4:termbin.com:9999' | 11:21 |
| omero | bummer, yes oerheks .. and no clues why ... dns record is there but nothing responds | 11:21 |
| omero | was handy to share commands output on the fly .. no browsers and tricky copy/paste | 11:24 |
| tomreyn | fwiw, termbin source code is at https://github.com/solusipse/fiche - and i think a regular here hosted a copy, too. | 11:33 |
| ravage | It's me! | 11:33 |
| omero | yes, did check github last night .. main problem being i dont have a static ip else would host it myself | 11:34 |
| ravage | You can use ddns | 11:34 |
| === neocharles is now known as Guest3743 | ||
| ravage | But a VPS is really cheap these days | 11:35 |
| omero | i suppose some kind of storage would be necessary too .. meaning: huge hdd | 11:35 |
| ravage | For some text? | 11:35 |
| * mgedmin pays under $4/month for a cheap OVH VPS that rarely gets set on fire | 11:35 | |
| omero | not every countries are alike mgedmin :Þ | 11:36 |
| omero | here providers dont even consider providing ... ipv6 | 11:36 |
| omero | which is silly | 11:37 |
| ravage | I think I clean up my pastes once a month or so | 11:37 |
| ravage | I don't think they take more than 1MB of space | 11:38 |
| ravage | It's not like that paste service is meant as a permanent storage | 11:38 |
| omero | oh? mmm ... i would use it just for debugging stuff and sharing some minor stuff | 11:39 |
| omero | will have a better look at the github tree (fiche) | 11:46 |
| === mativo_ is now known as mativo | ||
| === SherlockCat is now known as spinningCat | ||
| ravage | omero: https://git.buechner.me/nbuechner/fiche supports IPv6 and i added patches of some forks of the original github project | 13:03 |
| ravage | (if you are still here) | 13:03 |
| POPEYE85 | if I copy (move) a file from one mounted partition to another (or from one folder on mounted partition to another folder on the same partition) as "usar a", is it enough to run sync command as the same "user a" or I must run sync as root? | 13:15 |
| pragmaticenigma | POPEYE85: what purpose are you needing to manually call sync? | 13:17 |
| POPEYE85 | to prevent copied or moved data loss if there is power failure | 13:19 |
| POPEYE85 | for example | 13:19 |
| BluesKaj | Hi all | 13:19 |
| pragmaticenigma | POPEYE85: unless you live in a very power limited region, there is no reason to manually call sync. The system does that on its own schedule | 13:20 |
| ogra | POPEYE85, sync operates in-kernel, so you dont need sudo or anything, it just triggers the sync/fsync syscalls and the kernel does the rest | 13:26 |
| ogra | POPEYE85, for more info see: info '(coreutils) sync invocation' | 13:27 |
| POPEYE85 | running sync makes something according to drive light. But in many manuals is written to run sudo sync to sync everything | 13:28 |
| ogra | it shouldnt make any difference to use sudo here | 13:28 |
| pragmaticenigma | POPEYE85: you should read up on what the command does and how it works. The kernel regularly calls sync when it determines it needs to. All data is written to the drive. Sync ensures anything remaining in caches (rare in modern computers and journaling file systems) is ensure to be flushed to the storage media | 13:30 |
| ogra | you can call sync with arguments to have it only sync specific filesystems (i.e. the target device you copied to) or even files ... | 13:32 |
| pragmaticenigma | POPEYE85: all the sync command does is request the kernel perform it's sync function. It does not actually happen when a user enters the command. The light blinking is coincidental, and not reflective of being triggered by your action. | 13:32 |
| ogra | aqlso note that calling sync is part of the umount command, if you unmount the filesystem you copied to after copying the sync will happen in that step automatically, you don't need to call it separately | 13:34 |
| POPEYE85 | I am sure that blinking the light on the drive that has it's own light (right after copying / moving files on it), at the same moment when the sync command is running (the terminal command line is inaccessible while command running) is not coincidental | 13:41 |
| POPEYE85 | <ogra> I know about umount | 13:42 |
| POPEYE85 | But I prefer to sync after every change of the user files to minimize the risk of data loss | 13:44 |
| pragmaticenigma | POPEYE85: the light blinks because the driver controller was accessed, it does not reflect data being written to or read from the device | 13:45 |
| POPEYE85 | For example in one manual is written: "If you run the sync command without any arguments, it synchronizes all cached data for the current user to permanent storage" | 13:47 |
| POPEYE85 | and: "Using the sync command without arguments will synchronize the current file system. To synchronize all mounted file systems, execute the command with sudo" | 13:47 |
| leftyfb | POPEYE85: you asked for help about an application (sync) and were told your use case does not cover it's purpose. You can listen to the advice or continue wasting your efforts. It's your call. | 13:48 |
| ahmed | hello | 14:59 |
| === ahmed is now known as Guest8697 | ||
| === neocharles is now known as Guest9927 | ||
| tomreyn | termbin is back btw. | 16:54 |
| mdmbkr | I updated to 24.04.02 LTS (from 22) and now one of my nvme's is reporting a 4k sector size .. anyone seen something like that | 16:57 |
| leftyfb | mdmbkr: is there something wrong with that? | 16:57 |
| mdmbkr | 'nvme list' shows it as an 8TB drive, but it's a 1TB drive (4k is 8x 512 bytes) | 16:57 |
| mdmbkr | yes, it should be 512 bytes | 16:58 |
| leftyfb | mdmbkr: why? | 16:59 |
| mdmbkr | I failed to make that clear, the drive (like the other 3 nvme's in this system) has a 512 byte sector size ; after the update, the system thinks it's got a 4k sector size, which now makes the system see it as an 8TB drive | 16:59 |
| mdmbkr | why should it be 512 bytes? because that's how it was operating before, and now it doesn't work | 17:00 |
| leftyfb | mdmbkr: what is the model# of the drive? Modern drives use 4k | 17:00 |
| mdmbkr | WDS100T3X0C-00SJG0 | 17:00 |
| leftyfb | mdmbkr: I'm not sure how that's possible. Changing sector size destroys all the data | 17:02 |
| mdmbkr | all I've figured out so far is that it was working fine for about a year, then I updated to 24.x and it immediately stopped working, and 'nvme list' claims it's grown from 1TB to 8TB | 17:04 |
| leftyfb | mdmbkr: which kernel are you running? | 17:07 |
| mdmbkr | 6.8.0-53-generic | 17:07 |
| leftyfb | mdmbkr: try the 6.11 hwe kernel | 17:07 |
| tomreyn | cna you also discuss how "it is not working" now? | 17:09 |
| tomreyn | is there an error message, does it fail to boot, to mount file systems, can you still read the partition table? | 17:10 |
| tomreyn | is the system fully updated? is it using third party apt repositories? | 17:10 |
| mdmbkr | sure .. there are four nvme's in this system, 3 of them are on a pci-e x16 carrier designed to hold 4 drives, and one of them is on the motherboard | 17:11 |
| mdmbkr | I'd have to open the case to verify but I'd bet the one that seems abnormal would be the one on the motherboard | 17:12 |
| mdmbkr | the reason it's not on the carrier is because this bios won't do 4x4 bifurcation, only x8+2x4 | 17:12 |
| mdmbkr | the nvme's were all raided together using md to make a scratch area | 17:13 |
| === todi1 is now known as todi | ||
| tomreyn | that's good contextual info. however, i don't think you answered any of the questions. | 17:14 |
| mdmbkr | yeah sorry I'm working towards that =) | 17:15 |
| mdmbkr | block nvme0n1: No UUID available providing old NGUID | 17:18 |
| mdmbkr | that looks more like a sympton than a root cause | 17:21 |
| leftyfb | mdmbkr: did you try the 6.11 hwe kernel? | 17:21 |
| mdmbkr | I'm not really sure how to do that, a quick search shows it's part of 24.04.2? | 17:22 |
| mdmbkr | (which my system says it is currently running) | 17:23 |
| leftyfb | mdmbkr: sudo apt install linux-image-generic-hwe-22.04 # then reboot | 17:23 |
| mdmbkr | E: Unable to locate package linux-image-generic-hwe-22.04 | 17:23 |
| leftyfb | oh, sorry | 17:23 |
| leftyfb | mdmbkr: sudo apt install linux-image-generic-hwe-24.04 # then reboot | 17:24 |
| mdmbkr | I guess I ought to read things first heh | 17:24 |
| tomreyn | https://gist.github.com/tomreyn/8d7675840d7bc7389b32e4d8887ca449 if you'd like to read up on this, might be for later. | 17:25 |
| mdmbkr | ok thanks for that | 17:29 |
| mdmbkr | there's not really any new hardware in this system .. again, everything was working fine prior to 'do-release-upgrade' | 17:29 |
| mdmbkr | before I reboot to try this newer kernel: Diagnostics: The currently running kernel version is not the expected kernel version 6.11.0-17-generic. | 17:30 |
| mdmbkr | seems like the various packages installed just fine, I guess that's just a notice? | 17:30 |
| leftyfb | probably because that's the latest kernel installed now | 17:31 |
| mdmbkr | ok 6.11.0-17-generic is booted | 17:35 |
| mdmbkr | that nvme still showing 8TB instead of 1TB, and 4k sector size unlike the others which are 512 bytes | 17:36 |
| mdmbkr | smartd[1472]: Device: /dev/nvme2, WDS100T3X0C-00SJG0, S/N:19117C801590, FW:102000WD, 1.00 TB | 17:43 |
| mdmbkr | smartd sees it correctly as a 1TB drive ... | 17:43 |
| tomreyn | what about fdisk -l ? which capacity does the kernel report during boot? | 17:48 |
| mdmbkr | fdisk -l shows the correct capacity, but shows all 4k sector sizes, where the other 3 nvme's are all showing 512 bytes | 17:51 |
| mdmbkr | (so maybe the capacity being wrong in nvme list is just an unrelated bug?) | 17:51 |
| mdmbkr | fdisk -l for two 1TB nvme's: https://bpa.st/7FAQ | 17:53 |
| mdmbkr | how do I see what capacity the kernel reports during boot, my SAS/SATA devices show up in dmesg that way, but not the nvme's | 17:55 |
| leftyfb | mdmbkr: you said this drive was just part of a RAID array? | 17:56 |
| mdmbkr | yes | 17:57 |
| tomreyn | mdmbkr: maybe the capacity doesn't show for nvme's during boot, i lack experience with NVMe's | 17:57 |
| leftyfb | mdmbkr: so you don't care what's on it? | 17:58 |
| mdmbkr | yes, in theory I won't really lose any sleep if it gets wiped, however, I would prefer to understand what's going on | 17:58 |
| mdmbkr | because I really doubt this one drive is actually 4k while the others are 512, the change in capacity in 'nvme list' reinforces that notion, and therefore in my imagination there's probably a config change, command option, bios setting, or something like that to get it 'fixed' back to 512 bytes | 17:59 |
| === tmroland_ is now known as tmroland | ||
| mdmbkr | but, supposing I'm wrong, and it's been 4k all along, and the capacity always showed 8TB and I just never noticed it, then perhaps an alternate theory is something during the do-release-upgrade process nuked the md superblock or something | 18:01 |
| === neocharles is now known as Guest9849 | ||
| tomreyn | does this help? https://www.bjonnh.net/article/20210721_nvme4k/ | 18:07 |
| tomreyn | make sure you remove the array member before you format, in case you want to | 18:08 |
| mdmbkr | ok checking that | 18:10 |
| mdmbkr | LBA Format 1 : Metadata Size: 0 bytes - Data Size: 4096 bytes - Relative Performance: 0x1 Better (in use) | 18:11 |
| mdmbkr | hmmm | 18:11 |
| tomreyn | and the others, which are the same model (and same firmware?)? | 18:11 |
| mdmbkr | heh, of the four nvme's in this system, three support 4k sectors, three are using 512, and one is using 4k | 18:12 |
| mdmbkr | no they're all mixed up | 18:12 |
| mdmbkr | I will paste some stuff hang on | 18:12 |
| pickanick | mdmbkr: before you tear up your installation, please note, in many cases it might be smart to have the boot/OS drive have 512byte sectors and the /home partition have 4k sectors. | 18:12 |
| leftyfb | pickanick: why? | 18:13 |
| pickanick | because the OS has lots of small files. And the user data has lots of big files. | 18:13 |
| mdmbkr | https://bpa.st/LZDA | 18:15 |
| pickanick | If you use 4k on mostly small files, more space goes wasted. | 18:15 |
| mdmbkr | it's more of a media server setup .. there's an mpt2sas controller with a bunch of drives storing media on xfs, then nvme's raided together for a scratch area, and a tape drive for backup | 18:18 |
| pickanick | hold on, does Sabrent manufacture NVMe's, or just the connector to another manufacturer's NVMe's ? because *maybe* that middle hardware is confusing the situation. | 18:19 |
| mdmbkr | that nvme has a (physical) label on it that says sabrent, looks like any other m.2 | 18:20 |
| mdmbkr | looks a lot like this: https://cdn.idealo.com/folder/Product/6556/5/6556532/s3_produktbild_max/sabrent-rocket-nvme-2tb-m-2.jpg | 18:21 |
| mdmbkr | (the drive that seems to be the issue here is the WD one though) | 18:22 |
| pickanick | but is the WD a hard drive or NVME? | 18:23 |
| tomreyn | what does it look like? https://shop.sandisk.com/en-gb/products/ssd/internal-ssd/wd-black-sn750-nvme-ssd?sku=WDS100T3X0C-00SJG0 | 18:24 |
| mdmbkr | that's exactly what it looks like | 18:26 |
| xless1 | tomreyn: what capacity are you after? | 18:29 |
| xless1 | 4TB is a pretty euro | 18:29 |
| xless1 | "553.99", i guess if you can make one under the sandisk brand, you can charge $$$ | 18:30 |
| leftyfb | xless1: tomreyn is currently helping someone with an ubuntu issue | 18:31 |
| pickanick | for serving media (ie large files) 4k sectors might be slightly faster and save memory / directory structures. | 18:31 |
| leftyfb | xless1: if you'd like to discuss storage products, feel free to /join #ubuntu-offtopic | 18:31 |
| xless1 | leftyfb: oh ok, probably a dialog from before i joined | 18:32 |
| === tmroland_ is now known as tmroland | ||
| gordonjcp | ~. | 21:16 |
| tomreyn | Connection closed. | 21:18 |
| morgan-u | I have only 3 g of swap space. should that make g chrome problems? | 22:27 |
| morgan-u | 8g ram 22.04 | 22:27 |
| morgan-u | 2.5g ram used with only hexxchat. | 22:28 |
| morgan-u | erase that | 22:29 |
| leftyfb | 2024 Nov 07 11:42:15 <leftyfb>morgan-u: add the 8GB of memory you said you have. Do not rely on swap. swap != memory | 22:30 |
| bprompt | morgan-u: hmm well, keep in mind that hexchat doesn't use that much, the 2.5gbs used in ram is mostly the system load, plus some buffered ram and cached ram | 22:35 |
| pragmaticenigma | bprompt: they have repeated the same question (as lefty indicated) since November. | 22:40 |
| === travisghansen350 is now known as travisghansen35 | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
