Project

General

Profile

communication #51260 » 2019-06-04-heroes-meeting.txt

IRC meeting log - cboltz, 2019-06-04 21:26

 
2019-06-04 heroes meeting

[19:14:25] <thomic> hi all is cboltz already around and pjessen maybe?
[19:14:28] <thomic> I can'
[19:14:34] <thomic> I can't be there today at 20.00
[19:14:42] <thomic> I have another meeting starting at 20.10
[19:14:53] <thomic> If you have questions feel free to ask them now cboltz :)
[19:14:59] <thomic> otherwise we can talk tommorow as well
[19:22:30] * Eighth_Doctor waves
[19:28:24] <okurz> same for me, otherwise I would have joined. As an update to everyone in the channel (I might repeat this later): Regarding redirect problems I try to help by playing the bridge between openSUSE admins and Microfocus IT in https://progress.opensuse.org/issues/42599 . So far there is an MF IT ticket still open but no real technical progress so far
[19:34:28] <cboltz> thomic: maybe you have a quick status update for rsync.o.o?
[19:34:45] <thomic> oehm well it is syncing and working as http proxy like charme
[19:35:18] <thomic> but as soon as i enable rsync as well... it breaks down the disk i/o and it gets out of sync fast and everybody complains
[19:35:46] <cboltz> great :-/ - but still good to know
[19:35:46] <thomic> of course we all can look into this together but 8 hard disk with 4tb each dont make a good performance
[19:35:50] <thomic> aehm... 6
[19:35:53] <thomic> not 8
[19:36:10] <thomic> so we can have no rsync
[19:36:12] <thomic> or no http
[19:36:31] <cboltz> actually pjessen already came up with the idea of "no http" because we have enough http mirrors
[19:36:32] <thomic> but as rsync.o.o is our first line of defence http mirror
[19:36:58] <thomic> well we in fact not have that many push-mirrors right?
[19:37:06] <thomic> but lets discuss this with pjessen
[19:37:15] <thomic> im open for everything that helps
[19:37:31] <cboltz> exactly, pjessen should know more details about mirrors
[19:37:38] <thomic> best would be to move this to some $cloud
[19:37:40] <thomic> xD
[19:37:55] <thomic> instead of having hard disks somewhere spinning slow
[19:38:03] <lcp> everything in btfs and ipfs
[19:39:35] <thomic> lcp: ?
[19:40:11] <lcp> well, they are decentralized nodes for data sharing >:D
[19:40:37] <lcp> it's perfect for mirror infrastructure which you don't want to host >:D
[19:41:12] <lcp> https://www.bittorrent.com/btfs/ https://ipfs.io/
[19:43:58] <thomic> aehm...
[19:44:05] <thomic> yay
[19:44:14] <thomic> of course we dont want to host
[19:44:24] <thomic> thats what you got out of the conversation
[19:44:30] <thomic> try again ;(
[19:44:32] <thomic> ;)
[19:45:52] <lcp> thomic: I'm joking :P
[19:46:19] <lcp> ipfs is reeeeallly slow irl in my experience
[19:47:57] <cboltz> lcp: you know that sometimes the best ideas are based on jokes? First you laugh, then someone says "why not?" ;-)
[19:49:53] <thomic> what might work (might) is something like a lot of small VMs
[19:49:57] <thomic> with 500G storage each
[19:50:02] <thomic> having a distributed ceph-cluster
[19:50:04] <thomic> or so
[19:50:09] <thomic> but if this syncs in time
[19:50:15] <thomic> via 1G ethernet to the internet
[19:50:19] <thomic> and the traffic limitations
[19:50:22] <thomic> i guess not
[19:50:43] <thomic> so actually we need a provider who would be happy to share 20-30TB out of there CephCluster
[19:50:49] <thomic> *their
[19:52:27] <adamm> there was some CDN that wanted to share some space, but nothing ever happened with that AFAIK
[19:53:37] <cboltz> right, I think it was CDN77
[19:53:43] <cboltz> IIRC some testing (whatever that means) was done, but I never heard about results
[19:54:49] <pjessen> I don't know of anyone testing anyhting with cdn77 - am I late to the meeting?
[19:56:31] <cboltz> pjessen: we started earlier because thomic has another meeting at 20:10
[19:57:21] <bmwiedemann2> Hi
[19:57:56] <cboltz> the cdn77 testing was done > 2 years ago, and the person who did it is no longer around
[19:58:55] <tuanpembual> hi
[19:59:31] <pjessen> aha. okay.
[20:00:55] <cboltz> since it's 20:00 now, let's formally start the meeting ;-)
[20:01:42] <cboltz> since thomic has to leave soon, I'd say we move (waiting for) questions from the community to the end of the meeting
[20:01:49] <cboltz> and instead start with status reports
[20:02:13] <cboltz> so - does someone have a status report?
[20:02:18] <pjessen> let me start
[20:02:45] <pjessen> from a mirror pov, I think the 151 release went well, but there seems to be some issues wqith torrents.
[20:03:17] <pjessen> I don't know much about bittorrent, if any0onme wants to have a look, that would be great.
[20:04:11] <pjessen> Otherwise not much to report - I have a new mailing list that needs a virtual maps update, but i'll poke thomic off-line
[20:04:16] <bmwiedemann2> pjessen: the tracker does not know about the torrents, so only clients with DHT support can share the torrent
[20:04:29] <cboltz> I only know that lnussel asked some days before the release if bittorrent still gets used - as the tickets and the mailinglist noise show, the answer is yes
[20:05:59] <pjessen> yes, it looks like it. Feel free to update the tickets or reply on the lists - I won't, coz I just don't know.
[20:06:42] <cboltz> I also don't know anything about the technical side ;-)
[20:06:51] <bmwiedemann2> darix handled it in the old days and he gave me the hint that one has to login there and drop the torrent files in some place and restart the service
[20:07:51] <lcp> I should setup a peer myself on my server
[20:08:15] <cboltz> bmwiedemann2: do you know on which server it is running, and where we have to drop the torrent files?
[20:08:17] <pjessen> i used to run a feeder at Hetzner, but traffic was so low
[20:08:29] <pjessen> tracker.o.o ?
[20:08:33] <lcp> I don't really use torrent much, but not that I will pay extra for it
[20:09:15] <cboltz> pjessen: obviously yes, but instead of finding out which server does the actual work, asking is easier ;-)
[20:09:27] <bmwiedemann2> tracker.o.o is a cname for stage.o.o
[20:09:40] <bmwiedemann2> our admin wiki does not have details on it though
[20:10:21] <pjessen> there is sometrhing called "hefurd" running on pontifex. seems torrent related
[20:11:07] <lcp> https://github.com/abique/hefur
[20:11:26] <pjessen> I see some torrent files for the DVDs, but not for 15.1
[20:11:31] <bmwiedemann2> is there a /var/lib/hefurd ?
[20:11:40] <pjessen> yup, has the torrent files
[20:12:15] <bmwiedemann2> want to add http://download.opensuse.org/distribution/leap/15.1/iso/openSUSE-Leap-15.1-DVD-x86_64.iso.sha256.torrent
[20:12:16] <pjessen> has 3 torrent files - 42.2, 42.3 and 15.0
[20:12:21] <bmwiedemann2> err. without sha256
[20:12:33] <pjessen> okay. do I just drop it in there and restart?
[20:12:51] <bmwiedemann2> think so
[20:13:33] <pjessen> done
[20:14:19] <bmwiedemann2> I'll watch my rtorrent to see if it succeeds to track there now
[20:14:46] <bmwiedemann2> yay!
[20:14:57] <cboltz> :-)
[20:15:01] <pjessen> cool! gotta take a 5min break, pick up my wife from the sbahn. I'll be back.
[20:15:20] <tuanpembual> cool
[20:15:33] <tuanpembual> allow me be next.
[20:15:52] <tuanpembual> I still on progress.infra.o.o
[20:16:10] <tuanpembual> sorry, I feel my progress is too slow.
[20:16:27] <tuanpembual> still working on plugin redmine_ichain.
[20:16:39] <tuanpembual> http://progress.infra.opensuse.org
[20:16:51] <tuanpembual> cboltz done with nginx.
[20:17:22] <cboltz> nice :-)
[20:17:32] <tuanpembual> still problem when install theme like current progress. Will work from that.
[20:17:45] <tuanpembual> for ichain, need more sugestion.
[20:18:06] <tuanpembual> done from me.
[20:18:36] <cboltz> I prefer an ugly-looking new progress.o.o with the default theme over a nice-looking but terribly outdated one ;-)
[20:19:04] <cboltz> (besides that, for the theme you might want to talk to lcp so that we get the latest design while working on it)
[20:19:19] <cboltz> for ichain - which suggestions do you need?
[20:19:26] <lcp> I see there is a project for adapting bootstrap to redmine
[20:19:36] <lcp> but it's 5 yo and unmaintained
[20:20:00] <lcp> I will get on making something based on that then ;)
[20:20:26] <tuanpembual> maybe, tried to get in touch with ancrogs? -> Ancor Gonzalez Sosa
[20:20:38] <tuanpembual> kbabioch suggest me to do that.
[20:20:52] <tuanpembual> thanks lcp.
[20:21:04] <lcp> ancorgs might be busy with yast stuff, and some of his projects still are using older bootstrap :c
[20:21:14] <kbabioch> tuanpembual: yes, i think he is the original author of the plugin, so he probably knows a thing or two about it ;-)
[20:21:35] <lcp> I will need to get all of that up to date at some point
[20:22:31] <lcp> but when I looked into tsp for example, tw didn't have old enough ruby to run it so I gave up for now
[20:23:19] <cboltz> moving tsp to a new VM is already in ancorgs' queue - we'll see if some of the old code explodes on Leap 15.1 ;-)
[20:23:39] <lcp> it will most likely, it's rails 4 with old old ruby
[20:23:59] <lcp> but his daily stuff is ruby, he can manage >:D
[20:24:11] <pjessen> i'm back.
[20:24:12] * Pharaoh_Atem waves
[20:24:15] <Pharaoh_Atem> I'm back
[20:24:41] <cboltz> lcp: I think so, he'll "just" need some time to do it ;-)
[20:25:09] <lcp> "just" some time is scary with some things in openSUSE
[20:25:28] <cboltz> looks like you understand the problem ;-)
[20:26:24] <lcp> I mean, myself I try to get through with everything as fast as I can
[20:26:40] <lcp> but that's not possible with everything, so there is a huge backlog ;)
[20:26:54] <cboltz> same for me ;-)
[20:27:37] <cboltz> I did some preparations for updating to a new MediaWiki, but a) I'm waiting for a new VM for the new elasticsearch version and b) didn't really have time to work on it in the last weeks
[20:28:41] <cboltz> on the positive side, MediaWiki updates are usually boring (unless you have fun with an extension)
[20:29:26] <lcp> there were some suggestions to use markdown extensions in mediawiki, but I can only imagine the amount of work it would actually take
[20:29:56] <lcp> although wiki markup is plain confusing even when the explanation is always on screen :P
[20:30:46] <cboltz> actually you can already use markdown if you host the markdown file on github ;-)
[20:30:52] <cboltz> (we have an extension for that)
[20:30:57] <lcp> oh?
[20:32:54] <cboltz> https://github.com/JeroenDeDauw/GitHub#Usage
[20:33:26] <lcp> this is kind of amazing
[20:37:07] <bmwiedemann2> I'm off. gn all
[20:37:19] <lcp> night
[20:38:01] <tuanpembual> I'm off to, good morning all.
[20:38:03] <cboltz> tuanpembual, lcp - it seems there's an actively maintained bootstrap theme for redmine - "PurpleMine2" on http://www.redmine.org/projects/redmine/wiki/Theme_List
[20:38:38] <lcp> eh, uses some parts might not be good enough
[20:38:54] <cboltz> right, but it's probably better than something from 2014 ;-)
[20:39:12] <lcp> that's correct :P
[20:40:04] <lcp> gah, it uses node.js and npm to customize stuff
[20:40:13] <lcp> awful
[20:40:40] * lcp might be weird though, makefiles with calls to sass don't sound much better
[20:41:29] <lcp> cboltz: we should start on employing ipsilon to replace current openid stuff
[20:41:49] <lcp> that way we can move everything over to the local servers
[20:42:21] <lcp> and that can be plugged directly to existing freeipa server
[20:42:30] <cboltz> I know nothing about openid and ipsilon, but everything that includes getting rid of a server in Provo sounds good :-)
[20:42:43] <cboltz> note that freeipa is only used for the heroes accounts
[20:42:57] <lcp> yeah, I wonder how to handle this
[20:43:02] <cboltz> but not for the openSUSE/SUSE/Attachmate/Novell account
[20:43:20] <lcp> time to split openSUSE from the rest >:D
[20:43:34] <pjessen> question about the vpn?
[20:43:53] <cboltz> lcp: several people are working on that, and the plan is to have the split done in about a year
[20:44:13] <cboltz> and that also includes "a solution"[tm] for handling accounts
[20:44:24] <cboltz> (there are some ideas, but no decision yet)
[20:44:31] <cboltz> pjessen: ask ;-)
[20:44:39] <pjessen> if I log in from two different machines, I get the same IP address.
[20:44:46] <lcp> Pharaoh_Atem: where does RH keep their code for BZ?
[20:45:01] <Pharaoh_Atem> lcp: unfortunately, it's kinda private still
[20:45:01] <pjessen> which doesn't work
[20:45:08] <lcp> Pharaoh_Atem: dang it
[20:45:26] <lcp> open source </3 RH >:D
[20:45:27] <Pharaoh_Atem> lcp: but I'm pretty sure you can reach out to them about getting specific parts of it open sourced to incorporate on top of BMO (which is their base codebase)
[20:45:27] <cboltz> pjessen: right, the IPs are hardcoded for each user, so that isn't unexpected
[20:45:48] <Pharaoh_Atem> lcp: there's a BZ tracking open sourcing the whole thing
[20:45:51] <cboltz> looks like nobody thought about people using multiple machines to access the VPN ;-)
[20:46:09] <Pharaoh_Atem> lcp: https://bugzilla.redhat.com/478886
[20:46:11] <cboltz> if it helps you, I could setup a user "pjessen2" for you ;-)
[20:46:41] <lcp> Pharaoh_Atem: it would be nice to get BZ handled with mixed accounts (FAS/RH) into SUSE's new bugzilla
[20:46:48] <pjessen> it was an exception, and I can't be bothered to work with two ids. why isn't it just done with dhcp?
[20:46:56] <lcp> and I hope it's bugzilla otherwise imma go and murder
[20:47:27] <Pharaoh_Atem> lcp: I'm pretty sure we can get in touch with the right people and get that chunk of it opened up to use on SUSE's new bugzilla
[20:47:44] <cboltz> pjessen: I'd _guess_ the intention was to make it easier to track you, for example if you ssh root@$machine the IP makes clear who you are
[20:48:28] <Pharaoh_Atem> lcp: if you have an rhbz account, you can ask about progress in there :)
[20:48:32] <lcp> cboltz: could we maybe start deploying a login system for openSUSE, just so we have it, and from there ask SUSE for a mixed account access to SUSE new bugtracker?
[20:49:00] <pjessen> cboltz: yeah
[20:49:31] <cboltz> lcp: there's no final decision yet. I can tell you that bugzilla is one of the options, and bugzilla is the board's recommendation to SUSE
[20:49:35] <lcp> cboltz: although frankly for a new account system it runing on Fedora won't look good >:D
[20:49:56] <lcp> Pharaoh_Atem: what are the issues with running it on openSUSE actually
[20:50:20] <Pharaoh_Atem> shouldn't be much... just time and need help getting everything packaged
[20:50:35] <lcp> I mean freeipa, not the "rest"
[20:50:52] <Pharaoh_Atem> we need dogtag packaged in opensuse, and then the pki stuff needs to be made to work
[20:51:01] <cboltz> actually I like the irony of running parts of the openSUSE infrastructure on Fedora ;-)
[20:51:03] <Pharaoh_Atem> opensuse lacks pki paths and such
[20:51:18] <cboltz> (just wondering - does Fedora use an openSUSE machine? ;-)
[20:51:21] <Pharaoh_Atem> cboltz: I mean, for a while, we ran openqa on openSUSE until we got everything working
[20:51:36] <cboltz> :-)
[20:51:36] <lcp> cboltz: if SUSE continues with their path with OBS, it might be the next instance >:D
[20:54:19] <lcp> cboltz: I am afraid SUSE will go with something random, although maybe it will have a standard openid stuff so we can connect anyway >:D
[20:55:03] <cboltz> the login system is on the to-be-decided list
[20:55:29] <cboltz> and it's not even clear if we'll keep the shared login system or if SUSE will have its own
[20:55:43] <cboltz> (obviously SUSE employees would then also need an openSUSE account for bugtracking etc.)
[20:56:04] <lcp> I don't think we should, and I think we should deploy new login system before SUSE does
[20:56:56] <mstroeder> Sorry, I'm late but reading login I'm all ears.
[20:57:12] <cboltz> ;-)
[20:57:13] <Pharaoh_Atem> cboltz: it would not surprise me if we still have an openSUSE machine somewhere in the Fedora infra :)
[20:58:01] <lcp> like, crossbreading SUSE and openSUSE infrastructure doesn't make much sense at all, it's convinient to have the same account for bugzilla for both, but that can be fixed with two providers
[20:58:42] <cboltz> Pharaoh_Atem: the evil answer would be "so you have at least one machine with a good OS" ;-)) - but actually I'm a big fan of cross-distro collaboration, so running another distro where it makes sense is a good thing IMHO
[20:58:46] <lcp> but from the pov of SUSE, I wouldn't trust anything that happens in for example connect logins >:D
[20:59:15] <lcp> only god knows what truly happens there
[20:59:31] <Pharaoh_Atem> cboltz: one day the SELinux policy will work in openSUSE :)
[20:59:52] <cboltz> lcp: the good thing about connect is that it uses the login proxy, which means it only sees usernames, but no passwords ;-)
[21:00:17] <cboltz> Pharaoh_Atem: yeah, and Fedora will come with AppArmor support ;-)
[21:00:35] <cboltz> if you want to work into that direction, I'd be happy to help ;-)
[21:02:09] <lcp> cboltz: good question is, should a new login system be based on existing freeipa?
[21:02:46] <lcp> should a portion of accounts from SUSE db be imported into the new system?
[21:02:53] <cboltz> I don't know ;-)
[21:03:07] <cboltz> there's also the idea of using Ædir
[21:03:40] <lcp> me neither, those are good questions that are worth considering :/
[21:03:44] <cboltz> my personal impression is that its concept looks more secure, but the web interface of freeipa looks better ;-)
[21:04:15] <lcp> better web interface >>> secure
[21:04:19] <lcp> for sure >:D
[21:04:51] <cboltz> ;-)
[21:05:07] <mstroeder> Anyone here saw my Æ-DIR talk at oSC 2019?
[21:05:34] <kbabioch> <- not live, but watched the video :-)
[21:06:04] <lcp> maybe we could track that, establish a new login system for openSUSE, track suggestions in that regard and start working on it
[21:06:26] <lcp> establish that, get ipsilon running
[21:06:33] <mstroeder> The reason I'm asking: Æ-DIR differs quite a lot from other LDAP-based user management systems because it has a clear distinction of persons and their multiple user accounts.
[21:07:33] <lcp> and then deploy noggin when it's ready, because it clearly is wip
[21:07:41] <lcp> I should get myself over there and help a bit
[21:07:56] <lcp> mstroeder: I will have to watch it :P
[21:08:01] <mstroeder> There are two Æ-DIR test servers running in infra.o.o, see https://progress.opensuse.org/issues/39872
[21:08:50] <mstroeder> Whoever wants to test e.g. integrating a Linux machine with aehostd, just drop me a note or comment in the PoC ticket.
[21:08:52] <lcp> and via noggin we could manage memberships
[21:09:27] <lcp> although, with managing memberships, the quicker the connect replacement the better
[21:09:56] <cboltz> right, that shouldn't be blocked by the login system
[21:10:06] <lcp> yeah, that's problematic as is
[21:10:43] <lcp> although having it implemented right in loggin system is way better than having it seperate
[21:11:39] <Pharaoh_Atem> cboltz: well, nothing stops Fedora having AppArmor except for someone wanting to do it :P
[21:11:42] <lcp> cboltz: I wonder if we couldn't get logging system now, get required accounts from Provo now and import them into existing freeipa or aedir now
[21:11:53] <lcp> although, provo and now are very seperate
[21:11:54] <Pharaoh_Atem> as for SELinux, I've been interested in the past, though I was also told no one wanted it before :(
[21:12:15] <Pharaoh_Atem> (I had done it at least once before when I ported OpenShift to openSUSE)
[21:12:48] <Pharaoh_Atem> lcp: ipsilon, as it stands, should work with Æ-DIR or FreeIPA
[21:13:03] <lcp> that's good
[21:13:13] <Pharaoh_Atem> I'm not sure if CAIAPI (which is the basis of noggin) will support Æ-DIR as it is
[21:13:15] <lcp> support for aedir would probably need to be added to noggin
[21:13:27] <Pharaoh_Atem> the main assumption for Noggin is that it can offload _a lot_ to FreeIPA
[21:13:28] <lcp> not that it supports much now anyway >:D
[21:15:12] <lcp> Pharaoh_Atem: well, noggin has a long way to go anyway, might as well start now instead of waiting >:D
[21:15:13] <mstroeder> Could you please point me to CAIAPI (noggin) docs regarding login mechs?
[21:15:34] <Pharaoh_Atem> mstroeder: it's all in here: https://github.com/fedora-infra/noggin
[21:15:43] <Pharaoh_Atem> it's all in-progress and under active development
[21:16:01] <Pharaoh_Atem> so in theory, now would be a good time to look to support more than one IPA system
[21:16:18] <mstroeder> My idea for Æ-DIR would be: Whatever membership management system is used the project members should be synced to Æ-DIR as aePerson entries for which connected aeUser entries are added.
[21:17:22] <pjessen> do we have anything specific left for the Heroes? otherwise I'm gonna go watch telly.
[21:17:47] <mstroeder> So noggin is just a Flask web app and it should not be too difficult to add support for WebSSO mechs or plain LDAP.
[21:18:09] <Pharaoh_Atem> mstroeder: yeah, as it stands, it's pretty simple
[21:18:22] <lcp> pjessen: I mean, I have mentioned everything regarding basis of separate login system for openSUSE, the details will have to be mentioned later
[21:18:36] <Pharaoh_Atem> noggin is a frontend for CAIAPI, and ipsilon is the SSO component for the noggin system
[21:18:41] <lcp> we should have an issue for tracking development on this stuff
[21:18:52] <cboltz> pjessen: we have two more agenda items - the forum issues in the last days (not sure if we can do much about that) and Q&A (whoever has a question, just ask!)
[21:19:15] <cboltz> and I have the feeling that the login discussion will continue for some more time ;-)
[21:19:32] <lcp> it will be on our minds and on our keyboards >:D
[21:19:44] <lcp> that's a nerdy way to say that
[21:19:47] <mstroeder> I already looked at ipsilon. Despite its dependency on lasso libs I like the idea of an IdP written in Python.
[21:20:11] <lcp> and it would free us from novell openid provider >:D
[21:20:41] <lcp> this https://www.opensuse.org/openid/
[21:21:03] <lcp> and would expand it a little, just old openid is not enough nowadays >:D
[21:22:02] <mstroeder> But it's hard to tell for an outsider whether ipsilon is still actively developed.
[21:22:28] <lcp> "well enough"
[21:22:31] <mstroeder> Today one would like to use 'OpenID Connect' as WebSSO protocol anyway.
[21:22:41] <Pharaoh_Atem> yes, and ipsilon is an oidc idp
[21:22:47] <Pharaoh_Atem> as well as a saml2 and oauth2
[21:22:53] <pjessen> okay, so lets get back on track. what is up with the forums?
[21:23:32] <cboltz> there were several outages - both complete downtimes and users getting logged out while trying to post
[21:24:10] <pjessen> is it a matter of tech support? I havent seen anyone providing any useful answers.
[21:24:49] <knurpht> pjessen: that's most likely because the behaviour ( and resulting messages ) is inconsistent.
[21:24:56] <cboltz> so far, I only got a "will check" answer, but nothing really useful
[21:25:21] <cboltz> okurz: did you get something more useful in the tickets you opened?
[21:25:27] <pjessen> still, whoever is supporting the forum setup ought to be able to help out.
[21:26:37] <knurpht> pjessen: Kim G ( MF IT ) in the past said it's headers ....
[21:27:06] <knurpht> If I could get reproducible behaviour, I would have reported.
[21:27:47] <knurpht> In many occasions over the last couple of months this happens during weekends.
[21:28:28] <knurpht> Error messages vary
[21:28:49] <pjessen> I feel we (Heroes) have an obligation to help the community, but from where I am sat, I'm helpless.
[21:29:09] <knurpht> pjessen: so is the forums team.
[21:32:15] <pjessen> I have to ask - does anyone have a proposal how to proceed with this issues?
[21:33:51] <knurpht> pjessen: personal: move stuff away from provo, start a discourse instance ( with bridge to MLs ). But, I'm part of a team ...
[21:34:25] <cboltz> I'm afraid we'll have to go via the board -> SUSE management route
[21:34:42] <cboltz> and I fully agree with knurpht that we should setup something under our control
[21:34:46] <knurpht> cboltz: I already put it ( once again ) on the agenda
[21:35:26] <cboltz> maybe you should also send a mail to board@ and ask Richard to escalate it _now_ instead of waiting for the board meeting
[21:35:44] <pjessen> I agree, so we just need someone to do it.
[21:36:19] <lcp> https://github.com/openSUSE/chameleon-discourse just in case
[21:36:22] <lcp> mentioning
[21:36:54] <Pharaoh_Atem> lcp: no pictures? for shame!
[21:36:56] <knurpht> I'm already planning to install a discourse try out on my own VPS, just to see how it works.
[21:37:19] <lcp> Pharaoh_Atem: I posted A LOT to discord :P
[21:38:07] <knurpht> lcp: Discord is a different beast.
[21:38:19] <lcp> that's true
[21:38:35] <knurpht> and not FOSS
[21:38:44] <knurpht> discourse is.
[21:39:01] <knurpht> Not really clear from their website, but it is.
[21:39:29] <lcp> https://cdn.discordapp.com/attachments/407993213425680384/561953286786383912/Screenshot_from_2019-03-31_18-39-41.png
[21:39:38] <lcp> I had light theme version somewhere
[21:39:44] <lcp> also I changed that lower bar somewhat
[21:39:56] <lcp> but in general, it's a simple rebrand, with not that many changes
[21:39:57] <pjessen> how about gatewaying between discord/course and nntp? :-)
[21:40:14] <okurz> cboltz: no further useful info in the tickets, sorry
[21:40:19] <lcp> pjessen: we have matrix/irc bridges set up
[21:40:28] <cboltz> okurz: I'm not surprised
[21:40:31] <knurpht> pjessen: should be possible
[21:40:49] <knurpht> lcp: bridge to matrix works only half of the time
[21:40:57] <knurpht> if even
[21:41:03] <lcp> because it's not hosted on our infra
[21:41:09] <lcp> I really need to get back to this
[21:41:39] <knurpht> And does not offer a forums-like structure
[21:41:45] <pjessen> okay, I'm just asking, at first it was not easy with forum<>nntp, and with the nntp servers being deprecated, we also need someone to set new ones up.
[21:41:47] <knurpht> discourse does
[21:42:12] <knurpht> pjessen: yes, forums team is worried about this.
[21:42:15] <lcp> oh yeah, I'm not suggesting discord/matrix/irc as a replacement for forums :P
[21:42:21] <lcp> there is a place for both
[21:42:37] <knurpht> discourse could do forums+MLs
[21:42:45] <knurpht> maybe even more
[21:42:51] <lcp> yes
[21:42:56] <lcp> well, hyperkitty exists
[21:43:05] <knurpht> and IIRC it's able to import the forums content.
[21:43:19] <lcp> https://lists.fedoraproject.org/archives/
[21:43:19] <knurpht> and single sign-on
[21:43:33] <lcp> it's very forum like
[21:43:58] <lcp> but I need to fix it so it works without js
[21:44:12] <lcp> because the amount of js there just annoys the hell out of me
[21:44:13] <knurpht> https://www.discourse.org/features
[21:44:44] <pjessen> volunteers are welcome :-) I have not managed to find time to migrate away from mlmmj in the last 2-3 years.
[21:44:51] <lcp> knurpht: yeah, I'm talking about mailing lists now
[21:45:28] <Pharaoh_Atem> pjessen: lcp and I have been looking at it
[21:45:57] <Pharaoh_Atem> it's on my todo to deal with packaging it, and lcp plans on contributing some improvements to HyperKitty UI as well
[21:47:21] <lcp> yes, plans
[21:47:31] <lcp> plans
[21:47:34] <lcp> ehhh
[21:47:45] <Pharaoh_Atem> well, lcp _will_ do it soonish :P
[21:48:02] <knurpht> lcp: I plan to be 100 years old...... at 58 now.
[21:48:28] <Pharaoh_Atem> knurpht: I'd be impressed if you pulled that off with you smoking like that
[21:48:34] <lcp> I mean, I am gonna die from stress and caffeine before 30 at this rate
[21:49:38] <knurpht> Pharaoh_Atem: there's a slight difference between plans and reality. Reality says I will reach 94
[21:49:46] <knurpht> :P
[21:51:55] <pjessen> back to near reality perhaps - anyone able to help with setting an nntp server for the forums?
[21:52:29] <mstroeder> Is there still actively maintained NNTP server software?
[21:52:45] <mstroeder> inn still alive?
[21:53:25] <pjessen> afaik, yes. it's in leap 15.1
[21:53:34] <cboltz> I know nothing about nntp servers, but I can probably help on the salt side (if someone tells me what it should do ;-)
[21:54:45] <pjessen> I hate to admit it, but I can probably do the nntp side, I just don't know anything about the forum side.
[21:54:49] <mstroeder> My last INN setup was 20 years ago...
[21:55:05] <pjessen> yeah, mine was in 2006
[21:55:28] <Pharaoh_Atem> the last time I *touched* an nntp system was in 2000
[21:55:33] <Pharaoh_Atem> so I have no idea :P
[21:56:22] <cboltz> knurpht: can you help on the forum side?
[21:58:21] <knurpht> cboltz: sure, when- and wherever I can
[21:59:14] <pjessen> we would need someone to provide a lists of groups as well as configure the delivery to a new nntp server.
[22:03:19] <pjessen> it wud also be nice if someone would actually drive this migration, I would prefer not to take it on myself.
[22:08:06] <pjessen> massive silence :-) guys, I gotta go. cboltz and knurpht lets talk on email. have a nice Pfingsten everyone!
[22:08:25] <Pharaoh_Atem> well, I'm mainly here to talk about login things :)
[22:08:37] <Pharaoh_Atem> so there's not much for me to say about nntp :)
[22:08:47] <cboltz> pjessen: you'll always get silence if you ask for volunteers ;-)
[22:09:02] <lcp> with exception of login systems
[22:09:13] <lcp> too many people jumped on that >:D
[22:10:41] <cboltz> I'm quite sure the number of people will decrease once we decided on a solution and ask who will implement it ;-)
[22:12:04] <lcp> yeah, I really like that mstroeder suggested aedir, it's a great idea if implemented in noggin :P
[22:12:13] <lcp> but implementation will be a pain non-the-less
[22:12:40] <mstroeder> Which implementation is a pain?
[22:12:58] <lcp> implementation of login system in infra >:D
[22:13:31] <lcp> I have no idea how hard implementing aedir in noggin will be, haven't look deep enough to know
[22:15:13] <lcp> I think I will take a break from yast for a while and implement a theme for redmine now
[22:15:25] <mstroeder> Æ-DIR test systems are there upgraded to Leap 15.1. Installation is automated with ansible, including details like pretty paranoid AppArmor profiles. If the openSUSE project decides to use that I'm up to help with everything related to that.
[22:16:33] <lcp> meanwhile infra uses salt >:D
[22:16:51] <knurpht> pjessen: I'd gladly join the heroes team and help out, but I'd require some serious training / intro into our maintenance / infra
[22:17:25] <mstroeder> Yes, I know that you use salt. But Æ-DIR installation is based on ansible. The base config of the servers can be done with salt.
[22:17:43] <cboltz> lcp: in germany we have a saying "competition vitalizes the business" ;-)
[22:18:22] <lcp> cboltz: I don't think there is much competition, I'm absolutely fine with aedir if noggin implementation exists (although it does not currently)
[22:18:35] <knurpht> lcp: in the board we say "Let's have cboltz's couple of things first" :D
[22:18:52] <cboltz> "competition" was about salt vs. ansible ;-)
[22:19:07] <lcp> ah, yes
[22:19:18] <mstroeder> Æ-DIR's ansible roles do not apply base OS configuration. So there's no competition at that level.
[22:19:58] <cboltz> yeah, but it's still funny if you tell someone that you use salt _and_ ansible to manage the same server ;-)
[22:21:31] <lcp> add puppet to that
[22:21:37] <lcp> and chef
[22:21:46] <knurpht> lcp: and and and
[22:21:51] <cboltz> lol
[22:22:11] <cboltz> knurpht: I can create your heroes VPN account and get you started - let's talk about the details later or in the next days
[22:22:16] <lcp> and manage pip, apt, zypper, npm etc with the configs
[22:22:29] <knurpht> cboltz: Wow ...
[22:22:33] <knurpht> thanks
[22:23:36] <cboltz> after a while, you'll find out that I'm just evil and will offload work to you ;-)
[22:23:39] <knurpht> Don't have time the next days/evenings ( sigh, stupid kids ) but I gladly accept that
[22:24:02] <knurpht> cboltz: you know I'm good at ping-pong?
[22:24:23] <cboltz> ;-)
[22:24:56] <knurpht> consider me a walking learning curve.
[22:25:18] <knurpht> i.e. for others :P
[22:27:41] <cboltz> ;-)
[22:28:59] <knurpht> Haris: "Every community needs someone bad to show the rest are good people. Well, the bad guy is Gertjan" ......
[22:29:49] <lcp> I didn't get that impression
[22:30:17] <knurpht> lcp: I explained the same to his daughter, the other way around.
[22:30:24] <knurpht> In German
[22:30:37] <knurpht> She had a good laugh
[22:32:03] <knurpht> And to be honest: https://cloud.knurpht.nl/s/xCj5fkyzKbSHQQP

    (1-1/1)