#archlinux32 | Logs for 2019-08-17

Back
[00:00:38] -!- MrBIOS has joined #archlinux32
[01:21:19] -!- MrBIOS has quit [Quit: MrBIOS]
[04:23:13] <thePiGrepper> hi, the pentium4 is acting kinda weird. for instance, the linux package is at 5.2.6, however linux-headers is at 5.2.8. I think something like that happened or have been happening for the last couple of weeks.
[04:47:58] -!- thePiGrepper has quit [Ping timeout: 245 seconds]
[06:19:50] -!- soalokin has quit [Read error: Connection reset by peer]
[06:20:07] -!- soalokin has joined #archlinux32
[06:50:28] -!- soalokin has quit [Quit: Died!]
[07:31:44] -!- MrBIOS has joined #archlinux32
[07:44:46] -!- quequotion has joined #archlinux32
[08:14:48] -!- abaumann has joined #archlinux32
[08:15:06] -!- abaumann has parted #archlinux32
[08:15:09] -!- abaumann has joined #archlinux32
[08:15:42] <abaumann> things are heavily borked in the buildmaster database. It looks like publishing packages doesn't work at all..
[08:15:54] <abaumann> https://buildmaster.archlinux32.org
[08:15:56] <phrik> Title: result of archlinux32 build master's sanity checksanity of the buildmaster's mysql database (at buildmaster.archlinux32.org)
[08:16:30] -!- abaumann has quit [Client Quit]
[08:18:23] -!- abaumann has joined #archlinux32
[08:18:31] <abaumann> i486: broken dependencies:
[08:18:32] <abaumann> warning: cannot resolve "gcc-libs=9.1.0-2.3", a dependency of "gcc"
[08:18:33] <abaumann> warning: cannot resolve "device-mapper>=2.02.185", a dependency of "lvm2"
[08:18:33] <abaumann> :: The following packages cannot be upgraded due to unresolvable dependencies: gcc lvm2
[08:18:40] <abaumann> so this also seems not to work
[08:20:18] -!- abaumann has quit [Client Quit]
[08:21:35] -!- abaumann has joined #archlinux32
[08:21:49] -!- quequotion has quit [Quit: quequotion]
[08:21:51] <abaumann> mmh. the errors in the database are pentium4 only
[08:23:01] <abaumann> but the broken dependencies are 486 only
[08:23:12] <abaumann> or maybe not, I just have too few software installed.
[08:25:47] <abaumann> a52dec-0.7.4-10.2-pentium4 is in extra (on the mirror), but a52dec-0.7.4-10.1-pentium4 in the database
[08:25:56] <abaumann> so the last database update went wrong
[08:27:04] <abaumann> mmh. where are those mysql scripts
[08:33:45] <abaumann> aha. the pool contains akonadi-notes-19.04.2-1.0 and akonadi-notes-19.04.3-1.0, extra points to pentium4/extra/akonadi-notes-19.04.3-1.0-pentium4
[08:33:59] <abaumann> so, also deleting old packages in the pool failed
[08:34:26] <abaumann> sadly I didn't execute the db-update script into a log, nor does it write one
[08:38:23] <abaumann> the problem is also: file operations and database operations must succeed or fail in the same transaction. If a deadlock hits in the middle of a file move on the mirror, then we end up in this mess.
[08:38:54] <abaumann> I would also opt to do more small selects and steps and not move 1000nds of packages at once. So the mess (if it happens) affects just one package.
[08:59:04] -!- abaumann has quit [Quit: leaving]
[09:24:42] -!- abaumann has joined #archlinux32
[10:07:10] -!- MrBIOS has quit [Quit: MrBIOS]
[10:16:25] -!- davor has quit [Ping timeout: 246 seconds]
[10:17:01] -!- davor has joined #archlinux32
[10:18:47] <abaumann> aha. seems quite simple, I just need a script grepping the red lines and remove the files from the master mirror..
[10:27:30] <buildmaster> Hi abaumann!
[10:27:30] <buildmaster> !rq abaumann
[10:27:31] <phrik> buildmaster: <abaumann> The next step will be to distribute code to neighbour servers or use the Amazon cluster in the build script to run tests..
[10:27:35] -!- deep42thought has joined #archlinux32
[10:27:35] <buildmaster> Hi deep42thought!
[10:27:35] <buildmaster> !rq deep42thought
[10:27:36] <phrik> buildmaster: <deep42thought> good software is like a good meal: you cannot expect to get something really good if it's done in 5 minutes ...
[10:27:43] <deep42thought> Hi abaumann, sry, was hiking
[10:27:51] <deep42thought> ... will fix this mess shortly
[10:28:06] <deep42thought> and: you _do_ have access to the master mirror (via the buildmaster's sftp access)
[10:29:05] <abaumann> hi deep42thought
[10:29:08] <abaumann> no problem
[10:29:15] <abaumann> I'm on my way fixing the problem
[10:29:27] <abaumann> found the sftp helper in common-functions
[10:29:51] <deep42thought> $master_mirror_sftp_command
[10:29:59] <deep42thought> $master_mirror_sftp_directory or so
[10:30:25] * abaumann is just babbling in the chat to leave a trace of the disaster for deep42thought :-)
[10:30:36] <deep42thought> :-D
[10:30:38] <abaumann> there is even a safe ftp wrapper function :-)
[10:30:44] <deep42thought> yeah
[10:30:50] <deep42thought> but don't use that from the shell
[10:30:58] <deep42thought> it tries the command over and over (upto 10 times)
[10:31:07] <deep42thought> so if you have some typo => fun!
[10:31:07] <abaumann> if it fails, yeas.
[10:31:19] * abaumann doesn't make typos ;-)
[10:31:29] <abaumann> so, let's sanity check
[10:31:37] <abaumann> https://bbs.archlinux32.org
[10:31:39] <phrik> Title: New massive update broke system (pentium4)? / Installation / Arch Linux 32 Forums (at bbs.archlinux32.org)
[10:31:42] <abaumann> this sounds more alarming.
[10:31:56] <abaumann> though I updated a pentium4 machine and I didn't encounter those issues..
[10:32:31] * buildmaster resumes sanity.
[10:32:39] <abaumann> \o/
[10:33:59] <deep42thought> :-)
[10:34:15] * abaumann can not only destroy things, but also fix them (sometimes)
[10:34:21] <deep42thought> :-)
[10:34:29] <deep42thought> fix by deletion
[10:34:42] <abaumann> I really wonder, what went wrong yesterday
[10:35:05] <deep42thought> I think, something about the "detect older package" logic is wrong (but why?)
[10:35:25] <abaumann> a hint could be, that only pentium4 packages where affected
[10:35:42] <deep42thought> this can mean, that only pentium4 packages were moved at the time
[10:36:02] <abaumann> ah. yes.
[10:36:56] <deep42thought> might have been a connectivity issue, too
[10:37:09] <deep42thought> e.g. the connection died (more than 10 times) while things would have been moved
[10:37:25] <abaumann> that's quite likely
[10:40:21] <abaumann> ok sddm works on pentium4 only
[10:41:22] <abaumann> lightdm dito
[10:41:32] <abaumann> most likely again SSE2 micro-optimizations in gtk or qt
[10:41:44] <abaumann> xfce breaks, lxde works
[10:42:34] <abaumann> I thought yesterday: we can either be a stable distribution (not publishing packages for months) or be bleeding edge.. :-)
[10:42:39] <abaumann> so bleeding it is..
[10:44:48] <buildmaster> i686/grafana is broken (says nlopc46): https://archlinux32.org
[10:46:49] <abaumann> Aug 17 10:41:09 arch32-stable systemd[1]: lightdm.service: Scheduled restart job, restart counter is at 5.
[10:46:52] <abaumann> Aug 17 10:41:09 arch32-stable systemd[1]: Stopped Light Display Manager.
[10:46:55] <abaumann> Aug 17 10:41:09 arch32-stable audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=lightdm comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[10:47:09] * abaumann took the wrong window for copy-pasting
[10:47:29] <deep42thought> too many clipboards?
[10:48:02] <abaumann> Aug 17 10:41:09 arch32-stable systemd[1]: lightdm.service: Scheduled restart job, restart counter is at 5.
[10:48:04] * deep42thought reads through 200 new emails ...
[10:48:05] <abaumann> Aug 17 10:41:09 arch32-stable systemd[1]: Stopped Light Display Manager.
[10:48:06] <abaumann> darn
[10:48:08] <abaumann> Aug 17 10:41:09 arch32-stable audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=lightdm comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[10:48:13] <abaumann> Aug 17 10:41:09 arch32-stable systemd[1]: lightdm.service: Scheduled restart job, restart counter is at 5.
[10:48:16] <abaumann> Aug 17 10:41:09 arch32-stable systemd[1]: Stopped Light Display Manager.
[10:48:19] <abaumann> Aug 17 10:41:09 arch32-stable audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=lightdm comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
[10:48:32] <abaumann> Copy pasting from vnc via X server doesn't work
[10:49:15] <abaumann> kde doesn't work, as usual
[10:49:33] <abaumann> o wait. it's just that slow to start up :-)
[10:49:40] <deep42thought> :-D
[10:49:54] <abaumann> at least on pentium4 it works. i686 has again micro-optimizations everywhere
[10:50:26] * abaumann thinks it's really important for the KDE spashscreen to have SSE2 commands.. :->
[10:50:46] <deep42thought> !grab abaumann
[10:50:47] <phrik> deep42thought: Tada!
[10:53:54] <abaumann> it's really fun to test all window managers and login managers for all platforms..
[10:54:01] <deep42thought> :-)
[10:54:05] <abaumann> ..luckily there are none on i486..
[10:54:14] <deep42thought> you can measure some timings, too and make a nice table
[10:54:37] <deep42thought> "cannot recommend gnome - start .1s slower than kde"
[10:55:09] <abaumann> I thought more about a table like: on i686 use xfce or lxde, don't try KDE, Gnome or Mate
[10:55:30] <deep42thought> that would be helpful, too :-)
[10:55:43] <abaumann> oh. hang on. mate works on i686
[10:55:55] <buildmaster> pentium4/grafana is broken (says buildknecht): https://archlinux32.org
[10:56:00] <abaumann> who wants gnome and kde anyway..
[10:56:08] <abaumann> kde even copies security bugs from Windows :->
[10:56:28] * abaumann refers to KDE's latest disaster with clickable infenctions on the desktop
[11:01:09] <abaumann> ldd /usr/bin/xfce4-session | grep not libxfconf-0.so.2 => not found, xfconf was published, xfce4-session is linked against an older version. I trigger a rebuild..
[11:05:10] <abaumann> mmh device-mapper and gcc-libs where not moved on i486, I moved them now manually.
[11:06:59] * deep42thought waits with updating all his router a little longer, then
[11:07:18] * abaumann thinks this is a wise idea :-)
[11:07:33] <abaumann> I'm just updating some i486 test machines, let's see..
[11:07:42] <abaumann> ..but I'm not running DNS and stuff on them..
[11:07:51] <deep42thought> dns is not *that* importatnt
[11:07:59] <abaumann> :-)
[11:08:02] <deep42thought> I know (some) ip addresses of the outside world
[11:08:21] <abaumann> ipv4 addresses are not that many..easy to remember.. ;-)
[11:08:29] <deep42thought> !grab abaumann
[11:08:30] <phrik> deep42thought: 🎉
[11:08:33] <deep42thought> you know them all?
[11:08:34] -!- Dimtree has joined #archlinux32
[11:09:02] <abaumann> almost. ;-)
[11:09:38] <abaumann> https://srv0.tyzoid.com:8006 has an expired certificate
[11:11:21] <abaumann> ok, the i486 build slaves survived a reboot
[11:17:15] <deep42thought> on i486 new linux-api-headers but no new linux?
[11:17:17] <deep42thought> is this right?
[11:17:35] Alina-malina_ is now known as Alina-malina
[11:17:40] <abaumann> mmh.
[11:18:12] <abaumann> thePiGrepper said something similar also for pentium4
[11:20:36] * deep42thought updates anyways
[11:32:21] <abaumann> ok, my AMD machine (i486 branch) survived the update
[11:32:31] <abaumann> now my Alix..
[11:32:37] <deep42thought> my router is still generating its ramdisk
[11:32:47] <abaumann> lol
[11:32:48] <abaumann> yeah
[11:33:25] <abaumann> I always wondered if all those new compression algorighms are just faster on modern machines..
[11:33:45] <deep42thought> compression on gpu ;-)
[11:34:10] <abaumann> I'm sure some SSExx stuff is quite handy for compression
[11:35:38] <deep42thought> damn, compressing the ramdisk brings down the dns server O.o
[11:36:01] <deep42thought> ... but as long as I'm staying online, it's still routing :-D
[11:36:06] <abaumann> TMPDIR=/somewhere_on_the_SD_card ?
[11:36:34] <deep42thought> sure
[11:36:43] <abaumann> mmh
[11:36:45] <deep42thought> let me check (once I can log in)
[11:39:05] <deep42thought> TMPDIR is unset and /tmp is a tmpfs
[11:39:08] <deep42thought> so: should be fine
[11:39:25] * abaumann reboots his alix
[11:40:54] <abaumann> survived :-)
[11:46:48] <buildmaster> i486/spirv-llvm-translator is broken (says eurobuild6-7-i486): https://archlinux32.org
[12:32:06] <abaumann> oh well, almost all Xfce packages are linked against libxfconf-0.so.2, so I trigger a bigger rebuild.. we have enough build slaves.. :-)
[12:40:04] <buildmaster> i486/xfce4-taskmanager is broken (says eurobuild6-7-i486): https://archlinux32.org
[12:54:23] -!- deep42thought has quit [Quit: Leaving.]
[13:48:17] * abaumann hopes this is not a quite because of a lost router..
[13:48:20] <abaumann> *quit
[13:55:41] <buildmaster> i686/unison is broken (says eurobuild6-6): https://archlinux32.org
[13:55:47] -!- quequotion has joined #archlinux32
[13:57:18] <buildmaster> pentium4/unison is broken (says eurobuild6-4): https://archlinux32.org
[14:07:36] -!- thePiGrepper has joined #archlinux32
[14:24:13] -!- thePiGrepper has quit [Ping timeout: 245 seconds]
[14:50:03] -!- thePiGrepper has joined #archlinux32
[15:33:54] <abaumann> mmh. I priorized xfce4 and unison/ocaml stuff and everything is still on 0?
[15:34:05] <abaumann> most likely something is blocking the scheduled rebuilds.
[16:02:44] <buildmaster> i486/cups are broken (says eurobuild6-7-i486): https://archlinux32.org
[16:03:29] <buildmaster> i686/virtualbox-modules-arch is broken (says nlopc46): https://archlinux32.org
[16:06:20] <buildmaster> i486/cups-filters are broken (says eurobuild6-7-i486): https://archlinux32.org
[16:06:52] <buildmaster> pentium4/virtualbox-modules-arch is broken (says eurobuild6-5): https://archlinux32.org
[16:07:41] -!- quequotion has quit [Quit: quequotion]
[16:24:58] <buildmaster> pentium4/linux-zen is broken (says eurobuild6-5): https://archlinux32.org
[16:34:49] <buildmaster> i686/linux-zen is broken (says eurobuild6-3): https://archlinux32.org
[16:35:09] <buildmaster> i486/js60 is broken (says eurobuild6-7-i486): https://archlinux32.org
[16:45:34] <buildmaster> i486/ghostscript is broken (says eurobuild6-7-i486): https://archlinux32.org
[16:50:39] <buildmaster> i486/virtualbox-modules-arch is broken (says eurobuild6-7-i486): https://archlinux32.org
[16:59:09] <abaumann> too many broken zens..
[16:59:49] <buildmaster> i686/telepathy-accounts-signon is broken (says eurobuild6-5): https://archlinux32.org
[17:06:31] <abaumann> this is weird ./seed-build-list -w -p '^telepathy-accounts-signon$'
[17:06:31] <abaumann> ./prioritize-build-list -w <(printf '^telepathy-accounts-signon$\n' )
[17:06:41] <abaumann> telepathy-accounts-signon builds manually
[17:06:51] * abaumann has problems copy-pasting today
[17:08:59] <buildmaster> pentium4/telepathy-accounts-signon is broken (says eurobuild6-5): https://archlinux32.org
[17:10:12] <buildmaster> i686/telepathy-accounts-signon is broken (says rechenknecht): https://archlinux32.org
[17:42:48] <buildmaster> i686/kde-cli-tools are broken (says eurobuild6-5): https://archlinux32.org
[17:59:20] <buildmaster> i486/io is broken (says eurobuild6-7-i486): https://archlinux32.org
[18:00:38] <abaumann> warning: dependency cycle detected:
[18:00:38] <abaumann> warning: mesa will be installed before its libglvnd dependency
[18:00:38] <abaumann> warning: dependency cycle detected:
[18:00:38] <abaumann> warning: harfbuzz will be installed before its freetype2 dependency
[18:00:41] <abaumann> warning: dependency cycle detected:
[18:00:43] <abaumann> warning: usbmuxd will be installed before its libimobiledevice dependency
[18:00:46] <abaumann> warning: dependency cycle detected:
[18:00:48] <abaumann> warning: phonon-qt5-gstreamer will be installed before its phonon-qt5 dependency
[18:00:51] <abaumann> warning: dependency cycle detected:
[18:00:54] <abaumann> warning: x264 will be installed before its ffmpeg dependency
[18:00:56] <abaumann> kde-cli-tools
[18:00:58] <abaumann> nice
[18:02:47] <abaumann> this clusterf*k is in staging..
[18:02:50] <abaumann> ..I wonder.
[18:04:27] <abaumann> https://buildmaster.archlinux32.org
[18:04:29] <phrik> Title: Munin :: buildmaster :: buildmaster (at buildmaster.archlinux32.org)
[18:04:43] <abaumann> aha, the run times of db-update are much lower now, after so many packages have been pushed.
[18:04:55] <abaumann> most likely it's not a good idea to let packages accumulate too long
[18:19:39] -!- bill-auger has quit [Remote host closed the connection]
[18:22:50] <abaumann> mmh systemd-resolved is broken in testing but works on staging - the systemd version is 242.84-2.0 on both..
[18:28:18] <abaumann> now I have to read systemd-resolved to see, what binaries it's calling or what libraries could break ..
[18:32:42] <abaumann> starting /usr/lib/systemd/systemd-resolved by hand works
[18:37:23] <buildmaster> i686/massif-visualizer is broken (says rechenknecht): https://archlinux32.org
[18:37:57] <abaumann> Using system hostname 'arch32-testing'.
[18:37:57] <abaumann> Failed to listen on UDP socket 127.0.0.53:53: Operation not permitted.
[18:37:57] <abaumann> Turning off local DNS stub support.
[18:37:57] <abaumann> Failed to drop remaining caps: Operation not permitted
[18:38:02] <abaumann> running as systemd-resolved
[18:38:04] <abaumann> yeah
[18:39:49] <abaumann> libcap and libcap-ng are the same
[18:40:54] <abaumann> ah.
[18:40:59] <abaumann> now. I cannot do this.
[18:41:14] <abaumann> systemd is starting the service as root, then it drops privileges to systemd-resolved
[18:41:23] <abaumann> this makes testing failure so much easier :->
[18:50:38] <abaumann> systemd.debug-shell=1, very handy, a insecure shell backdoor to debug systemd issues
[18:50:46] <abaumann> on virtual console 9
[18:53:22] -!- infides has joined #archlinux32
[19:51:09] <abaumann> aha. the blue curve is on 2802
[19:51:11] <abaumann> https://archlinux32.org
[19:51:12] <phrik> Title: Buildmaster for Archlinux32 packages (i686, pentium4, any) (at archlinux32.org)
[19:51:27] <abaumann> agressive unblocking of failing packages is the clue.
[19:51:54] <abaumann> sadly the next one is now librsvg with rust issues *urgh*
[19:55:21] -!- buildmaster has quit [Remote host closed the connection]
[19:56:46] -!- buildmaster has joined #archlinux32
[19:56:46] <buildmaster> !rq buildmaster
[19:56:47] <phrik> buildmaster: <buildmaster> I might be insane, but never confused ... ;-)
[19:59:31] <buildmaster> i686/kalzium is broken (says eurobuild6-5): https://archlinux32.org
[20:01:53] <buildmaster> i686/cantor is broken (says eurobuild6-3): https://archlinux32.org
[20:06:07] -!- abaumann has quit [Quit: leaving]
[20:54:16] -!- bill-auger has joined #archlinux32
[21:00:51] <buildmaster> pentium4/broadcom-wl is broken (says nlopc46): https://archlinux32.org
[21:10:27] -!- thePiGrepper has quit [Ping timeout: 245 seconds]
[21:24:00] <buildmaster> pentium4/nvidia-390xx is broken (says buildknecht2): https://archlinux32.org
[21:31:43] -!- deep42thought has joined #archlinux32
[21:31:44] <buildmaster> Hi deep42thought!
[21:31:44] <buildmaster> !rq deep42thought
[21:31:45] <phrik> buildmaster: <deep42thought> my supervisor when I made my first shot with the laser: He handed me the remote and said: "Hey Erich, want to destroy a 20 Mio€ device?"
[21:32:42] <deep42thought> abaumann: we could automatize the aggressive-unblock part (I assume you meant "aggressively force build packages with dependencies which are currently broken"): we could just ignore broken dependencies at some point (when they failed several times to build)
[21:33:04] <deep42thought> but we would need to keep track of those packages somehow and reschedule them later again
[21:36:41] <deep42thought> the rescheduling is the main reason, why I did not yet implement something like that
[21:38:05] <deep42thought> in theory, we do have an auto-rescheduler which should exactly catch those cases, otoh it's currently disabled, because it seemed to reschedule too much (e.g. all of python was regularly rescheduled :-D)
[21:39:21] <buildmaster> pentium4/wire-desktop is broken (says eurobuild6-2): https://archlinux32.org
[21:39:40] -!- nit-picker has joined #archlinux32
[21:42:16] <buildmaster> pentium4/intel-opencl-clang is broken (says rechenknecht): https://archlinux32.org
[22:14:01] -!- deep42thought has quit [Quit: Leaving.]
[22:18:59] <buildmaster> pentium4/prometheus are broken (says eurobuild6-5): https://archlinux32.org
[22:24:09] <buildmaster> i686/linux is broken (says eurobuild6-3): https://archlinux32.org
[22:34:51] <buildmaster> i686/prometheus are broken (says eurobuild6-6): https://archlinux32.org
[23:03:44] <buildmaster> pentium4/rust is broken (says nlopc46): https://archlinux32.org
[23:49:40] -!- nit-picker has quit [Remote host closed the connection]
[23:50:01] -!- nit-picker has joined #archlinux32
[23:57:24] -!- infides has quit [Ping timeout: 272 seconds]