| Epx998 | keep getting no kernel modules on trying the 18.10 beta server image never seen that before | 00:26 |
|---|---|---|
| blackflow | Epx998: #ubuntu+1 is more suitable for that | 00:36 |
| Epx998 | yeah i keep forgetting | 00:36 |
| Epx998 | got passed the kernel error, see how it goes from here | 00:39 |
| === cpaelzer_ is now known as cpaelzer | ||
| lordievader | Good morning | 06:00 |
| cpaelzer | jamespage: coreycb: FYI bug 1789659 has a potential fix for UCA-Pike | 10:31 |
| ubottu | bug 1789659 in Ubuntu Cloud Archive " libvirt-daemon error "virHashSearch:727 : Hash operation not allowed during iteration"" [Undecided,New] https://launchpad.net/bugs/1789659 | 10:31 |
| jamespage | cpaelzer: ack - looking | 10:37 |
| rbasak | cpaelzer: welcome back! | 11:48 |
| rbasak | cpaelzer: bug 216847 came up in triage. | 11:49 |
| ubottu | bug 216847 in openssh (Ubuntu) "sshd will not start at boot if ListenAddress is set, because network interface is not yet up" [Low,Confirmed] https://launchpad.net/bugs/216847 | 11:49 |
| rbasak | cpaelzer: on network-online.target for openssh-server | 11:49 |
| rbasak | cpaelzer: following our previous discussion, is that now a Won't Fix for Ubuntu packaging? | 11:49 |
| rbasak | Or perhaps a topic for ubuntu-devel@ ML? | 11:50 |
| cpaelzer | rbasak: that is a won't fix unless upstream fixes it | 11:55 |
| cpaelzer | rbasak: this was discussed plenty of times, eventually there are two outcomes | 11:55 |
| cpaelzer | IIRC | 11:55 |
| cpaelzer | 1. the default config works always | 11:55 |
| cpaelzer | 2. if configured specially it might fail, but to implement that it would need IP_FREEBIND or netlink watching by upstream | 11:56 |
| cpaelzer | to pick up the interface late | 11:56 |
| cpaelzer | so per our discussion on the "general case" this would be Won't Fix as #1 is good and "please report upstream" for #2 | 11:56 |
| cpaelzer | but afaik that bug had an upstream report for quite some time | 11:57 |
| cpaelzer | yep it has | 11:57 |
| ahasenack | good morning | 12:10 |
| cpaelzer | hi ahasenack | 12:10 |
| ahasenack | hi cpaelzer | 12:14 |
| ahasenack | why is bind9 logging so hard :/ | 13:01 |
| blackflow | ahasenack: ? | 13:02 |
| ahasenack | it's not just add -d, or bump a log level parameter | 13:02 |
| ahasenack | one has to create channels | 13:02 |
| ahasenack | and categories | 13:03 |
| ahasenack | ugh | 13:03 |
| ahasenack | rbasak: cpaelzer: what are the next steps for the php7.3 removal, have josh comment on the bug? | 13:06 |
| blackflow | ahasenack: because it can quickly swamp all your IO if you're not careful and have a busy server. | 13:06 |
| blackflow | logging channels are great. you can separate xfr from queries from other events. | 13:07 |
| ahasenack | got any quick tips to debug this before I dig into pages and pages of documentation? | 13:08 |
| ahasenack | 12:48:59.821186 IP 10.0.3.1.50271 > 10.0.3.1.53: 54829+ [1au] A? fakehost.maas. (54) | 13:08 |
| ahasenack | 12:48:59.821319 IP 10.0.3.1.53 > 10.0.3.1.50271: 54829 Refused$ 0/0/0 (31) | 13:08 |
| ahasenack | querylog doesn't even show this query | 13:08 |
| ahasenack | I want to know why it was refused | 13:08 |
| blackflow | can you pastebin the zone? | 13:09 |
| ahasenack | no | 13:09 |
| ahasenack | the zone has just two A records | 13:09 |
| ahasenack | the bug is that I added a new interface to the host (10.0.3.1, virbr1) | 13:09 |
| ahasenack | but bind9, even while it acks the new nic, and says (logs) it's listening on it | 13:09 |
| ahasenack | doesn't answer queries on it | 13:09 |
| ahasenack | unless I restart it | 13:09 |
| ahasenack | netstat confirms it's listening on the new nic | 13:10 |
| cpaelzer | ahasenack: lets talk on standup about the removal | 13:10 |
| cpaelzer | I'm busy in a meeting atm - sorry | 13:10 |
| ahasenack | cpaelzer: ok, it's just that you and robie are eod by then | 13:10 |
| rbasak | ahasenack: looks like Steve's already done it? | 13:11 |
| blackflow | ahasenack: if you got refused, it's not networking issue but security or zone config issue. eg, there's no master zone defined and the server is forbidden from recursing (to your client IP), you'll get REFUSED. also try either raising severity to debug, or run bind with -d | 13:11 |
| ahasenack | blackflow: that's what I'm looking for, debugging logs | 13:12 |
| ahasenack | and this query was from localhost | 13:12 |
| cpaelzer | ahasenack: today I don't get my EOD like forever :-) | 13:12 |
| ahasenack | rbasak: indeed, looks like I hadn't subscribed to the bug | 13:13 |
| cpaelzer | well then, done :-) | 13:13 |
| ahasenack | cpaelzer: rbasak thanks :) | 13:13 |
| ahasenack | blackflow: running with -d is the equivalent of calling "rndc trace <level>", or just "rndc trace" to bump the debug level. The command works, but nothing is logged, and I figure it's because there are no "channels" or "categories" defined in named.conf.* | 13:14 |
| ahasenack | so, pages and pages it is | 13:14 |
| ahasenack | https://kb.isc.org/docs/aa-01526 "sample config", wow | 13:15 |
| blackflow | ahasenack: well yes you have to enable a channel for queries and uh... tbh, not sure which other chan, to see reasons for REFUSE. But I'm pretty sure if you get REFUSE, it's a security issue. by default named is not allowing recursive at all, | 13:15 |
| ahasenack | since a restart of named fixes it, it does look like a bug in the code that observes new interfaces and attaches to them | 13:16 |
| blackflow | needn't be a bug. named has ACLs so a different iface/ip address could've been unallowed | 13:17 |
| blackflow | btw 10.0.0.0/8 is not considered part of "localnets" ACL | 13:18 |
| ahasenack | as I said, a restart fixes it | 13:22 |
| cpaelzer | rbasak: ahasenack: we have no final freeze yet right, do you think one of you could quickly check https://code.launchpad.net/~paelzer/ubuntu/+source/virt-manager/+git/virt-manager/+merge/356334 ? | 15:12 |
| cpaelzer | like in the next 1-2 hours in between meetings as time permits? | 15:12 |
| ahasenack | let me see how big it is | 15:13 |
| ahasenack | since we have standup now and lunch later | 15:13 |
| ahasenack | s/we/I/ | 15:13 |
| cpaelzer | this usually (tm) builds and migrates without issues | 15:13 |
| cpaelzer | sure, only as time permits ahasenack | 15:13 |
| cpaelzer | please don't squeeze your lunchtime for it | 15:13 |
| rbasak | I'll do it unless ahasenack has already started | 15:13 |
| rbasak | Looks simple. | 15:13 |
| ahasenack | rbasak: go ahead please | 15:14 |
| rbasak | ack | 15:14 |
| === coconut is now known as coconut_ | ||
| DenBeiren_ | Hi all,.. i have two lines in fstab to mount shares from one share on a synology, the other on ubuntu server,.. one works the synology, the other doesn't,..the ubuntu one | 15:26 |
| DenBeiren_ | i am missing something, but can't put my finger on it | 15:26 |
| DenBeiren_ | https://pastebin.com/ee9eNhVB | 15:28 |
| DenBeiren_ | this is the format | 15:28 |
| DenBeiren_ | cifs utils is installed | 15:28 |
| cpaelzer | DenBeiren_: I used also ,vers=3.0 in the options for these actually | 15:34 |
| cpaelzer | not sure if that helps you | 15:34 |
| cpaelzer | my case was mount on Ubuntu a share exported by the NAS | 15:35 |
| DenBeiren_ | mine is mount ubuntu on a different ubuntu | 15:36 |
| DenBeiren_ | mount nas on ubuntu works :s | 15:36 |
| ahasenack | DenBeiren_: can you mount it manually from the command line? Is it just via fstab that it doesn't work? | 15:37 |
| cpaelzer | I formerly did ubuntu<->ubuntu and it worked, I'd also investigate on the server side config how the share is exported | 15:37 |
| cpaelzer | ahasenack: is the the right guy to ask | 15:37 |
| DenBeiren_ | it seems i can't mount it manually neither | 15:39 |
| DenBeiren_ | https://pastebin.com/wBrUMJ3Y | 15:40 |
| ahasenack | DenBeiren_: check if the share is exported | 15:42 |
| ahasenack | DenBeiren_: smbclient -L <host> | 15:43 |
| ahasenack | and repeat with -U user%pass if needed | 15:43 |
| ahasenack | and run testparm on the server, it will highlight some syntax errors in smb.conf if they exist | 15:43 |
| DenBeiren_ | grmbl, need to drive the kids to basketball,.. i'll get back on this! | 15:44 |
| DenBeiren_ | bbl to read and maybe get ideas :-) | 15:44 |
| ahasenack | cheers | 15:46 |
| ahasenack | I also have to leave, lunch | 15:46 |
| plm | Hi all =D | 16:00 |
| plm | $ sudo lxc launch ubuntu:16.04/armhf arm1 | 16:14 |
| plm | Creating arm1 | 16:14 |
| plm | error: Failed container creation: | 16:14 |
| plm | - https://cloud-images.ubuntu.com/releases: Requested architecture isn't supported by this host | 16:14 |
| plm | I following this url as TJ- told me - https://cloud-images.ubuntu.com/releases: Requested architecture isn't supported by this host | 16:14 |
| plm | What is wrong? | 16:14 |
| avu | plm: Are you running this on a armhf machine? | 16:15 |
| teward | plm: what's the main machine's architecture? | 16:16 |
| plm | avu: not, in a intel machine (xx86) | 16:16 |
| teward | if you intend to run this multiarch you can't do that with LXD containers | 16:16 |
| teward | plm: I asked this similar question to the LXC/LXD team about crossarch | 16:17 |
| plm | teward: x86 (64bit) | 16:17 |
| teward | and their ultimate response was "Not doable in LXD" | 16:17 |
| teward | https://discuss.linuxcontainers.org/t/lxd-containers-on-other-architectures/2331 | 16:17 |
| avu | plm: lxc is not a VM, it's just containers, you can't run another architecture | 16:17 |
| teward | plm: if you want to run ARMHF containers, you need to run LXD on an armhf machine | 16:17 |
| plm | teward: but here say that works https://askubuntu.com/questions/816886/how-do-run-an-arm-lxd-container-on-my-intel-host#816887 | 16:18 |
| teward | plm: if you intend to do virtual architectures you need to go a full VM route with qemu-kvm-static or such which can 'emulate' other architectures (but with a performance hit) | 16:18 |
| teward | plm: that's an incorrect post | 16:18 |
| plm | avu: i will use qmu static to emulate, see? | 16:18 |
| teward | plm: https://askubuntu.com/questions/1060089/lxd-running-an-image-with-a-foreign-architecture/1060114#1060114 trumps that | 16:18 |
| plm | teward: I actually has the ubuntu 16.4 ARM running on emu on my x86 host | 16:19 |
| teward | I'm going to operate on the assumption you don't care waht stgraber, one of the ***PRIMARY DEVELOPERS OF LXD*** has to say on the matter | 16:19 |
| teward | they're the one that replied to my thread I linked to on the Linux Containers dicussion site | 16:19 |
| teward | stgraber: ping, please assist ^ | 16:20 |
| teward | stgraber: because IIRC< oyu indicated cross-arch can't be done in LXD | 16:20 |
| teward | and I think plm is being stubborn and not listening :p | 16:20 |
| plm | teward: all right | 16:20 |
| avu | plm: read what that post teward linked says about wemu-user-static | 16:20 |
| teward | plm: what you were linking to was someone using ***LXC*** which had this support | 16:20 |
| teward | LXC support is no longer valid | 16:20 |
| avu | plm: it's still not a VM | 16:20 |
| teward | and LXC support is ***NOT*** for LXD | 16:20 |
| teward | LXD does not have qemu-static support | 16:21 |
| plm | avu: I read that | 16:21 |
| teward | and it's not an actual VM | 16:21 |
| teward | qemu-kvm-static *emulates* the architecture but that has limitations | 16:21 |
| teward | if you want armhf containers, you need armhf architecture | 16:21 |
| teward | for the machine itself | 16:21 |
| teward | for proper support | 16:21 |
| plm | teward: actually I already using qemu-arm-static, without run qemu vm, just starting the chroot | 16:21 |
| teward | qemu-arm-static isn't a VM | 16:21 |
| teward | it's emulation | 16:22 |
| teward | but point still stands | 16:22 |
| teward | (it's not a VM) | 16:22 |
| plm | teward: but with chroot I can to run the services, like as mysql etc, becouse I have the same IP like as the host | 16:22 |
| teward | but that's a chroot | 16:22 |
| teward | that's *not a VM or a container* | 16:22 |
| plm | becouse this TJ- suggest me that thread (url) | 16:22 |
| teward | then TJ- is wrong | 16:22 |
| teward | because that's LXC, not LXD. | 16:22 |
| teward | and LXC is more or less "obsolete" in favor of LXD | 16:23 |
| plm | teward: is possible I run a service in my chroot where has same ip of host? | 16:23 |
| teward | plm: *maybe*? Can't guarantee it'd work properly | 16:23 |
| plm | I see about LXD and LXC | 16:23 |
| teward | nor can I suggest that being a proper approach | 16:23 |
| teward | but that begs a question: | 16:23 |
| teward | what are you running that ***needs*** armhf? | 16:23 |
| plm | teward: yes, armhf ubuntu 16.4 | 16:24 |
| teward | 90% of all executables in existence work without armhf, even some armhf-compiled ones when compiled for standard Intel infrastructure | 16:24 |
| teward | plm: no, you're missing the point of my question | 16:24 |
| teward | ***WHY*** do you need armhf Ubuntu 16.04? | 16:24 |
| teward | if you have x86/x86_64 infra | 16:24 |
| plm | teward: becouse my device target is the armv7 | 16:24 |
| teward | I don't support using chroots to run services that're web facing | 16:25 |
| teward | but you can *try* | 16:25 |
| plm | teward: how? | 16:25 |
| plm | teward: iconfig on chroot I have the host ip | 16:25 |
| teward | and with that i'm going to walk away because I'm not going to explain how to run programs in chroots to you | 16:25 |
| teward | (and because I have to fix something at work) | 16:25 |
| plm | teward: I know hoe run rpograms on chroot, just run, problem is how run a service listening in a port | 16:26 |
| plm | teward: all right. anyway, thanks for the clarify =D | 16:26 |
| teward | plm: I said you can *try* but I didn't say it'd work | 16:26 |
| blackflow | plm: can you repeat the problem? | 16:26 |
| teward | blackflow: their core problem was they wanted to containerize armhf on their x86 infra, which can't be done :p | 16:27 |
| blackflow | someone told me the other day it can (qemu based LXD) | 16:27 |
| plm | blackflow: I'm running 16.4 armhf on qemu on my x86_64bit ubuntu | 16:27 |
| teward | blackflow: it can't | 16:27 |
| blackflow | which was very much news to me | 16:27 |
| plm | blackflow: I'm running as qemu static, so I mount rootfs as chroot | 16:27 |
| teward | blackflow: from stgraber themselves on the linuxcontainer site: https://discuss.linuxcontainers.org/t/lxd-containers-on-other-architectures/2331 | 16:28 |
| teward | and stgraber is LXC/LXD ***god*** | 16:28 |
| teward | blackflow: you could in LXC but it apparently was really slow | 16:28 |
| teward | LXC and LXD are different :p | 16:28 |
| plm | blackflow: and as chroot I do everything in my ubuntu16.4 arm, but I need to run services, like as mysql, http, etc, how to do that, becouse chroot has the same ip as the host, see? | 16:28 |
| blackflow | it must be slow, qemu is doing all the ISA translations | 16:28 |
| blackflow | so see, when I read things like "qemu chroot" my mind blows. how does thta work. either you virtualize the ISA or... chroot is just a filesystem namespace. | 16:29 |
| blackflow | plm: I do have to ask, why. why arm container on x86 cpu. what's the advantage? preparing the container before you push it out to real arm hardware? | 16:29 |
| teward | blackflow: they're testing something for am armv7 target | 16:30 |
| teward | which I don't agree with | 16:31 |
| teward | (I'd test on the hardware direct) | 16:31 |
| blackflow | so why not a qemu VM in full. it's gonna be equally slow. | 16:31 |
| blackflow | anyway.... chroots are just filesystem namespaces. if you want network namespace you need to set it up. that's why a bridge and a full blown qemu based VM would work best. | 16:32 |
| blackflow | without it, yes you're sharing the same network stack as the host. | 16:32 |
| plm | hmmm | 16:43 |
| plm | blackflow: < plm> teward: becouse my device target is the armv7 | 16:44 |
| plm | blackflow: all right, chroot works very fine, but I have no services | 16:44 |
| plm | blackflow: I think that chroot (qemu-arm-static) works faster thank a full qemu vm | 16:45 |
| blackflow | chroot only turns a directory into / . nothing else. if you need services inside it, you need to start them manually. | 16:45 |
| teward | plm: the speed will be equally slow as chroot or the full qemu VM | 16:46 |
| teward | and as blackflow said you'd have to start services manually | 16:46 |
| plm | teward: hmmm, so was just a perception =D | 16:46 |
| plm | blackflow: teward start services manually, how? | 16:46 |
| plm | blackflow: I start services manually on my chroot, but it run in the same port of host | 16:46 |
| blackflow | plm: you could copy your in-chroot systemd .service units to the host side and execute with RootDirectory set to the chroot | 16:47 |
| plm | *same IP | 16:47 |
| blackflow | plm: right because a chroot is just directory namespace, nothing else. | 16:47 |
| blackflow | if you want NETWORK namespace, you have to set it up. | 16:47 |
| plm | blackflow: hmm | 16:47 |
| plm | blackflow: "plm: you could copy your in-chroot systemd .service units to the host side and execute with RootDirectory set to the chroot" I not understand very well how to do that | 16:48 |
| blackflow | you use cp to copy, and then you modify it to add RootDirectory. but still, that will run within the same network namespace. | 16:48 |
| plm | blackflow: now problem, if I can access for example a port 8080 on my guest, is fine for me, for example | 16:49 |
| blackflow | you copy from /path/to/chroot/lib/systemd/system/foo.service to /etc/systemd/system/ for example. | 16:49 |
| blackflow | plm: there is no guest. chroot is a directory namespace. | 16:50 |
| plm | blackflow: sorry, yes, just my chroot | 16:50 |
| plm | "you copy from /path/to/chroot/lib/systemd/system/foo.service to /etc/systemd/system/ for example." will to try | 16:50 |
| plm | ^ that I was unkknow how to do =D | 16:50 |
| blackflow | plm: I'd really recommend you to set up a normal qemu based VM. libvirt will set up everything for you, including networking | 16:51 |
| plm | blackflow: ubuntu boot broken after I upgraded from 14.4 to 16.4 | 16:52 |
| plm | blackflow: becouse this I'm using chroot | 16:52 |
| plm | :( | 16:52 |
| ahasenack | cpaelzer: rbasak: if any of you are still around and have a moment: https://code.launchpad.net/~ahasenack/ubuntu/+source/squid/+git/squid/+merge/356351 | 17:18 |
| ahasenack | at least to glance. ppa is still building, same for dep8 runs | 17:18 |
| rbasak | ahasenack: done | 17:38 |
| ahasenack | rbasak: cpaelzer: thanks! | 17:40 |
| ahasenack | rbasak: https://bugs.launchpad.net/ubuntu/+source/strongswan/+bug/1795813 looks like it affects cosmic as well (via code inspection: the patch isn't there), and the package is in main | 18:45 |
| ubottu | Launchpad bug 1795813 in strongswan (Ubuntu) "backport fix for SIGSEGV when using mysql plugin" [Undecided,New] | 18:45 |
| DenBeiren_ | ahasenack: does this output help? https://pastebin.com/G4xPFFft | 19:02 |
| ahasenack | DenBeiren_: yeah, what is the share you are trying to mount again? | 19:02 |
| ahasenack | downloads? | 19:02 |
| DenBeiren_ | /home/downloads on 192.168.6.11 (skinner) | 19:03 |
| ahasenack | so what do you get when you try? mount //192.168.6.11/downloads /mnt -o username=<theuseryouhave> | 19:06 |
| ahasenack | (or some other target directory, doesn't have to be /mnt) | 19:06 |
| DenBeiren_ | found it! | 19:09 |
| DenBeiren_ | you can't use /home/downloads | 19:09 |
| DenBeiren_ | need to mount /downloads | 19:09 |
| DenBeiren_ | probably since you log in as a user, the /home is obsolete | 19:09 |
| ahasenack | ok | 19:10 |
| TJ- | DenBeiren_: no; it s because the share name defined on the 'server' is just "downloads" - it isn't the absolute path on the host itself. | 19:12 |
| DenBeiren_ | aha | 19:14 |
| rbasak | ahasenack: good point, thanks | 19:35 |
| rbasak | ahasenack: oh, I think I remember my thinking | 19:35 |
| rbasak | ahasenack: I wanted a test case first | 19:35 |
| rbasak | (in the hope the reporter can help with that | 19:36 |
| rbasak | ) | 19:36 |
| ahasenack | yeah, the test case might be tricky | 19:52 |
| ahasenack | rbasak: as they say it only happens under load | 19:52 |
| ahasenack | that, coupled with the fact there is no new release yet with the fix, made me skip it for cosmic inclusion | 19:53 |
| rbasak | ahasenack: perhaps I was hasty to tag it server-next - pending best effort reproduction steps from the reporter. | 20:04 |
| ahasenack | rbasak: we might never get such steps, at best a "it crashed before, now it works" from a site where the bug happens | 20:05 |
| rbasak | ahasenack: steps to configure it so that it may crash would still be useful I think - that would demonstrate what the configuration is at least so that others not familiar with the package could at least attempt to reproduce. | 20:06 |
| rbasak | Even if others can't actually reproduce it because of whatever race. | 20:07 |
| ahasenack | rbasak: strongswan (ipsec) with mysql, how hard could that be ;) | 20:43 |
| teward | ahasenack: ever hear of "death by a thousand cuts"? | 20:53 |
| teward | :P | 20:53 |
| ahasenack | this would be like a "quake axe murdered" scenario :) | 20:55 |
| ahasenack | cuts indeed | 20:55 |
| plm | TJ-: that url https://askubuntu.com/questions/816886/how-do-run-an-arm-lxd-container-on-my-intel-host#816887 not works | 22:41 |
| TJ- | plm: what's wrong with it? | 22:41 |
| plm | TJ-: I tried and put results here and people tellme that that thread is not correct | 22:41 |
| plm | TJ-: do yoy have log of this channel? | 22:42 |
| plm | TJ-: http://dpaste.com/12NTBR9 | 22:45 |
| plm | TJ-: Did you see? | 22:54 |
| plm | TJ-: I would like just to fix what is the problem eith my rootfs to boot in normal qemu, could you help me? | 22:55 |
| plm | s/eith/with | 22:55 |
| plm | TJ-: to me is very fine if I can just start 16.4 booting in full VM qemu. | 23:01 |
| plm | TJ-: around? =D | 23:12 |
| TJ- | plm: sorry... Stefan's reply to tew's question on the LXD issue tracker shows it is still possible with the newer LXD. I'm testing that method here | 23:27 |
| plm | TJ-: all right =D | 23:27 |
| plm | TJ-: I will wait for you :) | 23:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!