=== SinaCutie is now known as siina [06:21] Good morning [10:38] Hi guys. Do you think Pentium n4200 can candle 1Gb link? [11:06] mnms_: I would think so [11:18] https://pcengines.ch/apu2.htm comes with https://www.amd.com/en/product/7426 and they can do gigabit fine. your https://ark.intel.com/content/www/us/en/ark/products/95592/intel-pentium-processor-n4200-2m-cache-up-to-2-5-ghz.html looks more powerful in all aspects which would affect network data transfer. [11:19] (not counting performance degradation due to microcode updates / CPU bug counter measures) [11:32] tomreyn: Im also worried little about nic chipsets Realtek 8111G [11:57] that's a better reason to be worried [12:25] tomreyn: I see I can buy apu2c2, what workloads do you have? [12:26] cpu is heavily loaded? Do you use it as a router :)? [12:35] mnms_: i don't have such hardware now, just used to play with it a little. [12:36] but they can be used as routers, yes. [12:36] if it won't be doing tcp offload (routing) you need 1ghz per 1gbit [12:36] if it's bidirection, that is 2gbit traffic per a 1gbit port [12:37] that is only a 1.1ghz cpu [12:37] yee its not much [12:37] I would go with something with less cores and more ghz [12:37] hard to find this kind of sbc [12:37] cause drivers don't always irq balance between cores good [12:38] understand [12:39] up squared has realtek chipsets but Pentium n4200 [13:28] hi there! i am wondering if somebody has an idea: how could i "temporarily" extend the space of my /boot partition? i managed to maneuver myself in some unfortunate situation where i believe can get only out if i manage to increase the size of /boot [13:30] friendlyguy: you could copy everything to somewhere bigger and bind mount that bigger place back to /boot [13:30] friendlyguy: but perhaps easier to just move one probably-won't-be-touched thing out temporarily? [13:31] nah, i think apt wants to install dozens of old kernels, that wont fit in the 250mb partition [13:31] Why would apt want to do that? [13:32] good question :) i managed to screw it up badly i guess [13:33] i manually had to remove old kernels a bunch of times. i guess i just made an error [13:53] okay, lets c if that works :) i attached another vhdx, created a partition and mounted that. ran rsync, unmounted /boot, mounted the temporary /boot and now running apt-get install -f [13:54] might have been a good idea to create a snapshot first... but... already past that point# [14:05] okay, i think that worked. now i am wondering why apt-get autoremove didnt remove the old kernels from there? [14:28] okay... somehow things are screwed up even more: cant boot any more [14:28] grub is entering rescue mode [14:29] i think it "might" have used the uuid from the temporary disk or something [14:31] Yes you'll need to run update-initramfs manually etc. [14:31] ls in grub rescue only gives hd0, but no partitions? [14:31] You seem to have inferred far more than just my suggestion. [14:32] i moved the content of boot to another disk and mounted that as boot, isnt that what you suggested? [14:34] cant i just tell grub to use /dev/sda1? [14:34] because the content is there [14:45] Is there a kickstart file for cloud-init ubuntu image? [14:45] what a pain [14:48] okay, ls in rescue mode gives me only hd0. if i try to ls hd0 i get unknown filesystem [14:48] whats going wrong here? [14:48] sda1 only has two partitions: boot and root [14:48] so shouldnt grub display it as hd0,[1,2]? [14:55] hello? [14:59] friendlyguy: hello. [15:01] did you mean to say "sda only has two partitions: boot and root" (not sda1)? [15:01] :) thanks tomreyn, at least now i know my messages still go through. [15:02] sda has two partitions, yes. sda1 and sda2. sda1 is /boot [15:02] thats why i was trying to tell grub to use sda1 [15:03] can you sum up what you're trying to do overall, and what has already been done? [15:04] right now i try to unfuck it [15:04] i had space issues with /boot (seperate partition, 250mb total). so i asked if somebody knows a good way to "temporarly" increase the size of /boot [15:04] so you have a system which does not boot; grub loads but cannot find its configuration file and thus does not show the menu [15:05] ok, go on [15:05] its trying to boot of a disk with a uuid that is wrong... i guess [15:05] friendlyguy: generally, you can't do that, since the boot partition is generally sandwiched between the start of the disk or EFI partition and the rest [15:05] okay, going on... [15:06] its legacy boto [15:06] boot [15:06] no efi partition [15:06] friendlyguy: next time, don't use a dedicated boot partition ;) [15:06] nevertheless its sandwitched [15:06] ot was needed earlier, but that's quite some time ago [15:06] so rbasak came up with the idea to copy the content somewhere else and mount it as /boot [15:07] thats what i did: attached a second vhdx file to the vm, created partition and created fs on it, rsynced the content of boot to the temporary disk, mounted that disk [15:07] than ran a apt-get install -f to fix the "initial" problem [15:07] friendlyguy: just copy it to the rootfs and remove the boot partition - should work - some grub meddling may be needed, though [15:08] modern versions of grub can boot from large partitions - earlier it couldn't [15:08] i didnt want to move /boot permanently only temporary [15:08] you don't need it [15:08] really [15:09] what do i dont need? [15:09] friendlyguy: I did assume that you would put /boot back and weren't going to immediately reboot. [15:09] Then it would fix itself on the next update. [15:09] friendlyguy: a separate boot partition. I never use it anymore. it really isn't needed [15:09] You can run (IIRC, check the manpage) "update-initramfs -u" to fix up the initramfs images in /boot when you have restored /boot to the real one that matches everything [15:10] ah, yeah. but thats an very old system. so its in there, and it was done with setup defaults back then [15:10] -u -k all, iirc [15:10] so i guess there was a time it was a good idea === Ussat-1 is now known as Ussat [15:11] i think /boot still goes to a separate partition with default paritioning on every current ubuntu installer [15:11] okay, i now booted into the rescue system of some 16.04 iso [15:12] mounted the root, mounted the boot partition [15:13] so, only run "update-initramfs -u" now? [15:13] which ubuntu version do you have installed there? [15:13] 16.04.5 [15:13] you also need to mount virtual file systems [15:14] proc sys dev ... [15:14] isnt that being taken from root? [15:15] hmm, i don't understand what you mean [15:16] forget it. how do i mount the virtual file systems? [15:16] i'm saying that if you're going to run "update-initramfs" to update the initrd on your existing on-disk installation, you'll need to chroot into this installation and make sure the environment is similar enough to that of the booted system. [15:17] i thought thats what the rescue-system did [15:17] i took an 16.04.3 ios, booted it and selected repair a broken system (or something like that) [15:17] i then told it which is my root partition and mounted /boot [15:18] so i "guess" the rescue system chrooted me in there? [15:18] oh i see. i'm not sure how to work with this, always did it manually. [15:18] jup, same here [15:18] first try with the rescue system [15:18] https://help.ubuntu.com/community/LiveCdRecovery#Update_Failure is roughly the manal approach [15:23] so you'd mount / and any required fie systems of the standard system somewhere, say /target , then bind mount /proc /sys /dev /dev/pts onto the respective directories there and then chroot /target /bin/bash [15:35] i think i found the issue: during the replacement of /boot with the temporary location my grub.cfg was modified [15:35] the new grub config is pointing to the temporary disks uuid [15:36] so i need to fix the grub config to point to the old disk again [15:39] jup, that did the trick: the system is booting again [15:40] it would be awesome if somebody could help me to figure out why there are still so many old kernels in boot [15:42] kernel 4.4.0-154 is currently used. but there are still kernels in /boot ... 4.4.0-93, or an initrd.img... 3.13.0-70 [15:42] would be nice to remove those [15:42] i *could* remove it manually, but i`d prefer that apt cleans it [15:44] friendlyguy: try "apt --purge autoremove" - but check carefully when it prompts you [15:45] didnt do anything [15:45] I would remove them manually then. [15:45] Be careful to avoid your currently booted kernel and also the latest kernel of course [15:46] what i dont get: i have different number of files for abi, config, initrd and vmlinuz [15:46] like there are two files starting with abi, 8 or nine files starting with initrd, 3 vmlinuz,3 system.map [15:47] use "dpkg -S file" to see what package file comes from [15:52] that doesnt work for those files [15:52] Some may be generated [15:53] then they're not part of a package (or were generated during installation of a package and are not tracked) [15:53] Yes one of those two [15:53] Usually I "dpkg -l|grep linux" and manually return the ones I don't need. [15:53] Anything that's obviously a kernel version that I don't need. [15:54] wow [15:54] i think thats the main problem [15:55] to make auto-removal of kernel images (all but the latest two and the currently active one) work, you'll need to ensure that linux-generic and linux-image-generic are installed and that all the packages which contain a kernel version number as returned by dpkg -l linux-\* | awk '/^i/ {print $2}' are automatically, not manually installed. [15:55] let me show you something :) [15:57] https://pastebin.com/LJ4vicss [15:57] rbasak: i assume that command normally doesnt give you 240 lines back [15:58] the /etc/kernel/postinst.d/apt-auto-removal hook script runs whenever you install a kernel image and tracks (in /etc/apt/apt.conf.d/01autoremove-kernels) which kernel images (and initrds, headers) should be kept / removed [15:58] those 3.13 packages must be remainders from before your release upgrade to 16.04 [15:59] entirely possible === MassDebates_ is now known as MassDebates [15:59] shall i just purge linux-image-extra-3*? [15:59] Yeah that's rather large [16:00] as well as linux-image-3* [16:00] i guess you can sudo apt purge '^linux.*-3\.13\.0-.*' but make sure you check the output before you confirm [16:03] uuh:) down to 127 [16:03] is this system booted from disk now, or still chrooted in recovery? [16:03] booted from disk [16:04] which kernel is now running? [16:04] cat /proc/version or uname -r [16:04] 4.4.0-154 [16:05] good, that's not too far behind [16:06] so you can then purge all linux-* 4.4.0 packages which have a 2 digit patch level [16:06] i`m horrible with regex [16:06] :) [16:06] sudo apt purge '^linux.*-4\.4\.0-[0-9][0-9]-.*' [16:06] again, make sure you check the output before you confirm [16:07] sure [16:07] bryce: how's it going? [16:07] looking good to me, lets remove them [16:08] I usually process the dpkg -l output by hand in an editor and then feed that to dpkg -P [16:08] Saves doing regexes for a one off task and I get to confirm it's right [16:08] interesting: errors occured with linux-image-extra-4.4.0-93-generic [16:10] probably some leftover files? [16:10] ah no that'd been a warning [16:11] depmod: ERROR: could not open directory /lib/modules/4.4.0-93-generic: No such file or directory [16:11] so i guess i can ignore that [16:11] if it prevents purging the package then you probably can't ignore it [16:13] did it a second time: worked now [16:16] thats how it looks now: https://pastebin.com/1Atr59u3 [16:16] so you may want to get a fresh list of reamining installed (or leftover configuration files) linux-* packages and sort out which ones you want to keep / remove, or just mark all of them as automatically installed. [16:16] all of the versioned ones, that is. [16:17] can you take a look at the link i posted, thats what is currently installed [16:18] i did :) [16:19] looking much better to me :) [16:19] thanks for your help so far! [16:20] do you have an idea why it kept linux-headers-4.4.0-89 and 93? [16:22] manually removing them worked [16:23] nacc, heya, was away on vacation friday; will pick back up on php this afternoon or tomorrow I think [16:23] anyhow. need to get home. ill try to join back later [16:23] thanks for the help of all of you! [16:23] very kind [16:23] friendlyguy: i don't know why it kept these packages, no [16:24] i was just wondering why they were not catched by your regex [16:24] friendlyguy: you haven't prevented this issue from occurring again, yet [16:24] thought that they should have been catched [16:24] tomreyn: exactly!!! i would love to though [16:24] but i have to go home now [16:24] bryce: sounds good, thanks! [16:24] oh my regex had a dash behind those numbers [16:25] friendlyguy: see you [16:25] aaah, i c [16:25] :) [17:51] ahasenack: what is the process or is there one.. [17:51] for me to request a push of a tag for merge of https://code.launchpad.net/~rafaeldtinoco/ubuntu/+source/cloud-utils/+git/cloud-utils/+merge/370194 [17:51] i just uploaded rafael's b26a5aa7736d4e64c19f66fa563a12808b1c4ebd [17:51] as upload/0.31-0ubuntu1.1 [17:52] but i can't push there. [17:52] (pinging you because you're the only member of usd-import-team that is in current timezone normal working hours) [17:57] smoser: Andreas is at Debconf this week, so he may not be around. [18:01] fair enough [18:03] smoser: hey [18:03] * rafaeldtinoco reading [18:04] ahh its for andreas, ok, ill warn him (in front of me) [18:04] rafaeldtinoco: well, i think the ansewr is "there is no answer" [18:05] he is coming [18:05] =) [18:06] smoser: hi [18:06] smoser: there is none, let me look at that and push it for you, if there is still time to win the race [18:08] yeah, i'd guess you can win the race [18:08] as it is stuck waiting for approval [18:08] so i think you're racing a human [18:08] and they're slow [18:09] sru [18:09] ok [18:13] smoser: done [18:13] added info to the mp as well [19:36] hi. i have a 55GB database dump i need to move. I'm aware that compressing it might not be an option. should I just rsync it to the other server? [19:37] !crosspost | catbadger [19:37] catbadger: Please don't ask the same question in multiple Ubuntu channels at the same time. Many helpers are in more than one channel and it's not fair to them or the other people seeking support.