2019-10-01 10:51:04 algitbot: retry master 2019-10-01 11:05:03 that mesa failure on ppc64le smells like a parallel build issue 2019-10-02 07:04:24 should stop now 2019-10-02 07:18:49 algitbot: retry 3.9-stable 2019-10-02 07:22:21 algitbot: retry 3.9-stable 2019-10-02 07:28:11 algitbot: retry 3.9-stable 2019-10-02 07:30:52 _ikke_: i think we have an issue with mqtt 2019-10-02 07:31:38 lxc also acting strange 2019-10-02 07:32:35 ncopa: ping 2019-10-02 07:38:51 algitbot: retry 3.9-stable 2019-10-02 07:40:50 algitbot: retry 3.9-stable 2019-10-02 07:59:16 pong 2019-10-02 07:59:19 morning 2019-10-02 07:59:33 hi 2019-10-02 07:59:57 ncopa: after rebooting armv7 server it causes issues with builders 2019-10-02 08:00:27 one was in some loop and kept spamming mqtt 2019-10-02 08:00:35 ok 2019-10-02 08:00:40 i have seen that before 2019-10-02 08:00:49 not sure why it happens 2019-10-02 08:00:55 I restarted 2 of them 2019-10-02 08:01:04 but just stop mqtt-exec and start it again normally helps 2019-10-02 08:01:04 but now some packages dont want to get build 2019-10-02 08:01:11 on arm? 2019-10-02 08:01:14 or on others? 2019-10-02 08:01:21 looks like 3-10 builders hangs on python 2019-10-02 08:01:29 arm 2019-10-02 08:01:30 build-3-9-armv7 2019-10-02 08:01:38 it doesnt want to build exim 2019-10-02 08:01:45 and its not yet build 2019-10-02 08:03:36 i'll have a look 2019-10-02 08:03:54 i upgraded usa4 yesterday 2019-10-02 08:04:11 and i played with qemu on aarch64 2019-10-02 08:04:17 but its a bit... 2019-10-02 08:04:43 im not sure its smart to run docker in qemu on other arches 2019-10-02 08:08:37 what did you do? 2019-10-02 08:10:24 i logged into build-3-9-armv7, had a look at the running processes and installed packages 2019-10-02 08:10:27 uninstalled vim 2019-10-02 08:10:31 and rebooted 2019-10-02 08:10:42 i only installed vim 2 min ago 2019-10-02 08:10:47 ha! 2019-10-02 08:11:00 ok maybe 10 2019-10-02 08:11:03 :p 2019-10-02 08:11:04 Could not resolve host: hostap.epitest.fi 2019-10-02 08:11:13 looks like dns may be broke? 2019-10-02 08:11:17 i did not reboot the container again 2019-10-02 08:11:45 i dont know 2019-10-02 08:12:29 nah its aslso broke for me 2019-10-02 08:12:38 ERR_NAME_NOT_RESOLVED 2019-10-02 08:12:51 so its broken by upstream 2019-10-02 08:13:01 so you did nothing except just reboot build-3-9-armv7 2019-10-02 08:13:10 correct 2019-10-02 08:13:21 no problem analysis 2019-10-02 08:13:25 well, i first checked if it was "safe" to do so 2019-10-02 08:13:39 the mqtt-exec service was stopped 2019-10-02 08:13:45 nothing was running 2019-10-02 08:13:53 yes i was playing with it 2019-10-02 08:14:00 to enable logging 2019-10-02 08:14:09 i think we should turn that on anyways 2019-10-02 08:14:25 ok 2019-10-02 08:15:47 i have seen mqtt-exec go into some strange loop when host reboots 2019-10-02 08:16:06 i have never taken the time to analyze what is actually going on 2019-10-02 08:16:13 maybe thats when it previously crashed 2019-10-02 08:16:16 but not it will be restarted 2019-10-02 08:16:25 s/not/now 2019-10-02 08:16:48 it smells like some bug or race or something in mosquitto 2019-10-02 08:17:02 maybe its actually restarting all the time? 2019-10-02 08:17:09 didnt check 2019-10-02 08:17:17 no, i dont think so 2019-10-02 08:17:47 and i think it happens when there are two mqtt clients on same host starting at the same time or similar 2019-10-02 08:18:14 heh it actually still works? 2019-10-02 08:18:20 i was in the middle to edit the init.d 2019-10-02 08:18:23 it has leftovers 2019-10-02 08:18:54 sorry, i didnt know you were analyzing it 2019-10-02 08:21:39 ncopa: any comment regarding qemu? 2019-10-02 08:22:16 <@clandmeter> but its a bit... 2019-10-02 08:22:27 what is it a bit? compliated? 2019-10-02 08:22:35 limited 2019-10-02 08:22:39 and more complicated for sure 2019-10-02 08:22:41 how is it set up? 2019-10-02 08:22:53 undocumented also 2019-10-02 08:23:17 you create a vm where you run docker? 2019-10-02 08:23:52 i didnt even run docker yet 2019-10-02 08:24:04 qemu on arm works different 2019-10-02 08:24:07 had to do some digging 2019-10-02 08:24:13 so i ended up doing: 2019-10-02 08:24:16 qemu-system-aarch64 -m 32G -machine virt,gic-version=3 -enable-kvm -cpu host -smp cores=64 -nographic -nic user,model=virtio -drive format=raw,file=/dev/sdb -bios ./QEMU_EFI.fd -device virtio-rng-pci -device virtio-gpu-pci 2019-10-02 08:24:36 and it boots? 2019-10-02 08:24:41 sure 2019-10-02 08:24:46 so you use raw disk image 2019-10-02 08:24:51 you can boot very easy with kernel and initrd 2019-10-02 08:25:06 but if you want to use disk you need to use a boot loader 2019-10-02 08:25:11 thats first issue 2019-10-02 08:25:15 we dont not have one 2019-10-02 08:25:26 huh? 2019-10-02 08:25:47 efi boot loader for qemu 2019-10-02 08:25:55 we only build it for x86 2019-10-02 08:26:00 oh 2019-10-02 08:26:08 so i got it somewhere else 2019-10-02 08:26:25 so when it starts it didnt have display.. 2019-10-02 08:26:30 found out later why 2019-10-02 08:26:44 qemu aarch64 does not have a display by default 2019-10-02 08:26:51 needs -device virtio-gpu-pci 2019-10-02 08:27:33 and with aarc64 you need to specify machine 2019-10-02 08:27:51 which virt is the best option if you want kvm 2019-10-02 08:27:58 but that limits you to 8 cpus 2019-10-02 08:28:26 how do you connect to the display? 2019-10-02 08:29:22 you dont 2019-10-02 08:29:30 its stdout 2019-10-02 08:29:32 clandmeter: why do you use display on qemu? to see early boot messages? 2019-10-02 08:29:48 -nographic 2019-10-02 08:29:59 im doing it over ssh 2019-10-02 08:30:08 so you get serial console on stdout? 2019-10-02 08:30:19 yes 2019-10-02 08:30:41 i think with gpu option you can also run vnc 2019-10-02 08:30:45 but i found that out later 2019-10-02 08:31:33 to unlock thje 8 cpu limit i had to use another option 2019-10-02 08:31:46 gic-version=3 2019-10-02 08:32:02 i got it working with qemu u-boot but only serial console 2019-10-02 08:32:34 mps: yes its documented that qemu on aarch64 does not emulate a video adapter by default 2019-10-02 08:33:05 i didnt test it, but i think -device virtio-gpu-pci should fix it. 2019-10-02 08:33:28 iiuc, you are trying to make it for CI 2019-10-02 08:33:53 ncopa: what i noticed when using gic version 3 is that it boots slower 2019-10-02 08:34:32 and there is not much information to be found on this interrupt controller 2019-10-02 08:36:41 so it took me a few hours to get this far 2019-10-02 08:37:02 and i think it will take me days before getting this far on s390x and power9 2019-10-02 08:37:13 if its even properly supported 2019-10-02 08:43:30 im talking with mort___ here now.... 2019-10-02 08:43:50 ? 2019-10-02 08:44:10 he sits 1-2 meter from me. im in cambridge 2019-10-02 08:44:43 maybe we should move the discussion to #alpine-infra 2019-10-02 08:44:55 ah this is commits 2019-10-02 08:44:57 sure 2019-10-03 07:32:53 clandmeter, can you (or who have access) delete the downlaoded py3-aiofiles-0.4.0.tar.gz fomr distfiles ? 2019-10-03 07:33:04 Appears to be already downloaded in the past... 2019-10-03 07:33:27 but the new one, same version, has a different checksum 2019-10-03 09:54:28 algitbot: retry master 2019-10-03 10:12:12 algitbot: retry master 2019-10-06 21:33:37 mkinitfs:zfs [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130594238 2019-10-06 21:45:04 mkinitfs:zfs [failed] |Rasmus Thomsen| fixup! init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130594687 2019-10-06 21:49:24 mkinitfs:zfs [failed] |Rasmus Thomsen| fixup! fixup! init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130594888 2019-10-06 21:53:45 mkinitfs:zfs [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130595069 2019-10-06 21:56:24 mkinitfs:zfs [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130595164 2019-10-06 21:57:34 mkinitfs:zfs [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130595209 2019-10-07 18:01:46 tcely: regarding !296 yes I missed to upgrade post-upgrade 2019-10-07 18:02:00 wrong channel, sorry 2019-10-07 21:30:50 mkinitfs:zfs [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130814760 2019-10-07 22:03:31 mkinitfs:zfs [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130818291 2019-10-08 08:18:08 mkinitfs:zfs [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130870279 2019-10-08 12:17:48 huh? 2019-10-08 12:18:10 ncopa: ^ could you take a look at this looks like an issue with the previous build 2019-10-08 12:18:28 fails to cleanup the $srcdir 2019-10-08 12:31:22 nmeum: we have a fix for that 2019-10-08 12:31:27 check minio apkbuild 2019-10-08 12:32:23 oh 2019-10-08 12:32:39 just suprised because it didn't happen locally but maybe that's due to abuild rootbld 2019-10-08 12:33:05 thoses files are explicitly set to ro by go 2019-10-08 12:34:19 clandmeter: terraform already has the same cleanup function as minio 2019-10-08 12:34:29 i'll have a look at it after lunch 2019-10-08 12:34:46 thanks 2019-10-08 12:35:12 the entire go rebuild fiasco definitly takes more time than expected 2019-10-08 13:18:26 mkinitfs:zfs [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.com/Cogitri/mkinitfs/builds/130956574 2019-10-09 14:13:40 mkinitfs:master [failed] |Rasmus Thomsen| init: add support for force importing ZFS pool and ZFS native encryption | https://travis-ci.org/alpinelinux/mkinitfs/builds/595641254 2019-10-09 20:38:11 algitbot: retry master 2019-10-10 06:35:33 algitbot: retry master 2019-10-10 07:12:29 Cogitri: i think i know whats wrong 2019-10-10 07:14:57 Oh? :o 2019-10-10 07:15:17 Had hoped it fixed itself over night 😅 2019-10-10 07:15:57 its a computer, it will stay like this forever :p 2019-10-10 07:16:20 there is one little difference between builders and local build 2019-10-10 07:16:53 buidlers dont have a tty 2019-10-10 07:16:58 so there is no stdout 2019-10-10 07:17:08 Huh, but why would this only happen now then? 2019-10-10 07:17:34 Ninja worked well before 2019-10-10 07:17:44 check what has changed in the project 2019-10-10 07:18:08 Did a diff, nothing much that would trigger it 2019-10-10 07:18:25 what happens if you revert? 2019-10-10 07:18:47 Can try on the builders, but ninja debug tells me ninja explain: output build.ninja older than most recent input ../../../../../../../../dev/stdout (1570689391270569359 vs 1570689391677239578) 2019-10-10 07:18:55 which doesn't seem like something a project change would do 2019-10-10 07:19:10 maybe ninja got updated 2019-10-10 07:20:43 Last update was April 26th 2019-10-10 07:20:54 I'll try the diff method, let's see if that does something 2019-10-10 07:22:17 Cogitri: im not sure im right about it, but anytime a local build works and on a builder it doesnt and there is something shown about stdout it smells like this. 2019-10-10 07:24:59 i can also reproduce it locally if i detach the build 2019-10-10 07:25:17 Oh huh 2019-10-10 07:25:27 It's odd that this hasn't occured previously then though :o 2019-10-10 07:25:39 Thanks for looking into it. Not quite sure what to do about it now though 2019-10-10 07:27:46 found out where ninja explain: output build.ninja older than most recent input ../../../../../../../../../../dev/stdout (1570692001367809384 vs 1570692001587806708) 2019-10-10 07:27:47 happens 2019-10-10 07:28:33 i think its meson 2019-10-10 07:28:40 upgraded oct 8 2019-10-10 07:29:38 Ah, possibly 2019-10-10 07:30:19 i havent touched those a lot 2019-10-10 07:30:28 so im not your best resource 2019-10-10 07:30:43 I'll look into it later 2019-10-10 07:58:54 What do you use for detaching? nohup or smth? 2019-10-10 08:01:49 exec 2019-10-10 08:09:36 Ah, easy enough 2019-10-10 08:12:12 Huh, it works with exec 2019-10-10 08:14:15 (for me at least) ^ clandmeter 2019-10-10 08:14:37 how can i reproduce it locally? 2019-10-10 08:26:05 i did it with exec abuild -rf &> /tmp/build2.log 2019-10-10 08:34:11 build-edge-x86_64:~/aports/community/mutter/src/mutter-3.34.1$ sprunge < output/build.ninja 2019-10-10 08:34:11 http://sprunge.us/wECMny 2019-10-10 08:34:31 note that ninja says that /dev/stdout is a build depedency 2019-10-10 08:37:51 build build.ninja: REGENERATE_BUILD ../meson.build ... .../dev/stdout 2019-10-10 09:06:10 Hm, before I go into debugging this, it's not possibly to give the builders the TTY which would fix this, right? 2019-10-10 13:45:09 algitbot: retry master 2019-10-10 13:57:43 algitbot: retry master 2019-10-10 13:57:53 (the real master this time now that mutter has failed) 2019-10-11 20:42:19 <_ikke_> algitbot: retry master 2019-10-11 22:48:50 algitbot: retry master 2019-10-11 22:50:56 algitbot: retry master 2019-10-12 10:13:10 algitbot: retry master 2019-10-12 10:59:43 hm, somehow the builders are currently stack at "Signing the index..." 2019-10-12 10:59:57 or it just takes forever currently for some reason? 2019-10-12 11:07:11 <_ikke_> nmeum: Where do you see that? 2019-10-12 11:08:14 oh 2019-10-12 11:08:16 my fault sorry 2019-10-12 11:08:26 I thought they were stuck because an uploaded message was not output by algibot 2019-10-12 11:08:33 e.g. for kbd 2019-10-12 11:08:42 but that's probably the case because it's currently building something else 2019-10-13 00:05:31 apk-tools:master [errored] |nrybowski| test build | https://travis-ci.org/nrybowski/apk-tools/builds/597126014 2019-10-13 00:22:33 apk-tools:travis [errored] |nrybowski| Attempt to fix travis build | https://travis-ci.org/nrybowski/apk-tools/builds/597130700 2019-10-13 00:39:42 apk-tools:travis [errored] |nrybowski| Attempt to fix travis build x2 | https://travis-ci.org/nrybowski/apk-tools/builds/597133343 2019-10-13 01:02:17 apk-tools:travis [errored] |nrybowski| Attempt to fix travis build x4 | https://travis-ci.org/nrybowski/apk-tools/builds/597137812 2019-10-13 01:17:53 apk-tools:travis [errored] |nrybowski| Attempt to fix travis build x5 | https://travis-ci.org/nrybowski/apk-tools/builds/597139856 2019-10-13 01:29:00 apk-tools:travis [errored] |nrybowski| Quick fix of travis builds - ppc64le still failing | https://travis-ci.org/nrybowski/apk-tools/builds/597142200 2019-10-13 20:09:14 algitbot: retry master 2019-10-14 06:19:40 maxice8: ^ 2019-10-14 06:19:47 thanks 2019-10-14 06:19:55 my pleasure 2019-10-14 08:13:05 algitbot: retry 3.10-stable 2019-10-14 09:03:23 algitbot: retry master 2019-10-14 10:51:02 algitbot: retry master 2019-10-14 12:41:49 that's quite bad... 2019-10-14 12:41:58 kubernetes with 1.13 doesn't work :-/ 2019-10-14 12:42:42 https://github.com/kubernetes/kubernetes/issues/83124 2019-10-14 12:42:48 gotta disable for now.. 2019-10-14 18:59:09 algitbot: retry master 2019-10-15 14:13:42 ncopa: i pushed an updated version of ovmf 2019-10-15 14:13:48 i hope i didnt break things 2019-10-15 14:13:55 it includes aarch64 now 2019-10-15 14:14:11 clandmeter: \o/ 2019-10-15 14:14:59 👊 2019-10-15 17:13:08 apk-tools:master [errored] |nrybowski| Quick fix of travis builds - ppc64le still failing | https://travis-ci.org/nrybowski/apk-tools/builds/598251146 2019-10-15 17:17:03 apk-tools:master [errored] |nrybowski| Adding test case for --ignore flag | https://travis-ci.org/nrybowski/apk-tools/builds/598251865 2019-10-15 17:19:10 apk-tools:travis [passed] |nrybowski| travis fix attempt | https://travis-ci.org/nrybowski/apk-tools/builds/598251534 2019-10-17 07:55:42 <_ikke_> algitbot: retry 3.10-stable 2019-10-17 13:29:45 apk-tools:fix-travis [failed] |nrybowski| Fix travis build by updating alpine-chroot-install to v0.11.0 and Ubuntu distro to bionic | https://travis-ci.org/nrybowski/apk-tools/builds/599157085 2019-10-17 13:35:34 apk-tools:fix-travis [errored] |nrybowski| Fix travis build by updating alpine-chroot-install to v0.11.0 and Ubuntu distro to bionic | https://travis-ci.org/nrybowski/apk-tools/builds/599159709 2019-10-17 14:07:24 algitbot: retry master 2019-10-17 14:08:48 ACTION shakes fist at texlive's build race conditions 2019-10-17 15:57:41 <_ikke_> algitbot: retry master 2019-10-17 16:44:12 algitbot: retry master 2019-10-18 08:05:08 apk-tools:fix-travis [errored] |nrybowski| Fix travis build by updating alpine-chroot-install to v0.11.0 and Ubuntu distro to bionic | https://travis-ci.org/nrybowski/apk-tools/builds/599159709 2019-10-19 11:51:42 algitbot: retry master 2019-10-19 11:56:27 algitbot: retry master 2019-10-19 16:03:33 <_ikke_> wxWidgets wx/version.h file not found in 2019-10-21 13:39:28 algitbot: retry 3.10-stable 2019-10-22 10:15:07 algitbot: retry master 2019-10-23 14:38:26 foobar uploaded :D 2019-10-24 19:08:03 algitbot: retry master 2019-10-24 19:08:28 Damnit rustc 2019-10-24 19:21:37 <_ikke_> Cogitri: ok, now it seems to continue 2019-10-24 19:21:43 <_ikke_> What makes it flaky? 2019-10-24 19:31:20 _ikke_ Our rustc likes to SIGSEGV on projects using gtk-rs since that's such a massive project. Not sure what causes it, Void suffers from the same thing 2019-10-24 19:31:32 <_ikke_> ok 2019-10-26 18:50:30 <_ikke_> algitbot: retry mater 2019-10-26 18:50:32 <_ikke_> algitbot: retry master 2019-10-26 18:58:53 <_ikke_> algitbot: retry master 2019-10-26 19:01:34 <_ikke_> Strange, it seems like it still uses busybox cpio after installing the cpio package 2019-10-26 19:39:35 <_ikke_> algitbot: retry master 2019-10-26 19:42:27 <_ikke_> algitbot: retry master 2019-10-26 20:48:42 <_ikke_> algitbot: retry master 2019-10-26 21:01:04 <_ikke_> algitbot: retry master 2019-10-29 09:53:03 algitbot: retry master 2019-10-29 11:32:30 <_ikke_> maxice8: re nmap: http://tpaste.us/5Ev7 2019-10-29 11:32:39 <_ikke_> algitbot: retry master 2019-10-29 11:32:45 <_ikke_> algitbot: retry 3.9 2019-10-29 11:32:48 <_ikke_> algitbot: retry 3.9-stable 2019-10-29 11:33:53 <_ikke_> maxice8: hmm, ok, 2nd time it succeeded