| === JanC is now known as Guest16348 | ||
| === JanC_ is now known as JanC | ||
| hallyn | hm, https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1481723 is still happening to me in a zesty vm | 05:08 |
|---|---|---|
| ubot5 | Ubuntu bug 1481723 in linux (Ubuntu) "can not build kernel image from source package" [Medium,Confirmed] | 05:08 |
| === infinity1 is now known as infinity | ||
| pjf | apw: Thank you for looking into the mainline kernel repo! ( #1681183 ) | 13:33 |
| apw | pjf, np. very odd that the sync just stopped working, seems we lost a rule or something a few weeks back and as it is only am mirror things kept working | 13:34 |
| pjf | apw: I'm just glad I found the right place to report it. :) | 13:35 |
| leitao | I understand that the kernel for 17.04 GA is already on freeze since April 6th, right? | 13:36 |
| leitao | Any new patch will be only accepted in next SRU, rigth? | 13:36 |
| apw | leitao, it would normally take something boot essential to repin the kernel now in the release images. we do sometimes queue up a day0 SRU kernel if there are urgent non-boot-essential changes. otherwise we are in SRU mode, so in the next SRU cycle. | 13:39 |
| leitao | apw, got it. | 13:41 |
| apw | leitao, if you do have something you want considered we do that in the normal way, bug and email to the list | 13:48 |
| leitao | apw, sure, I think we can way the first SRU. | 13:50 |
| leitao | thanks! | 13:50 |
| acheronuk | Hi. is there any documentation explaining the switch to CFQ for 4.10? | 16:28 |
| apw | acheronuk, i believe that we did some significant testing on that, based on performance. we moved to deadline previously because performance sucked with cfq | 16:52 |
| apw | acheronuk, is it causing some sort of issue ? | 16:52 |
| acheronuk | apw: no problem. just a forum user posing the question... https://www.kubuntuforums.net/showthread.php?71686-Stll-wrong-i-o-scheduler-used-in-17-04 | 16:54 |
| apw | which "right" scheduler do they want | 16:54 |
| hallyn | apw: does anyone on kernel team do somewhat regular testing with ltp? | 16:54 |
| apw | hallyn, ltp is a pretty poor test suite in the main | 16:55 |
| acheronuk | apw: as usual,depends which random blog from someone who thinks they know you read :P | 16:55 |
| apw | acheronuk, indeed they are quoting cking who is the one who does those kinds of analysis | 16:55 |
| apw | acheronuk, and frankly they can change the default live iirc let alone at boot time | 16:55 |
| acheronuk | they wanted kubuntu to switch from CFQ to deadline, and I just pointed out that ubuntu had just gone the other way | 16:56 |
| acheronuk | so I was looking for a more detailed explanation | 16:56 |
| acheronuk | that is all | 16:56 |
| apw | i thought kubuntu was the one which was overriding us from deadline to cfq because you rely on some specific feature of cfq | 16:58 |
| apw | something to do with being able to hammer the disk and only use idle cycles | 16:58 |
| ogra_ | yeah, wasnt that nepomu (or some other tool with N) that was indexing the disk content | 16:59 |
| ogra_ | *nepomuk | 16:59 |
| acheronuk | apw: we have such an override in our setting, yes. that was the OP's point in the thread. they wanted us to stop that now | 16:59 |
| apw | which as you say would leve them as cfq anyhow | 17:00 |
| apw | but i think you needed it for reasons that deadline can never offer | 17:00 |
| acheronuk | that decision was before I was involved | 17:00 |
| apw | acheronuk, yeah, but i think i was involved on this side :) | 17:01 |
| acheronuk | I'm sure :P | 17:01 |
| hallyn | apw: yeah, that's what i thought. but i'm pushing extension of filecaps in kernel - current filecaps tests are in ltp. i could either extend those in ltp, or port them all to run standalone and add them to the kernel source. (or something else) | 17:01 |
| hallyn | advice? :) | 17:01 |
| acheronuk | apw: anyway, just wanted to know why the change on the other flavours to CFQ now with 4.10, so I could explain why kubuntu should probably not buck the trend once again | 17:03 |
| hallyn | think i'll port it all to linux/tools/testing/ | 17:04 |
| hallyn | which will slow me down a bit... | 17:05 |
| apw | acheronuk, there is a fair write up in the commit. cking did boot and performance testing | 17:09 |
| apw | acheronuk, for the issues we have moved away from cfq for. | 17:09 |
| apw | acheronuk, cfq is our natural default for our work-loads "nominally" | 17:09 |
| acheronuk | apw: yep. had just this second found it! http://kernel.ubuntu.com/git/ubuntu/ubuntu-zesty.git/commit/?id=af80b83a2b6184fea27f050948146fcd9a28070d | 17:11 |
| acheronuk | thx | 17:11 |
| apw | hallyn, there is some sanity in those being in the kernel self-tests in my mind | 17:12 |
| apw | hallyn, but i think ltp is one of the tests you can in theory run, and perhaps we can engineer a run with just hte tests you add if they end up in there | 17:12 |
| apw | hallyn, cking again is my expert at ading tests | 17:13 |
| apw | hallyn, so can perhaps give you a more targetted answer as to which is better | 17:13 |
| hallyn | apw: ok, thx. i know i rarely run ltp tests (despite having written quite a few), and havne't heard of anyone who does.. now if it once gets integrated into ubuntu automated tests then maybe that's a good thing, but... | 17:24 |
| apw | hallyn, the issue is it has a crap-ton of tests many of which fail and that is "right" and there is no way to know | 18:24 |
| apw | hallyn, well that _was_ last time i worked with it | 18:24 |
| hallyn | apw: yes, though you can easily specify the testsuites to run, and only run the ones you're interested in. But if most are not interesting then it doesn't really make sense | 18:42 |
| ogasawara | bug 1672850 | 19:19 |
| ubot5 | bug 1672850 in linux-aws (Ubuntu) "[MIR] linux-aws" [Undecided,Fix committed] https://launchpad.net/bugs/1672850 | 19:19 |
| ogasawara | cyphermox: ^^ random question, I see that is 'Fix Committed', what's the next step here? Is there anything I need to do to shepherd that over the goal line? | 19:20 |
| cyphermox | ogasawara: not really, aside from kicking slangasek to do his magic. From that point you need an archive admin to apply the change. | 19:21 |
| ogasawara | cyphermox: ack, thanks | 19:21 |
| cyphermox | I guess apw could do that too | 19:22 |
| slangasek | ogasawara, cyphermox: is it seeded in zesty? | 19:23 |
| ogasawara | slangasek: I don't know | 19:24 |
| slangasek | it's not in zesty at all, so that answers that | 19:24 |
| cyphermox | no itś for xenial AFAIk | 19:24 |
| slangasek | right | 19:24 |
| slangasek | so that would've been invisible to the AA team without a ping, fwiw - doing now | 19:25 |
| cyphermox | ah, good to know, thanks. | 19:25 |
| ogasawara | indeed, /me makes a note, thanks slangasek | 19:25 |
| Tahvok | Hey guys! | 20:32 |
| Tahvok | ubuntu-support-status is showing a wrong support time for hwe kernel. For example it shows that kernel 4.8 is supported till' April 2022. | 20:32 |
| Tahvok | I just can't find hwe-support-status package anywhere for 16.04, so maybe I used a wrong tool? | 20:33 |
| === acheronuk is now known as acheronUK | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!