| *** paulw has joined #baserock | 06:23 | |
| *** mike has joined #baserock | 06:54 | |
| *** mike is now known as Guest10066 | 06:54 | |
| *** mdunford has joined #baserock | 07:00 | |
| *** richard_1aw is now known as richard_maw | 07:07 | |
| *** ChanServ sets mode: +v richard_maw | 07:08 | |
| *** fay_ has joined #baserock | 07:35 | |
| *** jonathanmaw has joined #baserock | 08:02 | |
| *** bashrc_ has joined #baserock | 08:07 | |
| *** petefoth has joined #baserock | 08:14 | |
| *** gary_perkins has joined #baserock | 08:36 | |
| *** CTtpollard has joined #baserock | 08:53 | |
| pedroalvarez | hehe, one of the latest build times of mason: 123459s | 09:14 |
|---|---|---|
| pedroalvarez | almost! | 09:14 |
| rjek | haha | 09:14 |
| rjek | Make a very small optimisation. | 09:15 |
| pedroalvarez | :) | 09:15 |
| *** petefoth has quit IRC | 09:17 | |
| *** petefoth has joined #baserock | 09:20 | |
| *** rdale has joined #baserock | 09:21 | |
| *** petefoth has quit IRC | 09:52 | |
| *** petefoth has joined #baserock | 09:55 | |
| *** petefoth has quit IRC | 09:56 | |
| *** petefoth has joined #baserock | 10:26 | |
| paulsherwood | pedroalvarez: ybd :-) | 10:44 |
| pedroalvarez | tbf this was from scratch, 4 systems including the openstack one, and in a not-powerful VM | 10:47 |
| pedroalvarez | but yes, too slow | 10:47 |
| pedroalvarez | Not sure that ybd is more optimal than morph, altough I have to confess that I've never used it | 10:48 |
| pedroalvarez | meh, the partitioning change makes the flashing script for the jetson board fail :) | 11:05 |
| paulsherwood | pedroalvarez: does the release run *from scratch* or re-use existing cache artifacts if available? | 11:10 |
| pedroalvarez | well, that depends on where and how you do the release, but ideally it re-uses existing cache artifacts | 11:17 |
| radiofree | pedroalvarez: send a patch for the flashing script! | 11:33 |
| pedroalvarez | radiofree: yes! I was figuiring out how to do it | 11:34 |
| pedroalvarez | I was pondering using losetup, and then try to mount partition 1, and fall back to mount the whole disk | 11:35 |
| radiofree | pedroalvarez: it's a bit weird, you'd have to find out the where boot and the rootfs start (e.g parted -> u -> b -> p and note the Start bytes) | 11:36 |
| radiofree | then you could mount the img with mount -o loop,offset=STARTBYTES yourimage tmp/bootpartition | 11:37 |
| radiofree | copy everything in there... probably | 11:37 |
| pedroalvarez | :) | 11:37 |
| radiofree | although actually maybe if you just format the emmc and dd the whole lot over... that might work | 11:37 |
| pedroalvarez | but iiuc, the script extracts the u-boot file to flash it | 11:39 |
| pedroalvarez | I should read the flashing script before continue thinking about this | 11:40 |
| radiofree | you could rewrite the whole thing in something sensible like python if you want :D | 11:41 |
| pedroalvarez | haha | 11:41 |
| pedroalvarez | I wasn't plannig to spend a lot of time on this tbh :) | 11:42 |
| pedroalvarez | but it's a good idea | 11:42 |
| *** mdunford has quit IRC | 12:29 | |
| *** CTtpollard has quit IRC | 12:39 | |
| *** mdunford has joined #baserock | 12:44 | |
| *** rdale has quit IRC | 12:44 | |
| *** CTtpollard has joined #baserock | 12:54 | |
| *** mdunford has quit IRC | 12:58 | |
| *** mdunford has joined #baserock | 13:03 | |
| pedroalvarez | just "fixing" that wasn't enough. I've dd'ed the whole disk image (with 2 partitions) to mmcblk0p2 | 13:03 |
| pedroalvarez | It's probably easier to create a partitioning layout and then dd the whole thing | 13:04 |
| pedroalvarez | s/partitioning layout/partitioning layout for rawdisk deployments/ | 13:05 |
| *** petefoth has quit IRC | 13:18 | |
| *** petefoth has joined #baserock | 13:20 | |
| radiofree | pedroalvarez: if you have an image with a /boot / partition layout, you should be able to just dd it to mmcblk0 | 13:24 |
| radiofree | u-boot will look in mmcblk0p1:/extlinux/extlinux.conf | 13:25 |
| radiofree | so assuming your boot partition is sane it might work | 13:25 |
| *** petefoth has quit IRC | 13:41 | |
| *** zoli___ has joined #baserock | 14:08 | |
| pdar | I am aiming to do some patch reviews, but until I soe[B[B[B[B[B | 14:16 |
| pdar | apologies, | 14:17 |
| pdar | My terminal was frozen and then it suddenly wasnt | 14:18 |
| * pedroalvarez is waiting for something else after "but until I ..." | 14:26 | |
| pdar | Oh, it wasnt meant even for this channel, sorry pedro! | 14:28 |
| pedroalvarez | boo | 14:28 |
| * pedroalvarez continues reviewing on its own | 14:29 | |
| pdar | I'll do some reviews later if I have time, I'm promised to another project for now though | 14:31 |
| pedroalvarez | :) | 14:37 |
| paulsherwood | pedroalvarez: mason ci took 34.5 hours, is that correct? | 14:49 |
| paulsherwood | how (much) does it parallelise? | 14:50 |
| pedroalvarez | paulsherwood: is only one worker, so nothing | 14:51 |
| paulsherwood | last time i ran ybd to build all the x86 components on my mac a single instance took 07:13:37 | 14:51 |
| paulsherwood | pedroalvarez: so is that only x86? | 14:51 |
| pedroalvarez | all the x86_64 systems on clusters/ci.morph | 14:52 |
| rjek | paulsherwood: ISTR that your Mac has *significantly* superior IO than our cloud host | 14:52 |
| paulsherwood | i daresay. | 14:52 |
| pedroalvarez | indeed | 14:52 |
| paulsherwood | rjek: is there an amazon machine equivalent to 'our cloud host'? | 14:53 |
| rjek | paulsherwood: How do you mean? | 14:53 |
| paulsherwood | something of similar performance - otherwise i have no way to compare | 14:54 |
| rjek | I am not very familiar with AWS's offering, sorry | 14:54 |
| paulsherwood | what is the mason running on, then? | 14:54 |
| rjek | Or how their internal infrastructure works | 14:54 |
| rjek | paulsherwood: I had assumed OpenStack at DC | 14:55 |
| * rjek 's information may be out of date. | 14:55 | |
| paulsherwood | ah, so i could try the same | 14:55 |
| rjek | Ceph has some very nice properties, but performance and low latency are not among them | 14:55 |
| richard_maw | and it's not quite distributed btrfs, which I was disappointed with | 14:56 |
| pedroalvarez | It might be better/cheaper to have a super powerful VM that we only power of to do a build, and only take 1 hour to build everything, than having a slow one and taking 34 hours | 14:58 |
| rjek | Yeah | 14:58 |
| rjek | Also use a massive ram disc as a bcache :) | 14:58 |
| rjek | I know Collabora do than when building LibreOffice. | 14:58 |
| tlsa | doesn't mason build the chunks that are common between all four systems four times? If so, an obvious optimisation is to deduplicate that | 15:05 |
| *** jonathanmaw has quit IRC | 15:08 | |
| *** jonathanmaw has joined #baserock | 15:08 | |
| *** mdunford has quit IRC | 15:10 | |
| *** zoli____ has joined #baserock | 15:11 | |
| pedroalvarez | tlsa: nope, Sam fixed that a while ago :) | 15:11 |
| *** zoli_____ has joined #baserock | 15:12 | |
| tlsa | I don't think it's fixed | 15:13 |
| pedroalvarez | tlsa: hm.. why | 15:13 |
| tlsa | http://paste.baserock.org/nazeyayasu | 15:14 |
| *** zoli___ has quit IRC | 15:14 | |
| pedroalvarez | oh! | 15:14 |
| *** zoli____ has quit IRC | 15:14 | |
| SotK | :( | 15:14 |
| pedroalvarez | grrrrrrrrrr | 15:14 |
| tlsa | I picked glibc at random; its that same for everything I looked at | 15:14 |
| pedroalvarez | maybe it's fixed but the version used in mason doesn't have the full fix | 15:16 |
| pedroalvarez | I tested the fix, and it seemed to work :/ | 15:17 |
| tlsa | is it still the release-build script in definitions? | 15:17 |
| pedroalvarez | tlsa: thanks for pointing it out | 15:17 |
| pedroalvarez | tlsa: yes | 15:17 |
| tlsa | if so, it doesn't look like that can do any de duplication | 15:17 |
| tlsa | it just runs a separate morph build command for each system | 15:17 |
| pedroalvarez | tlsa: indeed, the fixes were for morph distbuild, so that it doesn't build twice something | 15:18 |
| tlsa | you probably need to do it in morph, so `morph build [list of systems]` | 15:18 |
| SotK | tlsa: I thought that script uses distbuild-morphology? | 15:18 |
| SotK | I hope it does anyway | 15:19 |
| tlsa | maybe, I didn't look deep | 15:19 |
| *** zoli_____ has quit IRC | 15:21 | |
| *** zoli___ has joined #baserock | 15:21 | |
| *** mdunford has joined #baserock | 15:25 | |
| * pedroalvarez tries to reproduce the error | 15:30 | |
| *** zoli___ has quit IRC | 15:40 | |
| *** zoli___ has joined #baserock | 15:42 | |
| pedroalvarez | reproduced \o/ | 15:42 |
| *** paulw has quit IRC | 15:45 | |
| *** zoli___ has quit IRC | 15:45 | |
| *** zoli___ has joined #baserock | 15:45 | |
| pedroalvarez | and still present in master /o\ | 15:46 |
| SotK | oh dear :( | 15:50 |
| pedroalvarez | looks like my testing-case wasn't good enough | 15:57 |
| *** zoli___ has quit IRC | 15:58 | |
| *** zoli___ has joined #baserock | 15:58 | |
| pedroalvarez | I'm not able to compete against ybd anymore | 16:06 |
| *** mdunford has quit IRC | 16:06 | |
| paulsherwood | lol | 16:08 |
| paulsherwood | ybd doesn't do stripping, and its cache algorithm is different, so i can't actually propose it as a 'solution' yet | 16:09 |
| pedroalvarez | IMO we should stop comparing and merge them | 16:10 |
| pedroalvarez | or create a third one :) | 16:11 |
| rjek | yabit; yet another build/integration tool | 16:13 |
| pedroalvarez | I knew this was going to happen :) | 16:13 |
| pedroalvarez | baserock-builder, or bb | 16:13 |
| rjek | haha | 16:13 |
| pedroalvarez | wait, no | 16:13 |
| CTtpollard | not bb | 16:13 |
| pedroalvarez | hehe | 16:13 |
| rjek | bakebits | 16:13 |
| paulsherwood | pedroalvarez: on aws, i believe ybd can do the whole x86 ci in under 2 hours | 16:35 |
| paulsherwood | (from scratch) | 16:35 |
| *** bashrc_ has quit IRC | 16:48 | |
| paulsherwood | there are some issues, though... for example http://paste.baserock.org/igedayasix | 16:51 |
| paulsherwood | after this is logged, my terminal goes blue text/underline :/ | 16:51 |
| paulsherwood | running on a mason-equivalent box now, 15-08-17 16:56:17 [CACHE-KEYS] Elapsed time for cache-key calculations 00:42:25 | 17:03 |
| paulsherwood | (that's the time to clone all the gits) | 17:03 |
| paulsherwood | will see how it does overnight | 17:04 |
| *** Guest10066 has quit IRC | 17:25 | |
| *** paulw has joined #baserock | 19:38 | |
| *** paulw has quit IRC | 19:42 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!