slangasek | ok seriously, what's going on with the autopkgtest queues? who put the handbrake on our x86 cloud? | 00:31 |
---|---|---|
slangasek | and also somehow arm64 | 00:31 |
mwhudson | arm64 comes pre-handbraked | 00:33 |
slangasek | mwhudson: yes, but compare https://cdo.kpi.canonical.com/d/000000030/ubuntu-foundations?panelId=19&fullscreen&orgId=1&from=1536109870476&to=1536242063433 to https://cdo.kpi.canonical.com/d/000000030/ubuntu-foundations?panelId=19&fullscreen&orgId=1&from=1536593296201 | 00:36 |
slangasek | is this the fault of a new toolchain? | 00:36 |
mwhudson | does grafana do derivatives | 00:37 |
mwhudson | the work rate on those two seems approximately the same? | 00:38 |
mwhudson | about 20 jobs / hour on arm64 | 00:38 |
slangasek | ah, fair point | 00:39 |
slangasek | https://cdo.kpi.canonical.com/d/000000030/ubuntu-foundations?panelId=19&fullscreen&orgId=1&from=1535119606275&to=1535447022050 shows a much higher work rate, but maybe the tests aren't comparable | 00:40 |
jbicha | slangasek: is it affected by the archive rebuild in progress? | 00:48 |
slangasek | not directly | 00:48 |
slangasek | it'd be affected in terms of noisy neighbor behavior on the clouds | 00:48 |
=== pleia2_ is now known as pleia2 | ||
=== simonquigley is now known as tsimonq2 | ||
=== 17SAAK0SL is now known as infinity | ||
Laney | slangasek: I'd check with scalingstack people; could be L1TF hitting performance or something (for x86). Also I would expect the rate to pick up once the KDE tests are through - they are mostly not very nice tests as they are C++ projects which rebuild themselves which takes ages. | 08:10 |
wgrant | There's a test rebuild ongoing, and overcommit is much worse than is traditional due to L1TF, yes | 08:11 |
* Laney said the magic word | 08:12 | |
-queuebot:#ubuntu-release- Unapproved: accepted gcc-7 [sync] (bionic-proposed) [7.3.0-27ubuntu1~18.04] | 08:55 | |
-queuebot:#ubuntu-release- Unapproved: accepted gcc-8 [sync] (bionic-proposed) [8.2.0-1ubuntu2~18.04] | 09:43 | |
-queuebot:#ubuntu-release- Unapproved: appstream (bionic-proposed/main) [0.12.0-3 => 0.12.0-3ubuntu1] (desktop-core) | 10:02 | |
bluesabre | Release team, can you please approve of the Xubuntu UIFe for the latest elementary-xfce? https://bugs.launchpad.net/ubuntu/+source/elementary-xfce/+bug/1792555 I've got some additional acks on the way as well | 10:34 |
ubot5 | Ubuntu bug 1792555 in elementary-xfce (Ubuntu) "[UIFe] elementary-xfce 0.13-1" [Undecided,New] | 10:34 |
Ukikie | Note that while this switches the theme from being svgs to pngs, this is actually what xubuntu-icon-theme did, only difference is optipng thrown in too. | 11:11 |
bluesabre | thanks Ukikie | 11:16 |
mfo | sil2100, hey! could you please trigger a rebuild of nodejs in cosmic-proposed for ARM 64 and i386? there's a test-case failure that seems to intermittent (parallel/test-net-listen-after-destroying-stdin), which I'm investigating, so a rebuild would help. note that s390x also FTBFS, but due to another test-case which i'm still working on. | 12:23 |
sil2100 | mfo: ok o/ | 12:33 |
mfo | sil2100, thanks o/ | 12:33 |
sil2100 | mfo: done, fingers crossed for success this time | 12:33 |
mfo | sil2100, heh, right. btw, do you happen to know whether it's possible to download the sbuild schroot used by LP somewhere? it's being too hard to reproduce this testcase failure in a number of environemnts, including a sbuild cosmic schroot running on xenial (as LP), so i'm trying to get the environment even closer. | 12:36 |
cjwatson_ | apt install sbuild-launchpad-chroot | 12:36 |
=== cjwatson_ is now known as cjwatson | ||
cjwatson | Or you can fetch it directly using the manage-chroot tool in lp:ubuntu-archive-tools | 12:37 |
cjwatson | Or from the chroot_url links in https://api.launchpad.net/devel/ubuntu/cosmic/amd64 etc. | 12:37 |
mfo | cjwatson, great, thank you! | 12:40 |
tkamppeter | Ghostscript is hanging in -proposed on the ocrmypdf autopkgtest, but ocrmypdf is only in Universe. | 12:49 |
tkamppeter | Current Ghostscript is an important security updated. What should we do here? | 12:49 |
-queuebot:#ubuntu-release- New binary: linux-signed [ppc64el] (bionic-proposed/main) [4.15.0-35.38] (core, kernel) | 13:14 | |
-queuebot:#ubuntu-release- Unapproved: gdm3 (bionic-proposed/main) [3.28.3-0ubuntu18.04.1 => 3.28.3-0ubuntu18.04.2] (ubuntu-desktop) | 13:28 | |
=== jdstrand_ is now known as jdstrand | ||
ginggs | tkamppeter: https://github.com/jbarlow83/OCRmyPDF/commit/517b385fe5cb2195023100a807e6f18dc7e6faea#diff-b61a6d542f9036550ba9c401c80f00ef | 14:04 |
ginggs | tkamppeter: i'm going to try cherry-picking that | 14:04 |
-queuebot:#ubuntu-release- Unapproved: zfs-linux (xenial-proposed/main) [0.6.5.6-0ubuntu24 => 0.6.5.6-0ubuntu25] (no packageset) | 14:44 | |
tkamppeter | ginggs, thank you very much in advance. | 14:53 |
ginggs | tkamppeter: that fixed one of the three failing tests, I'll look at the others soon | 15:10 |
tkamppeter | OK, thanks. | 15:32 |
tkamppeter | Another problem I am running into: cpdb-libs hangs in -proposed due to its autopkgtest failing only on arm64 and I have no way to debug it as the porter box for arm64 is broken. Can one make an exception here? AFAIR it never passed on this arch. | 15:37 |
Laney | that seems to be not a true statement: http://autopkgtest.ubuntu.com/packages/c/cpdb-libs/cosmic/arm64 | 15:39 |
Laney | those results to me look like 1.2.0-0ubuntu4 is broken and 1.2.0-0ubuntu2 works | 15:39 |
-queuebot:#ubuntu-release- New: accepted linux-signed [amd64] (bionic-proposed) [4.15.0-35.38] | 16:54 | |
-queuebot:#ubuntu-release- New: accepted linux-signed [ppc64el] (bionic-proposed) [4.15.0-35.38] | 16:54 | |
-queuebot:#ubuntu-release- Unapproved: gce-compute-image-packages (bionic-proposed/universe) [20180510+dfsg1-0ubuntu4~18.04.1 => 20180905+dfsg1-0ubuntu1~18.04.0] (ubuntu-cloud) | 17:15 | |
-queuebot:#ubuntu-release- Unapproved: grub2 (cosmic-proposed/main) [2.02+dfsg1-5ubuntu3 => 2.02+dfsg1-5ubuntu3] (core) | 18:30 | |
-queuebot:#ubuntu-release- Unapproved: grub2 (cosmic-proposed/main) [2.02+dfsg1-5ubuntu3 => 2.02+dfsg1-5ubuntu3] (core) | 18:55 | |
slangasek | Laney, wgrant: ok. So L1TF is still affecting performance that gravely even now that intel-microcode updates are released? | 19:16 |
-queuebot:#ubuntu-release- Unapproved: accepted grub2 [amd64] (cosmic-proposed) [2.02+dfsg1-5ubuntu3] | 19:49 | |
-queuebot:#ubuntu-release- Unapproved: accepted grub2 [arm64] (cosmic-proposed) [2.02+dfsg1-5ubuntu3] | 19:49 | |
-queuebot:#ubuntu-release- Unapproved: ibm-java80 (xenial-proposed/partner) [8.0.5.17-0ubuntu1 => 8.0.5.20-0ubuntu1] (no packageset) | 20:23 | |
-queuebot:#ubuntu-release- New binary: octavia [amd64] (cosmic-proposed/universe) [3.0.0-0ubuntu2] (no packageset) | 20:44 | |
wgrant | slangasek: intel-microcode only makes the L1 flush a bit more efficient; it doesn't mitigate the L1 flush's effects or make SMT usable (and we deployed the new microcode long before Ubuntu released it) | 22:04 |
wgrant | The permanent loss of SMT is the big problem | 22:04 |
wgrant | That's an easy >60% hit to capacity | 22:05 |
slangasek | ok | 22:10 |
-queuebot:#ubuntu-release- New binary: linux-signed-azure [amd64] (bionic-proposed/main) [4.15.0-1024.25] (kernel) | 22:19 | |
mfo | slangasek, hey, good news. the nodejs s390x test-case failure on cosmic-proposed is fixed with the libuv1 debdiff in LP #1792647. can i ask you for a review/upload when you have a chance, please? | 23:08 |
ubot5 | Launchpad bug 1792647 in libuv1 (Ubuntu) "fix nodejs test-case failure on s390x and LXD" [Undecided,New] https://launchpad.net/bugs/1792647 | 23:08 |
* mfo back later. | 23:14 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!