#archlinux32 | Logs for 2019-07-24

[00:07:26] <buildmaster> i686/qtpass are broken (says eurobuild6-2): https://archlinux32.org
[00:14:41] <buildmaster> i686/librecad is broken (says nlopc46): https://archlinux32.org
[01:47:09] -!- DCyrax has quit [Remote host closed the connection]
[01:47:35] -!- DCyrax has joined #archlinux32
[02:14:11] -!- thePiGrepper has joined #archlinux32
[02:51:44] <buildmaster> pentium4/vegeta is broken (says eurobuild3): https://archlinux32.org
[03:08:36] <buildmaster> i686/vegeta is broken (says eurobuild6-6): https://archlinux32.org
[03:18:23] <buildmaster> i686/ntk is broken (says buildknecht2): https://archlinux32.org
[04:53:45] <buildmaster> i686/cogl is broken (says rechenknecht): https://archlinux32.org
[06:38:00] -!- slacka123 has quit [Remote host closed the connection]
[06:51:14] -!- thePiGrepper has quit [Ping timeout: 248 seconds]
[06:54:41] <buildmaster> i686/plasma-maliit-framework is broken (says eurobuild6-5): https://archlinux32.org
[07:05:03] -!- titus_livius has quit [Remote host closed the connection]
[07:05:46] <girls> uagh, hopefully I didn't just strangle return-assignment :-/
[07:08:20] -!- titus_livius has joined #archlinux32
[07:12:34] <girls> looks good *phew*
[10:31:37] -!- chub has joined #archlinux32
[10:41:00] <chub> hello, anybody got problems with systemd-netword and bridge? my bridge is not starting up and now stuck in "configuring"
[10:42:35] <chub> systemd-networkd[465]: br0: Could not bring up interface: Invalid argument
[10:43:42] <chub> all I have to do manually each time the system restarts, I have to type in ip link set br0 up, and everything works, the IP is there. so I don't know why the interface is not able to go "up"
[10:44:11] <chub> and I don't seem to be the only one
[10:44:14] <chub> https://bbs.archlinux.org
[10:44:15] <phrik> Title: SOLVED: Bridge config with systemd-networkd / Networking, Server, and Protection / Arch Linux Forums (at bbs.archlinux.org)
[10:45:58] <girls> chub: I think, abaumann mentioned something similar (not for a bridge, though)
[10:46:25] <chub> https://bugs.archlinux.org
[10:46:27] <phrik> Title: FS#62347 : [systemd] DNS lookup broken after boot until restarting networkd after 242.0-1 update (at bugs.archlinux.org)
[10:46:30] <girls> https://mirror.archlinux32.org
[10:46:32] <phrik> Title: #archlinux32 | Logs for 2019-07-18 (at mirror.archlinux32.org)
[10:47:23] <chub> thank you girls
[10:48:53] <chub> alright, so I read the logs, but I guess there is no resolve yet?
[10:49:21] <girls> at least none which I'm aware of
[10:50:09] <girls> and we're beyond systemd 242.16 - so the "solution" from upstream arch also does not apply :-(
[10:50:39] <chub> ya
[10:50:55] <chub> I'm stumped, so is there maybe a way to run /etc/local.d scripts?
[10:51:22] <chub> I'm having power outages at home, and this is how the issue was highlighted to me
[10:51:26] <girls> you could set up a custom service
[10:51:42] <girls> systemd-service, that is
[10:52:54] <chub> I guess something like this: https://arashmilani.com
[10:52:56] <phrik> Title: Run a script during Arch Linux boot process (at arashmilani.com)
[10:53:17] <girls> yes
[10:53:57] <girls> you can fiddle with the dependencies of that service file to run it at the point, where you *can* enable the network interface and before it is needed by something else
[10:55:10] <chub> now to get my Google wording right
[10:58:47] <chub> https://superuser.com
[10:58:48] <phrik> Title: How do you make a systemd service as the last service on boot? - Super User (at superuser.com)
[10:59:07] <girls> you have no services, that require the network to be up?
[10:59:27] <girls> I'd try "after=network-pre.target" and "before=network-online.target"
[10:59:48] <girls> plus probably some "Wants=..." (have a look at `man systemd.service`)
[11:04:00] -!- panosdev has joined #archlinux32
[11:05:19] <chub> I have other networks that happens to start up so I guess those services are happy
[11:05:34] <girls> these are targets, not services
[11:05:59] <chub> girls: sorry, I guess I'm not familiar with systemd as I would like to think, so what would that imply?
[11:06:38] <chub> I have a mobile wifi device that is started up, but that happens because of dhcpcd hook scripts that gets the interface running
[11:06:41] <girls> the idea is, that any service which needs network (e.g. sshd) says "Wants=network-online.target" and "After=network-online.target", so your setup to bring up the network gets scheduled (first entry) and is executed before sshd (second entry)
[11:07:22] <girls> and you schedule whatever service is necessary to bring up the network as "before=network-online.target" and "wantedby=network-online.target" (I think)
[11:07:31] <chub> well, if my mobile wifi that is a usb ethernet interface, and is up, then that condition would suffice?
[11:07:57] <girls> the target is only considered started if *all* preconditions are met
[11:08:04] <girls> e.g. all your setup services have run
[11:08:06] <chub> I prefer my services to run regardless of which interface is up
[11:08:22] <chub> because some interfaces might take a while to get up and running
[11:08:33] <girls> and services which require them, should wait
[11:08:47] <chub> I understand
[11:09:01] <girls> ah, you mean, you don't need *all* interfaces to come?
[11:09:15] <chub> yes
[11:10:06] <girls> in that case, I'd suggest your original approach of bringing the slowest/least important network up after everything else (e.g. without condition "before=network-online.target")
[11:10:25] <chub> okay
[11:10:49] <chub> how about the example I'm using : https://superuser.com
[11:10:51] <phrik> Title: How do you make a systemd service as the last service on boot? - Super User (at superuser.com)
[11:11:04] <chub> where I simply set [Service] with Type=idle
[11:11:22] <chub> seems very similar to a local.rc or local.d behaviour
[11:11:24] <girls> would work too, but it looks too cautious to me
[11:11:40] <girls> you don't want to wait *starting* that device until everything else is done
[11:12:04] <girls> you simply want, that nothing waits *for your service*
[11:12:19] <girls> which you get already without the Type=idle
[11:12:34] <chub> well I think I have a chicken and egg problem maybe, because my custom script needs to run, but network.br0 is never going to reach "running" because it is stuck in configuring
[11:12:48] <chub> so if I use after? it will never reach that state?
[11:13:08] <girls> you need to tell whomever is configuring network.br0 not to do so - and do everything in your custom script
[11:13:30] <chub> so you are suggesting I change the configuration
[11:13:55] <girls> e.g. my approach is: systemd-standard-script-for-task-x does not work on my setup - so I disable it and replace it with a custom script, that *does* work on my setup
[11:14:49] <girls> e.g. you'd disable systemd-networkd for br0 (I'm not familiar with systemd-networkd - does it come with service files on a per-interface basis?) and put your own script in
[11:16:25] <chub> yes, but I have a lot of other network interfaces, and in my mind it feels like too much effort, to disable systemd-networkd and then moving all those other network interfaces over to a custom script. when I only have a problem with the bridge configuration, because I can see it adding the interfaces together, the IP is configured, it is just a matter of bringing the bridge interfaces up
[11:16:54] <chub> but you are probably right
[11:17:13] <chub> I guess I will have to restart and see if there is a race condition (what I think you are referring to)
[11:17:38] <girls> you can also try to modify the behaviour of systemd-networkd with drop-in files
[11:17:53] <chub> oh? I didn't know that could also be an option
[11:17:56] <girls> but for that, you have to dive into its config an look why it fails in the first place
[11:17:58] <chub> how do you mean with drop in files?
[11:18:58] * girls cannot find the right manpage ...
[11:19:18] <chub> well there is an idea, lets see what Google turns up
[11:19:27] <girls> it's that thing with the $name.service.d/ directory in /etc/systed/system
[11:19:50] <girls> where you can place additions to the original $name.service (from /usr/lib/systemd/system)
[11:20:16] <girls> this way you can add custom dependencies to schedule the service later (after a "ip link set dev br0 up" service, probably)
[11:20:51] <chub> are we talking about systemd-network specifically or a generic systemd service?
[11:20:59] <girls> works for each
[11:21:04] <chub> like a hook script for systemd or systemd-network
[11:21:09] <girls> but you should do it with systemd-network ;-)
[11:21:48] <girls> e.g. /etc/systemd/system/systemd-network@br0.service.d/bring-interface-up.conf
[11:22:01] <girls> where you put "after=bring-br0-up.service"
[11:22:28] <girls> and then you can add a bring-interface-up.service in /etc/systemd/system/ where you do exactly that: "ip link set dev br0 up"
[11:22:53] <girls> (additionally to "After=" you'll probably need a "Depends=", too)
[11:22:53] <chub> hmmmm, is this in the systemd wiki on arch?
[11:23:01] <girls> most probably: yes
[11:23:32] <girls> https://wiki.archlinux.org
[11:23:34] <phrik> Title: systemd - ArchWiki (at wiki.archlinux.org)
[11:23:46] <chub> but now how will it run? systemd-networkd is the one that fails to bring it up, how will it know now to run that script that you are suggesting, is it because systemd-network is now in a new state or something?
[11:23:53] <girls> wtf?? systemd even comes with an "edit" command for those O.o
[11:24:17] <girls> systemd-network will *depend* on your service, that brings the interface up
[11:24:46] <chub> but what is the sequence?
[11:25:04] <girls> sequence of what?
[11:25:04] <chub> will systemd-network run, up to a point, fail, and then call my script?
[11:25:20] <girls> no
[11:26:02] <girls> ah, you ask, because you want it to *assemble* br0?
[11:26:41] <chub> that is what I'm trying to achieve, hoping that systemd-network will start working again after another upgrade in the future, so this "temp fix" is only that, relay on the system as it is and will not break anything
[11:27:00] <girls> hmm
[11:27:05] <chub> girls: yes, I want systemd-networkd to continue doing what it is doing and then only call my script at the end
[11:27:10] <girls> systemd-networkd looks quite monolithic :-/
[11:27:24] <girls> I'm used to netctl - which creates one service per network interface
[11:27:38] <chub> okay I see, I never got netctl working
[11:27:48] <chub> so does it use the same files?
[11:27:53] <girls> no
[11:28:14] <girls> I think, the real issue which should be fixed, is systemd-networkd not bringing up the interface
[11:28:24] <chub> girls: yes of course
[11:28:29] <girls> but since I never used systemd-networkd, I'm not a big help here, sry
[11:28:49] <chub> well you gave me some interesting insight into systemd
[11:29:11] <chub> how do you go about netctl that makes a service per interface
[11:29:25] <girls> what do you mean?
[11:29:44] <girls> you end up with netctl@eno1.service, netctl@br0.service, ...
[11:29:51] <girls> and each can come with its own dependencies
[11:30:23] <chub> ah like in gentoo /etc/init.d/<interface name>.start
[11:31:01] <chub> I'm reading the wiki https://wiki.archlinux.org
[11:31:02] <phrik> Title: netctl - ArchWiki (at wiki.archlinux.org)
[11:31:08] <girls> you create profiles in /etc/netctl and enable/start them via "netctl enable/start $name" - which actually (almost only) sets up proper systemd services
[11:31:19] <girls> reading wiki is a good idea :-)
[11:31:29] <chub> and this is independent of systemd-networkd
[11:31:35] <girls> yes
[11:31:49] <chub> okay, well I started using systemd-network because it was in the beginners section for arch users
[11:31:55] <girls> well "independent" is maybe the wrong word: it is a different tool to set up networking
[11:32:03] <girls> e.g. you have to disable systemd-networkd, then
[11:32:06] <chub> okay, I'll look into netctl, maybe this is the way to go
[11:32:57] <chub> girls: thanks for the help
[11:33:12] <girls> chub: np :-)
[11:35:43] <chub> I have a question, so does this mean, your services have the before or after dependency based on each individual interface
[11:36:02] <chub> and that is why you had that approach?
[11:37:28] <girls> what service are you refering to?
[11:37:33] <girls> the one to bring the link up?
[12:04:08] <chub> girls: no I mean in general, I was wondering why you had a dependencies tied to your services based on the interface
[12:04:35] -!- z3ntu has quit [Write error: Connection reset by peer]
[12:06:51] <girls> the service will depend on the interface being available
[12:07:08] <girls> e.g. we have a vlan on top of a link-aggregation
[12:07:31] <girls> so we have the service for lag0 depending on eno3 and eno4
[12:07:42] <girls> and the service for lag0.1, lag0.2, ... depending on lag0
[12:07:56] <girls> then systemd does all the correct ordering for you :-)
[12:09:28] -!- eschwartz[m] has quit [Ping timeout: 276 seconds]
[12:22:23] <chub> ah I see, okay, yes that would make sense
[12:27:13] <buildmaster> pentium4/pony-stable is broken (says eurobuild6-1): https://archlinux32.org
[12:28:13] <buildmaster> i686/pony-stable is broken (says nlopc46): https://archlinux32.org
[12:29:56] <buildmaster> pentium4/npm is broken (says rechenknecht): https://archlinux32.org
[12:49:44] -!- eschwartz[m] has joined #archlinux32
[13:02:17] -!- T`aZ has joined #archlinux32
[13:02:30] <T`aZ> [ 6498.620247] nvidia: version magic '5.1.6-arch1-1.0-ARCH SMP preempt mod_unload 686 ' should be '5.2.1-arch1-1.0-ARCH SMP preempt mod_unload 686 '
[13:02:59] <T`aZ> looks like the nvidia binary driver needs to be rebuild against current new kernel :/
[13:03:33] <T`aZ> (im using nvidia-340xx)
[13:06:33] -!- AndrevS has joined #archlinux32
[13:17:42] <T`aZ> and ofc the makepkg fails :s
[13:18:08] <girls> nvidia-340xx was removed upstream
[13:18:25] <girls> we should do that, too
[13:19:34] <T`aZ> there is a package in aur anyway
[13:19:51] <T`aZ> the archilinux32 website gives 502 :/
[13:20:05] <girls> not again ...
[13:27:31] <T`aZ> let's try that nouveau driver then :/
[13:28:07] -!- z3ntu has joined #archlinux32
[13:47:14] -!- deep42thought has joined #archlinux32
[13:47:14] <buildmaster> Hi deep42thought!
[13:47:14] <buildmaster> !rq deep42thought
[13:47:16] <phrik> buildmaster: <deep42thought> doesn't one usually try to avoid rust on metals?
[13:47:39] <deep42thought> mysql on the webserver shows constant 200% cpu utilization :-/
[13:47:48] <deep42thought> this is the reason for all those gateway timeouts
[13:58:26] -!- chub has quit [Ping timeout: 248 seconds]
[14:16:46] -!- chub has joined #archlinux32
[14:36:10] -!- panosdev has quit [Quit: WeeChat 2.5]
[14:44:18] -!- Vollzornbrot has joined #archlinux32
[14:44:37] <Vollzornbrot> hello guys
[14:45:32] <Vollzornbrot> deep42thought you are here?
[14:47:52] -!- AndrevS has quit [Remote host closed the connection]
[14:47:58] -!- chub has quit [Ping timeout: 245 seconds]
[14:48:50] -!- Vollzornbrot has quit []
[15:06:07] <deep42thought> now I'm back
[16:03:37] -!- deep42thought has quit [Quit: Leaving.]
[16:36:17] -!- thePiGrepper has joined #archlinux32
[17:22:34] -!- abaumann has joined #archlinux32
[17:22:34] <buildmaster> Hi abaumann!
[17:22:34] <buildmaster> !rq abaumann
[17:22:35] <phrik> buildmaster: <abaumann> This will make guys happy, I actually checked the knowledge base of bugs first.. ;-)
[17:23:58] <abaumann> deep42thought: strange stuff happening, but the buildmaster and slaves seem all building fine, that's the important thing :-)
[17:25:10] * abaumann tries to reach a place not boiling at 35 degrees..
[17:25:13] -!- abaumann has quit [Client Quit]
[17:26:02] -!- WeS has joined #archlinux32
[19:33:32] -!- ornox has joined #archlinux32
[19:33:59] <ornox> DKMS module builds failing with:
[19:34:00] <ornox> cc1: error: incompatible gcc/plugin versions
[19:34:10] <ornox> i think kernel got built with a different version of gcc
[19:41:19] <ornox> this is with gcc linux and linux-headers package installed, nothing fancy done
[19:41:34] <ornox> only happened after updating
[20:04:10] -!- deep42thought has joined #archlinux32
[20:04:11] <buildmaster> Hi deep42thought!
[20:04:11] <buildmaster> !rq deep42thought
[20:04:12] <phrik> buildmaster: * deep42thought listens carefully, but doesn't hear any bells ringing over here
[20:04:22] <deep42thought> ornox: you think rebuilding the kernel would fix this?
[20:04:28] <deep42thought> or the kernel headers? or what?
[20:06:22] <ornox> maybe, im not entirely sure tbh, google said when they had the error it was because the kernel was built with a different version of gcc than what was available in package
[20:06:49] <ornox> so im not actually certain lol, and i dont really want to build the kernel to find out :S
[20:07:09] <ornox> system im running archlinux32 on is really slow and would take me ages
[20:08:04] <ornox> https://pastebin.com
[20:08:06] <phrik> Title: cc1: error: incompatible gcc/plugin versions cc1: error: fail to initialize plu - Pastebin.com (at pastebin.com)
[20:08:45] <ornox> on everything DKMS attempts to build
[20:08:55] <deep42thought> well, the question aimed at whether *we* should try to rebuild the kernel ;-)
[20:10:03] <ornox> oh yeah sure if it's not too much a pain, just make sure its using the same version of GCC available in package i suppose lol
[20:10:14] <ornox> if that doesnt fix it then :shrug:
[20:11:54] <deep42thought> you're running the default kernel?
[20:12:10] <ornox> yeah, does it on both linux and linux-zen actually
[20:12:35] <ornox> do you get the same problem
[20:12:39] <deep42thought> which architecture are you on? i686 or pentium4?
[20:12:49] <deep42thought> I just want to check some stuff
[20:13:13] <ornox> pentium4
[20:15:14] <deep42thought> and you don't have [testing] enabled?
[20:16:17] <ornox> no
[20:16:41] <deep42thought> hmm, linux-* is always newer than gcc-* in the repos
[20:16:50] <deep42thought> so rebuilding *that* won't probably fix it
[20:17:39] <deep42thought> maybe linux-api-headers is the bad guy?
[20:18:22] <deep42thought> I'll reschedule it - let's see ...
[20:19:38] <ornox> were you able to reproduce it?
[20:19:49] <deep42thought> I didn't try - I believe you ;-)
[20:23:07] <ornox> oh dude
[20:23:17] <ornox> Linux version 5.2.1-zen1-1.0-zen-zen (builduser@erich) (gcc version 9.1.0 (GCC)) #1 ZEN SMP PREEMPT Wed Jul 17 15:17:52 CEST 2019
[20:23:27] <ornox> gcc version 8.3.0 (GCC)
[20:23:41] <ornox> kernel was compiled with 9.1.0 apparently lol
[20:23:55] <ornox> wonder if thats wy
[20:24:16] <deep42thought> gcc 9 seems to be stuck in staging
[20:24:18] <deep42thought> :-/
[20:26:30] <ornox> you think that might be it then? the gcc in the package is 8.3.0, where is 9.1.0
[20:26:47] <ornox> i was hoping maybe it was in another package lol
[20:27:19] <deep42thought> it is in another repository
[20:27:26] <deep42thought> have a look in [staging]
[20:27:32] <ornox> cool, ill try it out
[20:27:40] <deep42thought> but you need to take gcc-libs and probably also glibc from there, too
[20:27:52] <ornox> yep, ill let you know how it goes
[20:30:59] <ornox> alright, testing now...
[20:31:39] <ornox> yep, that was it
[20:31:51] <ornox> added staging and only needed gcc and gcc-libs
[20:32:05] <deep42thought> testing, too?
[20:32:22] <ornox> nah i didnt add testing, nothing gcc or linux-* related was in testing
[20:33:03] <deep42thought> ok
[20:33:57] <ornox> ya works perfect now tho, thanks, but should probably have a solution to fix it that doesn't involve adding staging repo >.>
[20:37:01] <deep42thought> moving gcc to stable would be a solution :-D
[20:37:06] <ornox> sounds good to me!
[20:37:12] <ornox> looking forward to it
[20:37:54] <deep42thought> though, I wonder why it is not being moved automatically ...
[20:38:50] <ornox> does it usually o.o
[20:39:37] <deep42thought> usually, it means *something* (claims to) depend(s) on the old version ...
[20:41:22] <ornox> :: installing x265 (3.1.1-1.0) breaks dependency 'libx265.so=169-32' required by ffmpeg
[20:41:27] <ornox> is that related lol
[20:41:34] <deep42thought> no, it's not
[20:41:41] <ornox> ah kk cool
[20:41:55] <deep42thought> for some reason ffmpeg fails to build
[20:55:01] <deep42thought> ffmpeg -> gnutls -> libunistring
[20:55:06] <deep42thought> the last one fails in check()
[20:55:56] <deep42thought> "FAIL: test-thread_create" hmmm
[20:59:01] <ornox> weird
[21:09:33] -!- ornox has quit [Remote host closed the connection]
[21:11:19] -!- ornox has joined #archlinux32
[21:52:58] <buildmaster> i686/pbpst is broken (says nlopc46): https://archlinux32.org
[21:54:58] <buildmaster> pentium4/pbpst is broken (says eurobuild6-5): https://archlinux32.org
[22:06:09] <buildmaster> i686/helm is broken (says nlopc46): https://archlinux32.org
[22:11:19] <buildmaster> pentium4/helm is broken (says eurobuild6-5): https://archlinux32.org
[22:17:57] -!- thePiGrepper has quit [Ping timeout: 245 seconds]
[22:19:02] -!- thePiGrepper has joined #archlinux32
[22:46:02] <deep42thought> maybe we should switch from mysql to a google spreadsheet? https://xkcd.com
[22:46:03] <phrik> Title: xkcd: Spreadsheets (at xkcd.com)
[22:54:14] -!- deep42thought has quit [Quit: Leaving.]
[23:20:21] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.2019-07-24T23:09:49.g6Ahil.stdin"?.