apw | Noskcaj, that might be a plymouth issue which just got fixed | 00:30 |
---|---|---|
TJ- | apw: quick Q on behalf of a user: when is the Wily kernel matching the 4.2.8 stable mainline due? There's some BTRFS bugs the user has suffered from | 00:32 |
TJ- | User relies on AUFS too for LXC/LXD so can't simply switch to a mainline build. I suggests the Xenial 4.3.3 but that breaks user's ZFS DKMS module builds :) | 00:33 |
apw | TJ-, xenial should have zfs built-in though | 01:01 |
apw | TJ-, i would expect 4.2.8 to be in the next upload, but that is likely 5 weeks out | 01:02 |
TJ- | apw: I didn't know that; I'll tell the user next time they pop up :) | 01:02 |
TJ- | Yes, I thought it'd be next month for 4.2.8. | 01:03 |
=== _fortis_ is now known as _fortis | ||
PETSounds | Hi. Which kernel version best for Intel Skylake? | 14:34 |
PETSounds | i am currently on WIly kernel, but the performance is not that smooth. Just wonder if upgrading Kernel and Intel Graphics driver will improve the performance. | 14:39 |
apw | PETSounds, hmmm, well vivid was the first one to have any kind of support and that was backported, wily is intended to have native support | 14:47 |
apw | PETSounds, that said, it is very new even there, so you might find the xenial 4.3 kernel has some benefits, and we would be interested if so | 14:47 |
PETSounds | apw: thanks for your response. What do i need to do to make Broadcom driver doesn't stop working when upgrading Kernel? | 14:49 |
apw | PETSounds, bcmwl ? | 14:50 |
PETSounds | i upgraded | 14:51 |
PETSounds | i upgraded the kernel from http://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D and WLAN stops working | 14:51 |
apw | PETSounds, very likely, if it is bcmwl, you would want to get a copy of that from xenial as well as that is sync'd with the kernel in xenial | 14:51 |
apw | PETSounds, note i am talking about the ubuntu archive xenial 4.3 kernel not the ones in mainline-build respository | 14:52 |
PETSounds | apw: ahh ok. thank you. i'll try again and report back. | 14:53 |
=== cyphermox_ is now known as cyphermox | ||
yoasif | any chance anyone can tell me when we can expect a new kernel based on http://bugs.launchpad.net/bugs/1527359 ? | 15:19 |
ubot5 | Launchpad bug 1527359 in linux (Ubuntu Wily) "Wily update to v4.2.8 stable release" [Undecided,Confirmed] | 15:19 |
yoasif | looking for fixes http://kernel.ubuntu.com/git/ubuntu/ubuntu-wily.git/commit/?h=master-next&id=c7c784296cf04c0757f39aa00c77d12a4f2a5548 and http://kernel.ubuntu.com/git/ubuntu/ubuntu-wily.git/commit/?h=master-next&id=9e0d373b6959298b9a07cb2975d04a7aac7e0e23 | 15:20 |
apw | yoasif, that is already applied to master-next so it should be in the next sru cycle, which is about 5 weeks out with xmas in the way | 15:20 |
TJ- | yoasif: aha, I asked apq about this last night. You should be able to use the Xenial kernel since it has the ZFS driver built-in - no need for the DKMS | 15:21 |
apw | yoasif, is there a bug filed for your specific issue, is so i could spin you up a test kernel to confirm that the update is what you need | 15:21 |
TJ- | s/apq/apw/ | 15:21 |
yoasif | apw: ouch — so right now i'm running the mainline build from the kernel ppa page but there's no -extras package (so no aufs and docker doesn't work)… any chance there's an easy way to build an extras package for the mainline kernel? | 15:22 |
apw | (the bug gives me something to publish against | 15:22 |
yoasif | (the issue has been fixed in mainline http://kernel.ubuntu.com/%7Ekernel-ppa/mainline/v4.2.8-wily/ | 15:22 |
apw | yoasif, there are no extras packages for mainline builds, it is disabled, which means everything is builtin, but aufs is not in mainline so you don't get that | 15:23 |
yoasif | yeah :/ | 15:23 |
apw | everything is in the linux-image which exists | 15:23 |
apw | i can buid you a test kernel against that tip to use till the official one comes out in the new year if that is of use | 15:23 |
yoasif | but i should be able to use xenial's kernel with that extras package for the time being (since zfs is built in?) | 15:24 |
apw | xenials kernel contains the zfs bits nativly, but it might be simpler | 15:24 |
apw | for you to have a test kernel from me, which would sit version wise in the middle and upgrade correctly when | 15:24 |
apw | the real distro kernel comes out | 15:24 |
apw | if everything you need is in the .8 update | 15:25 |
yoasif | apw: i'd really appreciate that — but i'd need -extras too, since aufs isn't in mainline :/ | 15:25 |
apw | yoasif, extras comes out as part of the build automatically for the ubuntu builds | 15:25 |
yoasif | right now i'm running mainline, but i can't run docker, so i don't have any services running ;) | 15:26 |
yoasif | (been replacing disks on my btrfs raid, and the hard lockups were really annoying) | 15:26 |
apw | yep those are only there for problem isolation, they aren't production quality by any means | 15:27 |
yoasif | apw: so if you were to create a test kernel, would i be able to have aufs as a module (easily)? | 15:30 |
apw | yes, it would be an intermim build just the same as if i had uploaded the current state of the tree | 15:30 |
apw | yoasif, i assume you are needing an x86 kernel of one size or other | 15:31 |
yoasif | x64, yeah | 15:31 |
yoasif | apw: the mainline build is actually more "production quality" (in that it doesn't lock up) than the stable ubuntu build besides the fact that i don't have aufs as a module ;) | 15:38 |
apw | yoasif, well that is btrfs, that well known awsome filesystem | 15:40 |
yoasif | apw: i'm on the hype train for it… not as stable as zfs yet, but feature wise, it allows me to expand arrays without creating new vdevs — it's going to be fantastic once it is stable, but definitely experiencing some growing pains | 15:41 |
yoasif | also built the array prior to zfs being available in ubuntu without ppas | 15:42 |
apw | yoasif, "once" it is stable, it is years old now, it reallly should be there by now if it is going to be | 15:43 |
* apw strokes his ext4 filessytem gently, good filesystem | 15:43 | |
yoasif | been experiencing this issue for weeks now, even decided to build a new machine for it (not that i'm complaining) | 15:44 |
yoasif | never quite understood lvm (i should learn) and i was used to zfs, so been using btrfs; ext4 was never an option for this | 15:45 |
yoasif | (running ext4 on top of lvm for my root though) | 15:45 |
apw | yeah i am old school most of my boxes are whatever ordinary filesystems on top of lvm | 15:46 |
apw | which i am now at least at some kind of agreement with, it doesn't mess with me too much and i don't ask it to change much | 15:46 |
yoasif | well, there are supposed to be integrity guarantees with btrfs; unfortunately, i have lost data once with btrfs (ran a fsck on an array where one of the disks had dropped out and seen subsequent writes — destroyed the array for the most part) | 15:48 |
yoasif | never lost data with zfs though, been using an array for years now, even with bad disks | 15:48 |
yoasif | (should have scrubbed again and again before trying fsck) | 15:49 |
apw | fsck on btrfs has a bit of a reputation for being worse than not bothering | 15:49 |
apw | and hoping for the best ... | 15:50 |
yoasif | well, the mailing list is great … "if you don't have backups, you clearly didn't care about the data" | 15:50 |
yoasif | (which is true, but doesn't change the fact that this shouldn't happen) | 15:50 |
apw | yoasif, such an inspiring position :) | 15:50 |
apw | yoasif, http://people.canonical.com/~apw/stable-4.2.8-wily/ | 15:53 |
yoasif | but yeah, in case anyone wanted some unsolicited advice; do not fsck your btrfs array… try to do as many online scrubs as you can until you don't see errors | 15:53 |
yoasif | apw: not local to my machine right now, but i *really* appreciate you taking the time… will report back tomorrow morning (or tonight if you are around) | 15:54 |
yoasif | (no docker means no openvpn either ;) ) | 15:54 |
apw | yoasif, my nick is always arround, my body less reliably, but yeah let me know if that gets you going again | 15:55 |
yoasif | apw: definitely will do | 15:55 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!