=== AceLan_ is now known as AceLan [08:44] dmsmr_, that lists the exact file and line #, 65 === Guest75017 is now known as fmasi === dhkl is now known as fruitloop === fruitloop is now known as dhkl [11:19] xnox, hey ... would you expect me to be able to upgrade running systemd at all ? [11:20] apw: mostly yes. Some packages may fail in post-inst. E.g. those that try to do "invoke-rc.d start foo" where "foo" is only an upstart job, without systemd unit / init-script. [11:20] xnox, yeah that has just happened, in the stop for the existing installed one in fact [11:21] of course i am only doing that because of debhelpers heplfulness not directly [11:21] apw: you can add a symlink of that name to /dev/null [11:21] (locally) [11:21] i clearly need to get these work units in pronto [11:21] xnox, where do they live the work units ? [11:22] apw: packages install units into /lib/systemd/system/. local administrator into /etc/systemd/system [11:22] as right now i cannot reboot, because my kernel is half installed :/ [11:23] apw: $ sudo ln -s /dev/null /etc/systemd/system/foo.service [11:23] where foo is the service that postinsts try to start/stop/whatever [11:24] and $ systemctl reload [11:24] then try to dpkg --configure -a, or whatever [11:24] ok that worked, just adding the links and doing the upgrade again [11:24] \o/ [11:24] i'll get some work load units in here double quick, so i don't have this issue next time [11:25] they shold be about 3 lines === ara is now known as Guest77001 === fmasi is now known as Guest61073 === Guest61073 is now known as fmasi [14:57] bjf, gimme some love: https://lists.ubuntu.com/archives/kernel-team/2014-June/043914.html [14:59] rtg, looking [15:27] tyhicks: hello, i noticed that you worked on bug 1296459, which is fix released. I'm on apparmor version 2.8.95~2430-0ubuntu5 on trusty and seeing similar behaviour [15:27] bug 1296459 in apparmor (Ubuntu) "Upgrade from 2.8.0-0ubuntu38 to 2.8.95~2430-0ubuntu2 breaks LXC containers" [Critical,Fix released] https://launchpad.net/bugs/1296459 [15:29] psivaa: can you paste the results of [15:29] dmesg | grep DENEIED [15:29] psivaa: that should be DENIED [15:30] yeah [15:30] tyhicks: jjohansen1: oops sorry, that gives me empty list [15:30] psivaa: what is the similar behavior that you're seeing? [15:31] psivaa: okay do that on syslog [15:31] grep DENIED /var/log/syslog [15:31] I would assume that the container is failing to start [15:31] tyhicks: "lxc-start: command get_cgroup failed to receive response" when i run "sudo lxc-start -d -f /etc/lxc/default.conf --name juju-precise-template" [15:32] psivaa: on the grep did you fix my typo? [15:32] jjohansen1: yea i did and there is nothing on dmesg, but on syslog: [15:32] Jun 12 14:25:36 parapuva kernel: [415185.749263] type=1400 audit(1402579536.554:201): apparmor="DENIED" operation="mount" info="failed type match" error=-13 profile="/usr/bin/lxc-start" name="/" pid=4172 comm="lxc-start" flags="rw, rslave" [15:32] Jun 12 16:07:42 parapuva kernel: [421316.922238] type=1400 audit(1402585662.502:218): apparmor="DENIED" operation="mount" info="failed type match" error=-13 profile="/usr/bin/lxc-start" name="/" pid=8191 comm="lxc-start" flags="rw, rslave" [15:33] though the timings are not matching when i ran the lxc-start last though [15:34] and i did a restart at 16:10, so these errors are before the restart [15:36] should i ping stgraber then for a workaround? [15:37] psivaa: you weren't running with pre-release packages by any chance, were you? (someone recently hit a similar bug but was using the buggy apparmor package from the devel cycle) [15:38] jjohansen1: as a side note, it would be good to add the fstype to that audit message [15:38] psivaa: what is your uptime like? [15:38] psivaa: can you run lxc-start without the -d so you can actually see what's the error message? [15:38] tyhicks: indeed [15:39] tyhicks: i'm not running pre-release apparmor. the lxc is also from stable release (1.0.3-0ubuntu3) [15:40] jjohansen1: "16:39:58 up 28 min, 2 users, load average: 0.27, 0.28, 0.31" is the uptime [15:40] psivaa: so apparmor is 2.8.95~2430-0ubuntu5 ? [15:41] tyhicks: yes [15:41] psivaa: okay, that rules out stale policy in the kernel but not the cache [15:42] stgraber: the first line is "lxc-start: failed to attach 'vethU534EI' to the bridge 'lxcbr0' : No such device" [15:42] psivaa: can you do [15:42] sudo /etc/init.d/apparmor restart [15:42] to make sure that the policy is recompiled with the most recent apparmor [15:42] psivaa: ok, do you have a lxcbr0 device on that host? [15:42] stgraber: checking [15:45] stgraber: "sudo brctl show" shows nothing [15:45] should *I do anything about it :) [15:45] ? [15:46] psivaa: sudo status lxc-net [15:46] stgraber: lxc-net start/running [15:47] weird... can you pastebin /var/log/upstart/lxc-net.log? [15:48] stgraber: "dnsmasq: failed to create listening socket for 10.0.3.1: Address already in use" is the only line. [15:48] please bear in mind, this is the first time i am starting to use lxc so, i could be doing something silly [15:48] psivaa: this isn't feeling like bug #1296459, so I'm going to let you chase down the bridge device issue and then you can ping me if apparmor is still causing problems [15:48] bug 1296459 in apparmor (Ubuntu) "Upgrade from 2.8.0-0ubuntu38 to 2.8.95~2430-0ubuntu2 breaks LXC containers" [Critical,Fix released] https://launchpad.net/bugs/1296459 [15:49] tyhicks: ack, thanks and sorry for the noice [15:49] np [15:53] psivaa: hmm, do you have bind9, dnsmasq or some other dns servers running on that box? [15:55] stgraber: no http://paste.ubuntu.com/7634310/ is the grep '10.0.3.1' /var/log/syslog output [15:58] sforshee, are we ready for '-9.ucode for 3160, 7260 and 7265.' in Trusty ? [15:58] I think we're carrying the beacon filter patch in 3.13 [15:59] psivaa: you've got bind installed on that system [15:59] psivaa: (named) [16:00] rtg: not sure that trusty will even use it, but there's a patch coming down from 3.13 stable that we should get before we update from -7 I think [16:00] it's what Emmanuel thinks will get around the problems people saw with -8 on 7260 [16:00] sforshee, the Utopic LTS kernel will use it, right ? [16:01] rtg: yeah, but let me double check trusty too [16:01] psivaa: so I'd suggest you remove the bind9 package from your system then reboot, that should do it [16:02] psivaa: unless you actually need named for some reason, but if that's a standard desktop or server system, you don't. You pretty much only do if you are running a DNS server on there. [16:02] sforshee, if that patch is already in stable, then I can cherry-pick it and add the buglink [16:03] rtg: it was in the most recent set that kamal sent out for review [16:03] sforshee, do you recall the commit subject ? [16:03] sforshee, rtg which patch? [16:03] rtg: so trusty kernel will only use -8 right now, so I guess no harm in adding -9 [16:03] rtg: one sec [16:04] stgraber: ack, will remove it and rerun. thanks for the help [16:04] rtg: iwlwifi: mvm: disable beacon filtering [16:05] 7bacc782270ff7db3b9f29fa5d24ad2ee1e8e81d [16:05] sforshee, rtg: yup, that'll be in the 3.13.11.3 that I will release tomorrow [16:06] rtg: after that we _should_ be able to move back to -8 ucode, but we've actually only had one tester confirm that it helped [16:06] rtg, fyi, the sha in the 3.13-stable branch for that is f30f889c iwlwifi: mvm: disable beacon filtering [16:07] kamal, I'm gonna cherry-pick it in advance and add a buglink [16:07] rtg, wfm [16:08] rtg: bug number is 1293569 in case you don't have it [16:08] sforshee, thanks [16:09] rtg: there's another patch there that emmanuel suggested adding to 3.13, but I haven't had time to follow up yet [16:09] rtg: 12d423e816c69b0b4457bc047dda9a0a1c1a53c1 iwlwifi: mvm: Add a missed beacons threshold [16:09] kamal: ^ [16:10] sforshee, not marked for stable [16:11] sforshee, I'll stand by on that one until somebody (you, or him) to ask me to stuff it into 3.13 [16:11] s/to ask me/asks me/ [16:12] kamal: ack [16:12] sforshee, it does apply fine, btw [16:13] kamal: cool, thanks. My hesitation is that it seems to be more of a "this might help too" than a "this will definitely help" kind of suggestion [16:14] stgraber: tyhicks: now, after removing bind9, i see the apparmor DENIED message: [16:14] [ 274.185989] type=1400 audit(1402589384.417:111): apparmor="DENIED" operation="mount" info="failed type match" error=-13 profile="/usr/bin/lxc-start" name="/" pid=3388 comm="lxc-start" flags="rw, rslave" [16:14] sforshee, yeah, I'm happy to wait for you to feel less hesitant about it :-) [16:14] and the lxc-start command throws 'lxc-start: Permission denied - Failed to make / rslave' [16:14] psivaa: odd, can you paste /proc/self/mountinfo? [16:15] stgraber: http://paste.ubuntu.com/7634401/ [16:16] ok, what kernel and apparmor do you have on that machine? [16:17] stgraber: apparmor: 2.8.95~2430-0ubuntu5 and kernel: 3.13.0-29-generic #53- [16:19] hmm, that's pretty weird because you're not supposed to hit that code path on Ubuntu... [16:19] can you start the container with -o debug -l debug and pastebin the debug file? [16:25] stgraber: http://paste.ubuntu.com/7634437/ a question: do i not have to run lxc-create before ? :) [16:26] psivaa: ok, let's try a new clean simple container for now :) [16:26] psivaa: lxc-create -t download -n test -- -d ubuntu -r trusty -a amd64 && lxc-start -n test [16:30] stgraber: so the creating went fine, but "sudo lxc-start -f /etc/lxc/default.conf -n test" outputs "lxc-start: Permission denied - Failed to make / rslave" [16:32] with similar apparmor DENIED message in dmesg === psivaa is now known as psivaa-afk === ayan_ is now known as ayan === ayan is now known as Guest78419 [19:13] Hi all. I am using an USB3 attached SSD, and for TRIM support, I'm looking for UAS driver. But it seems that it was removed somewhere between quantal and raring. Can you advise me what is the nicest way to get it for myself [19:25] ha1dfo, http://pastebin.ubuntu.com/7635162/ [19:36] bjf, Thank you! [19:37] ha1dfo, nothing to really thank be for, it's busted and we're not going to enable it. you'll have to build your own kernel to do so [19:37] well, you saved me a couple of hours of searching for the reason, for sure! [19:37] ha1dfo, heh, well your welcome for that :-) [19:37] ok, so if uas is broken, then is there any other way to get TRIM working via usb3? [19:38] ha1dfo, that i don't know. cking? ^ [19:43] that I don't know about :-/ [19:44] was worth a shot