#archlinux32 | Logs for 2018-07-11

[00:51:14] elibrokeit is now known as meltdown_spectre
[00:59:13] meltdown_spectre is now known as elibrokeit
[01:29:19] -!- thePiGrepper has joined #archlinux32
[01:47:54] -!- isacdaavid has quit [Quit: Leaving.]
[02:49:21] -!- NoobAlice has joined #archlinux32
[04:12:50] -!- davor_ has joined #archlinux32
[04:13:35] -!- davor has quit [Ping timeout: 240 seconds]
[04:13:35] davor_ is now known as davor
[06:52:51] -!- eduardoeae has quit [Ping timeout: 240 seconds]
[06:58:49] -!- NoobAlice has quit [Quit: Leaving.]
[07:07:47] -!- thePiGrepper has quit [Quit: leaving]
[12:01:47] -!- eduardoeae has joined #archlinux32
[12:32:04] -!- deep42thought has joined #archlinux32
[12:32:04] <buildmaster> Hi deep42thought!
[12:32:13] <deep42thought> Hi buildmaster, what's up?
[12:32:13] <buildmaster> up? I'm up for 2 days, 5 hours, 41 minutes, load average: 0.68, 0.83, 0.84 ... and I'm insane :-D
[12:57:57] * buildmaster resumes sanity.
[13:15:43] <deep42thought> :: Retrieving packages...
[13:15:43] <deep42thought> device-mapper-2.02.... 269.6 KiB 27.8K/s 00:10 [######################] 100%
[13:15:43] <deep42thought> linux-4.17.5-1-x86_64 36.7 MiB 33.5K/s 17:07 [###########-----------] 52%
[13:15:56] <deep42thought> the buildmaster is _really_ slow lately
[14:07:00] -!- bill-auger has quit [Quit: No Ping reply in 90 seconds.]
[14:07:36] -!- bill-auger has joined #archlinux32
[14:14:38] -!- belanthor has joined #archlinux32
[14:45:49] -!- davor has quit [Ping timeout: 260 seconds]
[14:50:31] <deep42thought> How is "archlinux32" officially spelled?
[14:50:37] <deep42thought> All lowercase?
[14:50:45] -!- davor has joined #archlinux32
[15:21:22] <tyzoid> deep42thought: I've been using archlinux32 in all code references
[15:21:32] <tyzoid> but IIRC it was supposed to be "Arch Linux 32"
[15:22:02] <deep42thought> [Aa]rch\( \?\)[Ll]inux\132
[15:22:29] <tyzoid> deep42thought: https://www.archlinux.org you can see brtln referred to it as "Arch Linux 32"
[15:22:29] <phrik> Title:Arch Linux - News: The end of i686 support (at www.archlinux.org)
[15:22:51] <deep42thought> I prefer the capitalized version
[15:23:08] <deep42thought> although I'm using the small version myself :-D
[15:40:06] -!- belanthor has quit [Quit: Leaving]
[15:41:14] -!- eduardoeae has quit [Ping timeout: 260 seconds]
[15:43:42] -!- eduardoeae has joined #archlinux32
[16:05:29] <bill-auger> capital letters are pretentious
[16:05:51] <bill-auger> <-- notice my nick is not BillAuger
[16:06:40] <bill-auger> capital letters are for ClassNames and CONSTANTS
[16:07:57] <bill-auger> i would endorese 'ARCHLINUX32'
[16:08:49] <bill-auger> but if you put a space between "arch" and "linux" RMS will send a pack of wild GNUs to sneer at you
[16:09:19] <bill-auger> i wrote an article about that https://wiki.parabola.nu
[16:09:20] <phrik> Title:The ArchLinux Misnomer - ParabolaWiki (at wiki.parabola.nu)
[16:19:10] <deep42thought> for anyone wanting to help to repair python2: https://bugs.python.org https://bugs.python.org
[16:19:10] <phrik> Title:Issue 34095: [2.7] test_idle fails with: /usr/bin/xvfb-run: line 181: 3617 Segmentation fault - Python tracker (at bugs.python.org)
[16:19:19] <deep42thought> my motivation is depleted
[16:19:26] -!- deep42thought has quit [Quit: Leaving.]
[16:43:13] <elibrokeit> Well, Arch Linux is formally Arch Linux, except when it is archlinux
[16:49:22] <tyzoid> bill-auger, elibrokeit: Exactly this ^. deep42thought was asking it in reference to a formal document.
[17:10:39] -!- abaumann has joined #archlinux32
[17:10:40] <buildmaster> Hi abaumann!
[17:11:46] <tyzoid> wb
[17:11:51] <abaumann> hi.
[17:12:20] <abaumann> any clue, what the user-agent 'Transmission/2.84' could be?
[17:12:32] <abaumann> my bet is a download software of sorts.
[17:12:41] <abaumann> basically, it's clogging my mirror. :-)
[17:19:31] <tyzoid> abaumann: Yeah, that's the web seeding
[17:19:37] <tyzoid> you'll probably see it's only to isos
[17:19:59] <abaumann> yeah, I only see partial request (206) on the isos.
[17:20:20] <abaumann> I was considering to move my mirror to a proper hoster, my home network is too expensive to expand.
[17:20:48] <tyzoid> basically, the .torrent files for the isos have several http endpoints, which allow it to get chunks of the data without needing other peers
[17:21:07] <abaumann> ah. that's neat. didn't know.
[17:21:11] <tyzoid> ideally, peers will start peering to each other if there's enough of them
[17:21:30] <tyzoid> but it could just be one or two clients that are grabbing chunks
[17:21:39] <abaumann> This also explains, why I see accesses to the 2018-07-01 ISO only.
[17:22:01] <tyzoid> yeah, and the 206s are because it requests several chunks in the middle of the file
[17:22:19] <tyzoid> ideally, it'll randomize which web seeds get requested with chunks
[17:26:28] <tyzoid> abaumann: It should be possible to get a cheap VPS and configure varnish on it to sit in front of your home machine?
[17:27:09] <tyzoid> lol, that started out as a statement and ended as a question
[17:27:29] <abaumann> * abaumann is slightly puzzled :-)
[17:27:44] <tyzoid> abaumann: varnish is a caching http reverse proxy
[17:27:53] <abaumann> yep. I know that one.
[17:28:04] <abaumann> but, can it handle 206 partial requests?
[17:28:08] <tyzoid> It should
[17:28:25] <abaumann> that would be cool.
[17:28:53] <abaumann> i'll try that.
[17:29:02] <tyzoid> yeah, looks like it does
[17:29:45] <tyzoid> Looks like you can also configure varnish to use both disk and memory caching
[17:30:20] <tyzoid> Only thing to be wary of is to configure the repo dbs to not be cached
[17:30:28] <tyzoid> or to cache them only for a short time
[17:31:07] <abaumann> ..and I'll have to add some QoS on the router for my uplink.
[17:31:34] <tyzoid> yeah, a bw limit from your mirror side should fix things if you have a nice vps cache in front externally
[17:31:41] <abaumann> btw: https://archlinux32.andreasbaumann.cc is up with Let's Encrypt certs.
[17:31:43] <phrik> Title:Index of / (at archlinux32.andreasbaumann.cc)
[17:31:50] <tyzoid> Nice.
[17:31:53] <abaumann> true.
[17:32:09] <tyzoid> the other option is to just move the mirror entirely to a vps, but that requires ~50gb space
[17:32:24] <abaumann> oh. the web requests get through my http proxy, which currently is a Raspberry Pi B+ :-)
[17:33:36] <abaumann> Ah.. all i486 pacman tests run just fine.. the cross-compiled version had some issues..
[17:34:00] <abaumann> I'm currently patching 'packages' in a -i486' branch to get the 'i486' slave closer to producction..
[17:34:03] <abaumann> *production
[17:34:54] <abaumann> mmh. varnish first. other proxy server, if that doesn't help, VPS, yes, thanks tyzoid, good plan. :-)
[17:36:14] <abaumann> ok. have to rush. cu.
[17:36:17] -!- abaumann has quit [Quit: Page closed]
[17:52:40] -!- deep42thought has joined #archlinux32
[17:52:41] <buildmaster> Hi deep42thought!
[17:52:50] -!- abaumann has joined #archlinux32
[17:52:51] <buildmaster> Hi abaumann!
[17:52:58] <deep42thought> Hi!
[17:53:09] <abaumann> hi
[17:53:42] <abaumann> this is me on the cell phone. so expect slow typing.
[17:53:54] <abaumann> my https mirror is up.
[17:54:19] <deep42thought> haven't you said so some days ago, already?
[17:54:29] <abaumann> yes
[17:54:29] * deep42thought questions his memory
[17:54:38] <abaumann> might be. sorry :-)
[17:54:40] <deep42thought> np
[17:54:55] <deep42thought> I was just wondering if I missed something
[17:55:02] <abaumann> nono
[17:55:36] <abaumann> python2: i'm builing it and have a look too..
[17:55:50] <deep42thought> btw: the proposal for spi is out and to-be-decided in the august meeting
[17:55:55] <deep42thought> abaumann: thanks
[17:57:22] <abaumann> i'm curious about who they decide. :-)
[17:58:13] <deep42thought> yeah, me too
[17:58:32] <deep42thought> they have bigger fish on their support list
[17:58:55] <deep42thought> but otoh, there were some projects I hadn't heard of before, too
[18:12:23] -!- abaumann has quit [Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )]
[18:23:32] -!- isacdaavid has joined #archlinux32
[19:33:31] <tyzoid> abaumann: If you're interested, I've got an affiliate link with DigitalOcean that gives $10 of free credit, if you wanted a vps free for a few months
[20:11:14] -!- abaumann has joined #archlinux32
[20:11:15] <buildmaster> Hi abaumann!
[20:11:20] <tyzoid> wb
[20:11:25] <abaumann> hi.
[20:11:57] <abaumann> mmh. python2 build now on my slave.
[20:19:19] <abaumann> mmh. some problems with git.archlinux32.org?
[20:20:23] <tyzoid> I'm able to access it. Is it slow?
[20:20:34] <deep42thought> abaumann: are these ^ two statements related?
[20:20:36] <tyzoid> hmm...
[20:20:47] <abaumann> yes. seems to be very slow.
[20:21:09] <tyzoid> noticed it was a bit sluggish, but now it appears close to frozen
[20:21:10] <deep42thought> abaumann: the build of python2 is fine, it's the testsuite, that fails
[20:21:19] <tyzoid> Ah, right. I'm running a system-wide backup
[20:21:23] <abaumann> deep42thought: ah.
[20:21:25] <abaumann> I see.
[20:21:26] <tyzoid> disk based operations will be slightly slow
[20:21:41] <abaumann> tyzoid: ah, that's explains it.
[20:22:20] <tyzoid> "slightly" seems to be the understatement of the year, though.
[20:22:44] <abaumann> everything besides the backup basically freezes.. :-)
[20:24:31] <tyzoid> abaumann: cgit appears to be working better than gitea atm: https://git.archlinux32.org
[20:24:34] <phrik> Title:Git repository browser (at git.archlinux32.org)
[20:25:27] <deep42thought> abaumann: only the test suite being broken is pretty convenient - it makes rebuilds obsolete for reproducing with different options :-)
[20:25:47] <tyzoid> Yeah... not really working out as hoped. I'll reboot the box once I get into it
[20:25:59] <deep42thought> "Monitor is DOWN: Archlinux32 Git" hehe
[20:26:00] <tyzoid> even on tty, logins are timing out
[20:26:33] <tyzoid> Arch Linux 4.15.15-1-pve (lxc/tty1)
[20:26:34] <tyzoid> arch32-git login: tyler
[20:26:35] <tyzoid> Password:
[20:26:35] <tyzoid> login: timed out after 60 seconds
[20:27:29] <abaumann> but. I still have a working shell on the i486 VM. :-)
[20:27:40] <tyzoid> No, I just mean the git vm
[20:30:15] <deep42thought> hmm, strange: on i686, makepkg compiles inside src/ and packages to pkg/, but on x86_64 it compiles inside $pkgbase/src and packages to $pkgbase/pkg
[20:34:19] <tyzoid> deep42thought / abaumann: Do either of you have cronjobs that fetch every minute?
[20:34:41] <deep42thought> I have cronjobs
[20:34:46] <deep42thought> not sure if every minute, though
[20:34:50] <deep42thought> but it might add up to
[20:36:07] <tyzoid> gitea's ssh service is eating up all the ram / cpu
[20:36:18] <tyzoid> indicating it was getting hit relatively hard
[20:38:05] <deep42thought> there are several forks (on the buildmaster, the slaves, my box, ...) which pull frequently from the git repositories
[20:39:14] <deep42thought> just checked: the buildmaster does not access it via https
[20:40:06] <tyzoid> [tyler@arch32-git ~]$ sudo reboot
[20:40:07] <tyzoid> Failed to talk to init daemon.
[20:40:07] <tyzoid> [tyler@arch32-git ~]$
[20:40:11] <tyzoid> Well, isn't that just great
[20:40:36] <abaumann> yeah. the joys of systemd. :-)
[20:40:37] <deep42thought> standard configuration of the slaves is to access via https, though
[20:49:04] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.stdin.2018-07-11T20:48:59.kEycAJ"?.
[20:49:04] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.stdin.2018-07-11T20:49:00.m5oOIJ"?.
[20:51:21] <tyzoid> huh, so it rebooted itself?
[20:51:36] <deep42thought> who?
[20:51:46] <tyzoid> git.archlinux32.org
[20:51:54] <tyzoid> must've finally processed the reboot command
[20:51:59] <deep42thought> :-D
[20:52:37] <deep42thought> it's like shutting down your laptop and closing the lid before it completes the shutdown, so it enters hybernation - and if you wake it next time, it completes the shutdown :-)
[20:53:11] <tyzoid> hate when that happens
[21:05:17] * buildmaster resumes sanity.
[21:19:07] -!- abaumann has quit [Quit: leaving]
[21:22:14] * buildmaster failed to execute a mysql query - can you have a look at "tmp.mysql-functions.query.stdin.2018-07-11T21:16:20.1Gpk8q"?.
[21:40:22] -!- titus_livius has joined #archlinux32
[21:41:13] -!- buildmaster has joined #archlinux32
[21:52:20] <elibrokeit> deep42thought: not strange at all, it's a makepkg bug
[21:52:34] <deep42thought> oh, what triggers it?
[21:52:40] <elibrokeit> if $PWD is a symlink, makepkg gets confused and thinks you're using a custom $BUILDDIR
[21:52:49] <deep42thought> ah, ok
[21:52:58] <elibrokeit> fixed in git master
[21:53:18] <elibrokeit> https://git.archlinux.org
[21:53:19] <phrik> Title:pacman.git - The official pacman repository (at git.archlinux.org)
[21:53:25] <deep42thought> thanks for the explanation
[21:55:10] <deep42thought> I'm indeed working on x86_64 on a symlink and on i686 not
[21:55:23] <elibrokeit> since using $BUILDDIR makes makepkg build from $BUILDDIR/$pkgbase/src instead of $PWD/src (to prevent multiple packages clobbering each other)
[21:56:57] <deep42thought> ah, yeah, makes sense
[21:57:10] <deep42thought> I never used this feature on archlinux, though, but it's still a nice one :-)
[23:11:07] -!- deep42thought has quit [Quit: Leaving.]