2020-07-12T02:14:15 *** okurz_ is now known as okurz 2020-07-12T12:51:23 kl_eisbaer: our monitoring must be drunk ;-) 2020-07-12T12:51:31 Updates CRITICAL : discontinued OS Release openSUSE 15.2 2020-07-12T12:53:03 cboltz: https://cdn.discordapp.com/attachments/366989996101730304/731855201774076004/unknown.png and https://cdn.discordapp.com/attachments/366989996101730304/731855314672156782/unknown.png show that the autoindex skin is more or less ready ;) 2020-07-12T12:54:09 I reported https://progress.opensuse.org/issues/68863 to fix the issues that arise when using it, either from my side or server side (although I can't fix everything from my side) 2020-07-12T12:57:06 lcp79: did you run highstate on jekyll.i.o.o for the monitor.o.o main page? 2020-07-12T12:57:25 I just tried to switch to it (the haproxy way), and got the news.o.o content 2020-07-12T12:58:49 I didn't 2020-07-12T12:59:02 news is the default server, so that's why 2020-07-12T12:59:35 guessed so ;-) 2020-07-12T13:00:06 please tell me when the monitor.o.o main page is ready so that I can remove a # in the haproxy config 2020-07-12T13:00:55 for download.o.o - you mentioned in the ticket that you need to adjust jekyll-theme to not use https, but I'm not sure if that is really needed 2020-07-12T13:01:17 it is, firefox/chrome will refuse to load https content on http page 2020-07-12T13:01:28 and vice versa for that matter 2020-07-12T13:02:20 really? that makes things more interesting 2020-07-12T13:02:49 kl_eisbaer: after 15.2 update on some servers highstate doesn't work because of faulty salt in main repos, it should be switched to openSUSE:infrastructure version 2020-07-12T13:02:50 I hope redirects to https aren't a problem? 2020-07-12T13:03:03 they sure are 2020-07-12T13:03:19 that's why I mentioned to have static accessible from http too 2020-07-12T13:03:44 cboltz: "Updates CRITICAL : discontinued OS Release openSUSE 15.2" - it's not drunk, just a "bug" in the official package (too old version) 2020-07-12T13:04:27 "drunk" sounds better ;-) 2020-07-12T13:04:42 about the highstate on server: looks indeed like a salt problem. 2020-07-12T13:05:13 cboltz: I found 2 other bugs with monitoring packages in 15.2 already. Currently just collecting them to make one big update ;-) 2020-07-12T13:05:38 nice ;-) 2020-07-12T13:06:31 lcp79: thanks, btw., for your work on the download-o-o stuff. IMHO the https://build.opensuse.org/package/show/openSUSE:infrastructure/mirrorbrain-icons package should contain all available icons. If not, it's a bug ;-) 2020-07-12T13:07:15 FYI: ...and for mirrors.opensuse.org, I plan to extend https://build.opensuse.org/package/show/openSUSE:infrastructure/mirrorbrain-webpages with the template file as well 2020-07-12T13:07:24 so this could also end up in a git repo, if wanted. 2020-07-12T13:09:58 on which servers do you see the highstate failure? (one or two examples are good enough) 2020-07-12T13:10:34 jekyll a moment ago, I fixed it with `zypper in -f --from openSUSE:infrastructure salt*` 2020-07-12T13:10:49 The following "single page" pages might also worth an overhaul: https://kubic.opensuse.org/ https://monitor.opensuse.org/ https://osc-collab.opensuse.org/ https://rpmlint.opensuse.org/ https://yast.opensuse.org/ https://gcc.opensuse.org/ ;-) 2020-07-12T13:10:59 ah, and I forgot https://debuginfod.opensuse.org 2020-07-12T13:11:02 kubic is kinda on my radar 2020-07-12T13:11:12 yast is good imo 2020-07-12T13:11:24 and monitor is done >:D 2020-07-12T13:11:49 it is on jekyll.infra.opensuse.org already 2020-07-12T13:12:14 I made the current design of kubic, so that's why it's on my radar 2020-07-12T13:12:16 oh!? thanks! 2020-07-12T13:12:49 where is gcc hosted btw? 2020-07-12T13:12:58 on gcc.infra.opensuse.org 2020-07-12T13:13:36 I meant like git repo or something 2020-07-12T13:14:58 uh: that's something for the main operators 2020-07-12T13:15:41 aw, I had no idea setloc.js makes the links in breadcrumb happen, I would have rewrote it to better work with the theme 2020-07-12T13:15:50 => mmarek@o.o 2020-07-12T13:15:52 I kinda wish it was done on the server side too 2020-07-12T13:16:07 alright, I will contact them later 2020-07-12T13:17:38 kl_eisbaer: somewhat related - we have a pillar/id/gcc2_i_o_o ("future replacement for gcc.o.o") - but gcc2.i.o.o goes NXDOMAIN. 2020-07-12T13:18:00 is this a superfluous pillar/id/ file (= no replacement of gcc.i.o.o needed), or something else? 2020-07-12T13:18:13 cboltz: that one can go, as we managed to upgrade the current gcc to the "latest" Tumbleweed. 2020-07-12T13:18:41 ok, I'll delete it 2020-07-12T13:18:51 (the last update is a month old, so the next "zypper dup" is needed for the machine - that's why I say "latest" ;-) 2020-07-12T13:18:57 thanks. 2020-07-12T13:19:32 cboltz: if you are on it: I guess after the latest discussions with SUSE-IT, it might make sense to remove the daffy machines from salt as well - as SUSE-IT wants to manage them "on their own". 2020-07-12T13:19:55 while doing the tumbleweed update, please also check why ssh login as user doesn't work on gcc.i.o.o 2020-07-12T13:22:14 hm, ok. 2020-07-12T13:23:08 for daffy* - they are (at least partially, for example keepalived) salt-managed, so IMHO it would make sense to keep them in salt 2020-07-12T13:23:54 kl_eisbaer: I assume you don't know how the osc-collab front page is done? the osc-plugin-collab repo has an even older version 2020-07-12T13:24:22 cboltz: if I understand bmwiedemann and jdsn correctly, they did not want to have any openSUSE hero on those machines. That in turn would make it impossible to have them controlled by a salt-master under heroes control. But I might be wrong - so better ask back. 2020-07-12T13:24:47 cboltz: monitor on jekyll.i.o.o is ready 2020-07-12T13:25:00 lcp79: no, not really. I just look at those machines when i update/upgrade them or if something is broken that I may be able to fix if the original maintainer is not available 2020-07-12T13:25:16 alright, will ask dimstar then 2020-07-12T13:25:26 who apparently is here >:D 2020-07-12T13:25:30 but I like to stay away from all dirty details of those machines, if possible ;-) 2020-07-12T13:25:58 I also have cooverview to update too, it's a lot of stuff 2020-07-12T13:26:18 lcp79: thanks, haproxy config updated. The monitor.o.o main page now gets served from jekyll.i.o.o 2020-07-12T13:26:48 god bless I haven't finished the obs theming code, I would have way more work to do 2020-07-12T13:26:51 lcp79: are you also trying to provide Grafana dashboards ? 2020-07-12T13:27:01 In that case, I have some requests :-) 2020-07-12T13:27:17 not currently, but you can line up to have that done as well >:D 2020-07-12T13:27:54 hehe: I have no problems to line up - especially, if I get something at the end :-) 2020-07-12T13:28:49 we should have a request tracker for websites 2020-07-12T13:29:21 lcp79: which one would you prefer? 2020-07-12T13:30:01 I don't know tbh, can I request a pagure instance? >:D 2020-07-12T13:30:39 * kl_eisbaer is thinking about a "o-o startpages" Github repo and subdirectories for all of the sub-domains... 2020-07-12T13:31:02 that sounds like a nightmare tbh 2020-07-12T13:31:29 lcp79: feel free to request one - just line up to have that done as well :-D 2020-07-12T13:31:47 hm: really? Why? 2020-07-12T13:31:56 I will deploy pagure one day 2020-07-12T13:32:06 That would at least help to keep an overview about all the different sub-domain pages 2020-07-12T13:32:21 it would end up as unmaintained as everything else >:D 2020-07-12T13:32:54 jip, but right now, the pages are also unmaintained - and on top of that, nobody knows where their original source comes from... 2020-07-12T13:33:35 kl_eisbaer: pillar/role/jekyll_master.sls has a list ;-) 2020-07-12T13:33:41 Which makes it even harder to contribute. With a repo on github, people could at least do spell checking or other stuff ;-) 2020-07-12T13:34:21 cboltz: can I btw. remind you to think about a documentation for a "simple heroes Salt setup" for testing at home ? 2020-07-12T13:34:44 kl_eisbaer: mmarek's openSUSE email leads to novell.com address 2020-07-12T13:34:51 lol... 2020-07-12T13:35:43 kl_eisbaer: that's already hidden in our internal wiki ;-) (with the exception of my additional salt files, maybe I should just upload the tarball?) 2020-07-12T13:35:44 ...and I can't log in on gcc.infra.o.o to find out who logged in lately :-/ 2020-07-12T13:35:57 cboltz: yes, please do that. 2020-07-12T13:37:00 kl_eisbaer: if I wanted to touch setloc.js from mirrorbrain-icons, should I also do it in the download-o-o repo? 2020-07-12T13:37:30 but I guess that matz@s.d might be able to help out 2020-07-12T13:38:07 lcp79: both are so far completely independent from each other. So feel free to do whatever is needed. 2020-07-12T13:38:32 I have no clue how that's even loaded tbh 2020-07-12T13:39:25 oh yeah, cboltz, would you mind giving me permissions to download-o-o, I lost them when moving the repo to openSUSE namespace 2020-07-12T13:40:04 lcp79: the pages on download.o.o are generated by apache2-mod_mirrorbrain 2020-07-12T13:40:27 well, mod_autoindex_mb, no? 2020-07-12T13:40:40 lcp79: you are correct, yes 2020-07-12T13:41:01 There should be a "README.html" and a "HEADER.html" in the package 2020-07-12T13:41:29 lcp79: I upgraded you from "write" to "admin" 2020-07-12T13:41:34 ...and in addition some js, which is loaded by the "IndexHeadInsert" command 2020-07-12T13:41:57 alright, I will replace that with head.html file to load instead then 2020-07-12T13:42:46 lcp79: just have a look at the /etc/apache2/mod_autoindex-mirrorbrain.conf file on pontifex 2020-07-12T13:43:18 and the files that end up in /usr/share/apache2/mirrorbrain-icons/ 2020-07-12T13:43:18 alright, I will, if I can ssh there >:D 2020-07-12T13:43:53 lcp79: IMHO you should - as normal user + sudo 2020-07-12T13:52:19 kl_eisbaer: https://progress.opensuse.org/projects/opensuse-admin-wiki/wiki/Testing_with_local_saltmaster#section-5 2020-07-12T13:54:56 thanks! Will start testing, once I have my server back with nouveau :-/ 2020-07-12T13:57:46 different question: csync2 on anna says "Updating /etc/pdns/recursor.conf on elsa ..." -> "File stays in dirty state. Try again later...". Any idea what could cause this? 2020-07-12T13:58:09 and, bonus question, how to handle it? 2020-07-12T14:28:24 cboltz: this is a result of the latest upgrade: after that, the pdns-recursor.conf permissions were wrong. Before I figured that out, I decided to install and configure dnsmasq on anna instead, to get us up and running again. 2020-07-12T14:28:57 darix promised to look at his "baby" (aka: pdns-recursor) on elsa (to not interrupt the working dnsmasq setup on anna) 2020-07-12T14:29:08 but I did not hear back from him. 2020-07-12T14:30:02 tell him that I'll replace it with unbound if he doesn't fix it ;-) 2020-07-12T14:30:07 As dnsmasq would allow us to even have a simple DHCP server in the intra network, I am currently thinking of making the temporary solution (dnsmasq on anna) a permanent one - and obsoleting pdns on elsa as well. 2020-07-12T14:30:29 cboltz: feel free to tell him yourself, once he is online again ;-) 2020-07-12T14:30:38 ;-) 2020-07-12T14:30:59 But about your question: I expect the pdns-recursor.conf on elsa to the correct one, that should be copied over to anna. 2020-07-12T14:31:51 But: on pdns is currently disabled on anna anyway in favor of dnsmasq, so - whatever you do - it should not influence any running service at the moment. 2020-07-12T14:32:10 ok, then I'll just ignore the csync2 warning 2020-07-12T14:32:30 just try again now 2020-07-12T14:33:16 no complaints anymore :-) - thanks! 2020-07-12T14:34:05 ok - and elsa is rebooting now for kernel update 2020-07-12T14:40:09 * cboltz has to leave, see you later! 2020-07-12T15:07:15 lcp79, hi, any idea why forum response is slow get png files? 2020-07-12T15:54:33 malcolmlewis: what do you mean? 2020-07-12T16:00:30 lcp15, https://susepaste.org/7088823 2020-07-12T16:00:56 for me the forum is running oh so slow..... 2020-07-12T16:02:25 15 plus seconds to get to any page 2020-07-12T16:17:23 malcolmlewis: ditto, very slow 2020-07-12T16:19:21 pjessen, what do you think it might be, or does it need a restart? 2020-07-12T17:07:05 pjessen, lcp15 ^^ 2020-07-12T17:07:14 malcolmlewis: dunno - I would suspect database access. 2020-07-12T17:07:57 pjessen, any logs to look at? 2020-07-12T17:09:46 malcolmlewis: plenty :-) the apache log will basically confirm the issue 2020-07-12T17:10:14 pjessen, no access here, do you? 2020-07-12T17:11:52 yup. 2020-07-12T17:12:06 seems to be running fine 2020-07-12T17:12:14 go figure 2020-07-12T17:12:47 I only get long waits when I access something I haven't accessed before ? 2020-07-12T17:13:46 pjessen, yes, but it's every thread I look at 10-20 seconds for it to load 2020-07-12T17:15:24 malcolmlewis: Yeah, I see that too. 2020-07-12T17:22:37 pjessen, something has changed in the last 24-36hrs... 2020-07-12T17:23:12 pjessen, stuck maintenance cron jobs? 2020-07-12T17:24:09 pjessen, or does that re-indexing need to occur? I think I posted the screenshots for you? 2020-07-12T17:43:09 malcolmlewis: tbh, dunno. 2020-07-12T17:47:01 pjessen, can you see this page on our forum? https://www.vbulletin.com/docs/html/main/acp_settings_options_serverandoptimization 2020-07-12T21:26:16 cboltz: https://cdn.discordapp.com/attachments/366989996101730304/731982614390505533/unknown.png >:D 2020-07-12T21:26:41 I can promise you right now that if we ever get Bugzilla 6 deployed, I will make this happen :P 2020-07-12T21:30:50 looks good :-) - so let's wait for someone to update bugzilla 2020-07-12T21:31:07 well, Bugzilla 6 is also unreleased yet >:D 2020-07-12T21:31:23 it has been in development for the last 2ish years 2020-07-12T21:32:00 I'm kinda involved tho 2020-07-12T21:32:25 because of https://bugzilla.lcp.world/ and https://planet-bugzilla.lcp.world/ 2020-07-12T21:36:44 nice :-) - I start to wonder if there's something you are _not_ involved ;-) 2020-07-12T21:37:30 It's very easy to get me on board