#archlinux32 | Logs for 2023-06-21

Back
[00:44:10] -!- KitsuWhooa has quit [Server closed connection]
[00:44:37] -!- KitsuWhooa has joined #archlinux32
[07:20:56] -!- drathir_tor has quit [Ping timeout: 240 seconds]
[07:27:48] -!- drathir_tor has joined #archlinux32
[09:20:52] <KitsuWhooa> Uh, we might have an issue
[09:21:05] <KitsuWhooa> I just got 10:34:52 <buildmaster> Your buildslave "build-tasossah-2" uploaded a package with a not fully-trusted signature:
[09:22:55] <KitsuWhooa> Also, I wonder if something is wrong. Most of the time my builders stay idle because they're waiting for the build lock
[11:36:28] -!- abaumann has joined #archlinux32
[11:36:28] <buildmaster> Hi abaumann!
[11:36:28] <buildmaster> !rq abaumann
[11:36:29] <phrik> buildmaster: <abaumann> it's modern software in the end.. it _has_ to compile that long in order to get good.. ;-)
[11:36:32] <abaumann> eurobuild3-i486     abaumann     i686/firefox
[11:36:41] <abaumann> a i686 package built on a 486 VM, this is not good.
[11:36:51] -!- abaumann has quit [Client Quit]
[14:14:11] -!- abaumann has joined #archlinux32
[14:14:12] <buildmaster> Hi abaumann!
[14:14:12] <buildmaster> !rq abaumann
[14:14:12] <phrik> buildmaster: <abaumann> We should really not build a package more than 99 times. ;-)
[14:20:45] <abaumann> deep42thought: yes, this is ./schedule-for-rebuild -w -p '^openssh$':
[14:20:50] <abaumann> openssh 3ba51e1910a349b9ec038b8b1821d3b04ed4b3c9 413119cea02c94719617ae640f22435423dc9efb core
[14:20:53] <abaumann> openssh afeaefd117434d7f0673a640d01b4f82688c5f41 413119cea02c94719617ae640f22435423dc9efb core
[14:20:56] <abaumann> gzip: stdin: not in gzip format
[14:20:58] <abaumann> tar: Child returned status 1
[14:21:01] <abaumann> tar: Error is not recoverable: exiting now
[14:21:04] <abaumann> ==> ERROR: pkgrel is not allowed to be empty.
[14:21:06] <abaumann> ==> ERROR: pkgver is not allowed to be empty.
[14:21:08] <abaumann> ==> ERROR: pkgname is not allowed to be empty.
[14:21:11] <abaumann> "make_source_info" had empty output - eh, what?
[14:21:13] <abaumann> basically, it doens't work :-)
[14:21:16] <abaumann> maybe there is a strange package in the pipe?
[14:21:19] <abaumann> I do a './get-package-updates -w'
[14:21:27] <abaumann> to see, what gets into the buildmaster
[14:21:55] <abaumann> also: caveat, openssh has an ABI mismatch on pentium4 with openssl-3.1, stick to openssl-3.0 for now (or you cannot log in). :-)
[14:23:09] <KitsuWhooa> one of the builders is asking for a sudo password
[14:23:14] <KitsuWhooa> time to edit sudoers again
[14:23:25] <KitsuWhooa> Also, can you please block pdnsd from being built?
[14:23:36] <KitsuWhooa> it keeps trying to build, but the source is not available and blocks forever
[14:24:03] <KitsuWhooa> http://members.home.nl
[14:24:06] <KitsuWhooa> unless it's only for me
[14:26:34] <abaumann> ah, I'll have a look at pdnsd
[14:27:21] <abaumann> BTW. Did you update you keys on the build machines? Because you might have an unsigned private package signing key there (hence the buildmaster rejects your package being not trustworthingly signed)
[14:27:41] <KitsuWhooa> I ran pacman -Syu, yes
[14:27:48] <KitsuWhooa> it only sent that message once
[14:28:01] <KitsuWhooa> unless I have to do something else
[14:29:27] <abaumann> ah, then this is solved.
[14:30:08] <KitsuWhooa> I mentioned it yesterday, but is it normal for the builders to sit idle for this long trying to lock the build list?
[14:30:17] <abaumann> bill-auger: we are only signing arch32 packages with signing packages from the arch32-keyring. the arch-keyring from upstream is completely optional and we shipped it, so that people already have the right keys in case they do something with upstream packages on a 32-bit system.
[14:30:29] <abaumann> no, this is not normal
[14:30:38] <abaumann> there are several issues which can cause this:
[14:30:58] <abaumann> a) the buildmaster is unsane (but then you see in #archlinux32-devops a message like "buildmaster is unsane")
[14:31:19] <abaumann> b) get-package-updates runs for a very long time, can happen, if a lot changes upstream
[14:31:28] <KitsuWhooa> ah, looks like that's what is going on now
[14:31:37] <KitsuWhooa> get-package-upd 3079653 master 9w REG 9,2 0 175898683 /home/master/builder/work/build-list.lock
[14:31:37] <abaumann> c) a package of the toolchain is being built for that subarchitecture, then all other packages are blocked from building
[14:32:00] <abaumann> d) bugs, like a locking problem (but this is quite rare, had it once or twice)
[14:33:16] <KitsuWhooa> thanks for the explanation
[14:34:03] <abaumann> there might also be an e) or so.. ;-)
[14:34:29] <abaumann> uh. my mirror got stuck..
[15:38:51] <abaumann> so, I found a working download location for pdnsd, it doesn't build though..
[15:41:50] <KitsuWhooa> how did upstream build it :p
[15:42:30] <abaumann> quite a while ago.. we are currently rebuilding everything, so some failures of this kind are to be expected.
[15:42:31] <KitsuWhooa> > Build Date: 2020-05-25 14:53 UTC
[15:43:01] <abaumann> the software is _way_ old, so. this is a recent build date. :-)
[15:44:07] <abaumann> aha, get-package-updates finished, fine.
[15:44:12] <abaumann> no errors there.
[15:46:20] <KitsuWhooa> one of my builders is still complaining about not being able to lock the build list
[15:46:23] <KitsuWhooa> the other two started building
[15:46:42] <KitsuWhooa> oh, another one built a package and now is waiting again
[15:50:21] <abaumann> there more build clients, the more waiting you get.
[15:50:22] <abaumann> probably.
[15:50:35] <abaumann> eurobuild6-3 is also idle ATM
[15:50:44] <abaumann> eurobuild6-8-i486 too
[15:52:09] <abaumann> ATM I'm running a get-pacakge-updates on the master, so the build slaves are finished their current package, but they don't get a new one handed out.
[15:52:47] <abaumann> what worries me more is that again I get a lot of updates with get-package-updates - ah, all packages on the blacklist..
[15:53:26] <abaumann> so, this happens on every update. We should maybe not reconsider packages on the blacklist, but then we must explicitely enable them again for building when we want to reactivate them
[16:00:59] <KitsuWhooa> Yeah there is a lot of waiting
[16:34:32] <abaumann> motto of myu life ;-)
[16:48:50] -!- abaumann has quit [Quit: leaving]
[17:44:36] <bill-auger> i was only wondering why pacman asked to import KitsuWhooa's key, ratherthan simply installing it from hte keyring
[17:44:36] <bill-auger> i gave the archlinux-keyring only as an example - i dont remember ever being asked to import a key which is (or should be) in the keyring
[22:21:07] -!- socksinspace has quit [Ping timeout: 260 seconds]
[22:21:59] -!- socksinspace has joined #archlinux32