=== JanC is now known as Guest21552 === JanC_ is now known as JanC === cpaelzer_ is now known as cpaelzer [08:07] where should I file bugs for the lts-yakkety from the ppa? [08:16] jtaylor, https://bugs.launchpad.net/ubuntu/+source/linux-lts-yakkety/+filebug i guess [08:16] jtaylor, but do tell us the bug # here [08:27] hm is launchpad broken? [08:27] get an oops on submit ... [08:28] lovely... try fileing it against linux itself, and i'll sort it out afterwards [08:28] jtaylor, ^ [08:31] apw: bug 1631298 [08:31] bug 1631298 in linux (Ubuntu) "lts-yakkety 4.8 cannot mount lvm raid1" [Undecided,New] https://launchpad.net/bugs/1631298 [08:34] jtaylor, can i assume this is installing into an existing working system ... if so could you pastebin me a lsinitamfs of the working and not working initrds please [08:36] 4.8 http://paste.ubuntu.com/23288094/ [08:36] 4.4 (working) http://paste.ubuntu.com/23288095/ [08:37] thanks [10:08] jtaylor, could you boot up this test kernel for me so we can find out what the heck the feature flags are set to ... http://people.canonical.com/~apw/lts-backport-yakkety-xenial/ [10:09] jtaylor, not expecting it to fix anything but should drop an APW: in the dmesg with some info [10:23] jtaylor, ok naive testing of raid1s created on 4.4 and booted with 4.8 seem to assemble for me [10:23] apw: my raid is probably a lot older [10:23] jtaylor, so we need to get that debug output to see what differs on yours to mine [10:23] maybe 3.13 or more ;) [10:23] jtaylor, probabally and that is a little scarey :) [10:23] I'll reboot soon [10:23] jtaylor, thanks [10:35] apw: hm all zeros http://paste.ubuntu.com/23288447/ [10:37] jtaylor, oh you are right i have the same error, and yet the lv is there, oh but is it running [10:38] it is displayed but not active [10:38] though I haven't tried activating it manually, I just figured if it puts me in an emergency shell that doesn't work [10:38] i would doubt it will start [10:38] I actually forgot to mention that, the system doesn't boot, despite / and /boot being there [10:38] ok i am reproducing in fact, just not seeing what it is saying [10:46] jtaylor, i am going to assume the data you ahve in your raid1 is something you care about (given it is raid1d) [10:46] jtaylor, ie this is not something you can trivially test :/ [10:47] yes, though I have backups [10:47] but destroying it would still be inconvinient [10:47] jtaylor, do you have a test environment at all? obviously i can test a fix here but this is pretty scarey code to change [10:48] jtaylor, anyhow i have a theory and i am testing it, and if that works i'll ask upstream for safety [10:49] no, this is my home setup I only have this one raid1 system [10:50] jtaylor, though i guess in theory we could pull off one of the mirrors as a backup, hrm anyho [10:50] lets get to that once we have any idea if this is right [10:50] if you are reasonably confident in a fix I can update my backups and test it === mhcerri_ is now known as mhcerri [13:56] jsalisbury: updated comments on the bug report. The last kernel is good, second-last is bad. [13:59] om26er, great, I'll build the next one. Only one or two left. [13:59] jsalisbury: ack. [15:38] rtg, something going on with tangerine ? compiling arm64 Error: selected processor does not support `staddlh w0,[x1]' [15:40] rtg, looks like assembler messages from xenial-amd64 chroot [15:43] bjf, apw ^^ anyone know ? trying to build 4.8.0-21.23 [15:45] I am also seeing ==> Error: attempt to move .org backwards [15:47] build works fine with yakkety-amd64 chroot.. .something broken with xenial-amd64 chroot ? [15:56] jsalisbury: Hi! that's a good kernel. [15:56] om26er, ack [15:58] jsalisbury: how big is the diff from this stage ? (or number of commits) [16:01] om26er, this next kernel will narrow it down between 5 sched commits: [16:01] 55e16d3 sched/fair: Rework throttle_count sync [16:01] 599b484 sched/core: Fix sched_getaffinity() return value kerneldoc comment [16:01] 8663e24 sched/fair: Reorder cgroup creation code [16:01] 3d30544 sched/fair: Apply more PELT fixes [16:01] 7dc603c sched/fair: Fix PELT integrity for new tasks [16:02] jsalisbury: I am betting on the first one [16:03] om26er, :-) We shall know shortly [16:45] om26er, next kernel is posted [17:24] jsalisbury: that's a bad kernel [17:30] jsalisbury: do we finally have a commit blame ? === om26er is now known as om26er1 === om26er1 is now known as om26er [18:20] om26er, we have to test only one more. Then I'll bulild a test kernel with the reported commit reverted. The next kernel should be done in a few moments [18:20] hi. i saw gresurity related utilities but no grsecurity kernel. how so? [18:20] jsalisbury: great [18:42] om26er, next test kernel is posted [18:42] jsalisbury: thanks, testing it now. [18:42] om26er, ack [18:49] jsalisbury: that's a good one. [18:49] om26er, The bisect reports this as the offending commit: [18:49] commit 3d30544f02120b884bba2a9466c87dba980e3be5 [18:49] Author: Peter Zijlstra [18:49] Date: Tue Jun 21 14:27:50 2016 +0200 [18:49] sched/fair: Apply more PELT fixes [18:50] om26er, I'll build a test kerne with it reverted to see if it fixes the bug. [18:51] jsalisbury: you mean 4.8 release minus that commit ? [18:51] om26er, I'll build a yakkety kernel, minus that commit [18:51] jsalisbury: great [18:54] feels like a small diff https://kernel.googlesource.com/pub/scm/linux/kernel/git/tip/tip/+/3d30544f02120b884bba2a9466c87dba980e3be5%5E%21/ [19:19] om26er, kernel posted. Note with this kernel, you need to install both the linux-image and linux-image-extra .deb packages. [19:30] jsalisbury: that's it. [19:30] latest kernel is good [19:30] om26er, good news. I'm going to ping upstream and the patch author and get their feedback [19:31] jsalisbury: ok, the test case is `stress -c $your_total_cpu_cores` [19:32] om26er, thanks. Can you post that in the bug? I'll be pointing upstream there. === leftyfb_ is now known as leftyfb [19:34] done. === sts is now known as Guest59171 === xnox_ is now known as xnox === Darcy is now known as Spydar007 === kloeri_ is now known as kloeri === bdmurray_ is now known as bdmurray