| disposable2 | in 20.04, am i supposed to start using nftables instead of iptables? i am pretty sure libvirt relies on iptables. | 06:30 |
|---|---|---|
| disposable2 | for context - i would like to switch to nftables. just asking if virtualisation is ready for it. | 06:31 |
| cpaelzer | disposable2: libvirt still depends on iptables | 06:45 |
| cpaelzer | disposable2: the last bigger eval I've seen was around libvirt 5.1 (about a1.5y ago) and there nftables didn't support all features needed yet | 06:45 |
| cpaelzer | but iptables is done using nftables these days (yeah - complexity) so since this v5.1 libvirt can deal with that | 06:46 |
| cpaelzer | disposable2: thre are a few words about is in https://www.libvirt.org/news.html | 06:47 |
| cpaelzer | but this particular change was mostly for firewalld+nftables of which the former isn't enabled anyway | 06:48 |
| disposable2 | apart from libvirt, is there any other significant software in 20.04 that hasn't made a full transition to nftables? (i know lxd4.0 is fine with nftables) | 06:49 |
| disposable2 | cpaelzer: thank you for the detailed answer | 06:49 |
| cpaelzer | disposable2: reverse-depends --release=focal src:iptables is still rather long, but I'm unsure to what extend those might be iptabels|nftables kind of dependencies without looking further | 06:53 |
| cpaelzer | disposable2: looking through the list ufw, firewalld and docker.io to name a few | 06:54 |
| cpaelzer | interestingly systemd also depends on libip4tc2 which is part of iptables as well - not sure what it is used for thou | 06:56 |
| disposable2 | cpaelzer: this is weird, firewalld is supposed to be able to use nft since version 0.6.0. but thank you. this reverse-depends is a truly magical command | 06:57 |
| disposable2 | i would've normally used aptitude for that | 06:59 |
| cpaelzer | "aptitude why" can help why hings are on a local system, but I like the above for a global overview | 07:02 |
| danshearer | There was a grub bug introduced in a security update yesterday, or thereabouts | 08:48 |
| danshearer | I looked through the irc logs for #ubuntu-server, #ubuntu-devel and others and didn't see reference to it | 08:49 |
| danshearer | I expect the right people are on to it, but here is a report | 08:49 |
| danshearer | https://askubuntu.com/questions/1263125/how-to-fix-a-grub-boot-error-symbol-grub-calloc-not-found | 08:49 |
| danshearer | It probably affects all ubuntu and ubuntu spins; certainly server, desktop and mint. | 08:51 |
| cpaelzer | danshearer: that is at https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/1889509 | 11:13 |
| cpaelzer | and yes people are on it | 11:13 |
| ubottu | Launchpad bug 1889509 in grub2 (Ubuntu) "grub boot error : "symbol 'grub_calloc' not found" [Undecided,Confirmed] | 11:13 |
| oerheks | cpaelzer, actually i expected the message to reboot after the grub2 update, good thing i have not rebooted yet. | 11:20 |
| === Wryhder is now known as Lucas_Gray | ||
| === compdoc_ is now known as compdoc | ||
| rcj | xnox: I have a question on your grub MP because I'm not familiar with the install_devices_empty case where we still break | 16:27 |
| rcj | https://code.launchpad.net/~xnox/grub/+git/grub/+merge/388383 | 16:27 |
| xnox | rcj: install_devices_empty should never be reached, as it cc_groub_dpkg always sets a value. | 16:34 |
| xnox | rcj: install_devices_empty means "i chose not have package installed, but not actually install grub anywhere" which is fine non-interactively. | 16:34 |
| xnox | bah | 16:34 |
| xnox | rcj: install_devices_empty means "i chose _to_ have package installed, but not actually install grub anywhere" which is fine non-interactively. | 16:34 |
| xnox | because i want to have grub-install available to "rescue" other drives for example. | 16:35 |
| xnox | rcj: which ami did you boot on xenial to test this? i will have xenial built with that patch soon, and we can test things. | 16:35 |
| rcj | xnox: thanks for the explanation for the patch. I booted Xenial in AWS (us-west-2 ami-060d1be0dd4526759 built on 20200611) | 16:54 |
| rcj | and I booted an m5.large or m5a.large instance so I have an nvme root | 16:55 |
| xnox | ack | 16:56 |
| === tds7 is now known as tds | ||
| aloini | I have been having some rather unfortunate issues with NFS crashing or hanging at the Kernel level, and am not sure how to further debug if: https://paste.ubuntu.com/p/3QXRD7DDNF/ - Would anyone be able to provide further support? | 18:57 |
| aloini | Well, that was unexpected... | 19:01 |
| aloini | So just to re-ask... I have been having some rather unfortunate issues with NFS crashing or hanging at the Kernel level, and am not sure how to further debug if: https://paste.ubuntu.com/p/3QXRD7DDNF/ - Would anyone be able to provide further support? | 19:01 |
| RoyK | I just wonder which module is tainted here | 19:02 |
| RoyK | https://www.kernel.org/doc/html/latest/admin-guide/tainted-kernels.html | 19:04 |
| RoyK | the kernel or some module is non-gpl | 19:04 |
| aloini | Well, I am running ZFS and that is predominantly where the NFS shares are being used. | 19:05 |
| aloini | So that is likely a cause | 19:06 |
| RoyK | probably | 19:06 |
| RoyK | but that hang isn't from ZFS as far as I can see | 19:06 |
| RoyK | whih kernel/distro is this? | 19:07 |
| aloini | Ubuntu 20.04, 5.4.0-42-generic | 19:07 |
| matthias_arch | Hello, I hope this channel can answer my DPDK question. I want to attach a dpdk pdump process to my running dpdk application, however I get the following error: EAL: Cannot initialize tailq: RTE_FIB (full error: https://bpa.st/DKYQ) I'm calling rte_pdump_init() after rte_eal_init() as done in the testpmd app. I get the same error when modifying the | 19:10 |
| matthias_arch | l2fwd example for pdump. | 19:10 |
| aloini | I am using a FUSE filesystem shared out via NFS as well, that could be it? But afaik FUSE is a part of the kernel now, and isn't non-standard. | 19:10 |
| sarnold | aloini: oh that feels pretty plausible to me | 19:11 |
| sarnold | that'd be worth testing separately from zfs if you cant | 19:11 |
| sarnold | s/t$// | 19:11 |
| aloini | I can easily stop sharing the zfs nfs exports. Nothing should be using them, they were mainly backups in case something else wasn't working. | 19:12 |
| aloini | Problem is that it happens randomly, usually after 24 hours. | 19:12 |
| aloini | You asked for the full trace RoyK: https://paste.ubuntu.com/p/RY6RDM2gH3/ | 19:28 |
| RoyK | aloini: no idea, but it looks like it's related to unicode or something | 19:42 |
| RoyK | aloini: erm - is this zfs-fuse? | 19:42 |
| aloini | No | 19:45 |
| aloini | rclone/mergerfs FUSE | 19:45 |
| === Napsterbater is now known as Guest52776 | ||
| === Napsterbater_ is now known as Napsterbater | ||
| EmberCrest | Hey all! I have an HP ProLiant G7. It's connected to the net via Ethernet Port 1. | 21:23 |
| EmberCrest | However, I can't connect to this machine (nor from it) until I run this command: | 21:24 |
| EmberCrest | sudo dhclient -r eno0; sudo dhclient eno0 | 21:24 |
| EmberCrest | Not sure why the DHCP client isn't loading. but is there any log, or configuration that I should check as to why my connection isn't working after every startup? | 21:24 |
| EmberCrest | (Not sure that it isn't loading/starting, but that's my best guess based on the solution) | 21:25 |
| tomreyn | EmberCrest: you didn't state which ubuntu release you're running, but all the supported ones other than 16.04 have journalctl -b | 21:35 |
| tomreyn | on xenial (16.04) you'd look at syslog instead | 21:35 |
| tomreyn | where configurations ar likle ystored also depends on your ubuntu release | 21:36 |
| mwhudson | tomreyn: xenial was the first lts with system fwiw | 21:43 |
| mwhudson | +d | 21:44 |
| tomreyn | mwhudson: oh right, just not persistent journal, but journal | 21:56 |
| mwhudson | ah yes | 21:56 |
| tomreyn | shoudn't matter here, though | 21:57 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!