[19:00:34] so let's start? [19:00:40] gogogo [19:00:45] i have 20 min left [19:00:46] go! [19:01:19] roll call please [19:01:29] here [19:01:43] * cboltz hides [19:01:47] Per, do you want to moderate today? [19:02:14] here [19:02:55] Theo, do you know something about Lars? [19:03:17] Lars is off for the weekend as i know [19:03:23] ok [19:03:26] so he maybe will not join [19:03:30] ticket is here https://progress.opensuse.org/issues/17272 [19:03:57] anything else that is NOT on the topics' list on that ticket that needs to be discussed? [19:04:04] no tnx, I might have to be going to and from. [19:04:23] Have we got any questions by new people in the Heroes team? [19:05:30] So the first topic is finished. [19:05:34] I don't think so [19:05:52] Nect topic: connect.o.o [19:05:59] so second: connect.o.o [19:06:25] |miska| / plusky around? [19:06:28] What is the status here? [19:06:52] plusky is listed on our user list. [19:07:15] yes but he might know something about connect [19:07:53] okay let's skip it, I don't think there is anything to discuss [19:08:01] I know, that the Travel Support Application should get an approvement by GSoC. And that's part of connect. [19:08:11] they implemented some temp solution for the elections [19:08:15] not exactly [19:08:31] tsp is a ruby app, that is running under the connect.o.o domain [19:08:39] but appart from that, it is a totally separate webapp [19:08:46] ok [19:08:48] what kind of approvement it needs? [19:08:56] it is already running for quite some time [19:09:30] "Make the Travel Support Application useful for more organizations" is the topic on 101 [19:10:07] this is the url https://connect.opensuse.org/travel-support [19:10:12] you have a link for that? [19:10:26] http://101.opensuse.org/ [19:10:32] Ctrl-F "travel support" [19:10:33] http://101.opensuse.org/ under others [19:11:53] okay found it [19:12:06] anyway, let's proceed, there's nothing for us to discuss here I think [19:12:14] are there (besides TSP) other parts on connect.o.o which are not part of the PHP code? [19:12:14] never even knew that site existed [19:12:31] no, only tsp [19:12:54] next topic: DNS [19:13:50] What is the status here? [19:14:00] last what i heard from lars [19:14:06] it's in preparation [19:14:13] o.o will be in NUE at near future [19:14:15] but [19:14:21] we won't get the PTR-records [19:14:28] so we're taking it over from MF-IT? [19:14:33] yup [19:14:38] Yes. "In preparation" in Provo? [19:14:40] o.o zone [19:14:43] sounds good [19:14:48] but not PTR zone [19:15:12] and where it will be? in atreju? [19:15:16] because PTR is mixed IP-range between several MF services [19:15:24] Which time is planned for that until the next escalation? [19:15:24] tampakrap: i think so... [19:15:28] PTR is probably our smallest problem ;-) (except for servers sending out mails) [19:15:55] cboltz: PTR is of course still possible :) [19:16:04] but MF IT must do it ... [19:16:07] good [19:16:28] don't know a time frame - topic is handled by lars [19:16:42] cboltz: mail is MX and PTR resolving IPs [19:17:00] right [19:17:07] Ada_Lovelace: i think cboltz knows about that :) [19:17:23] PTR is important for almost anything including checks like mailservers spam check [19:17:33] the point is - if a server does _not_ send out mails, nobody cares about the PTR [19:17:38] or pfs [19:17:47] cboltz: maybe for PFS [19:17:55] maybe for https [19:18:00] i think browsers check ptr if you turn on PFS [19:18:41] I have > 100 domains on one server, so which PTR should I set? ;-) [19:18:56] so I slightly ;-) doubt a "wrong" PTR will break anything in the browser [19:19:03] without SNI, you'd need 100 IPs too [19:19:14] :) [19:19:15] that's why I use SNI ;-) [19:19:41] anyway, we're straying off-topic [19:20:24] We'll give MF-IT some time to do their work. Adrian is the new person for escalations... [19:20:53] will the o.o. zone be delegated to NUE ? [19:21:10] yes [19:21:55] okay so I imagine that this will be probably the SUSE-IT's responsibility, and not the openSUSE heroes though [19:22:31] we will see tampakrap maybe we can integrate it with the o.o FreeIPA. but hey first let the zone come [19:22:34] so, it won't be in the same saltmaster as the other heroes machines [19:22:43] correct [19:22:55] yes, but MF-IT will need somebody for kicking :P [19:23:01] but yeah I like it, it would be still a huge improvement [19:23:22] Ada_Lovelace: Lars is communicating directly with MF IT about that topic [19:23:27] keep calm everybody [19:23:49] next topic? [19:23:50] so let's move, we are making only assumptions here and I don't like it :) [19:24:04] next topic: monitoring and status page [19:24:26] Lars gave me access to the monitoring system. [19:24:49] That's a Icinga and I add all checks I think we need. [19:25:19] which is an atreju VM? [19:25:22] Ada_Lovelace: which domain does it run? [19:25:24] we need to add it in salt? [19:25:34] What is with messages for SUSE-IT admins? Do you want to get mails or alarms to any admin phone? [19:25:34] or is it already? [19:25:36] * tampakrap checks [19:26:02] http://monitor.opensuse.org/icinga/ [19:26:09] Ada_Lovelace: nope. no phone at all here. if i got the domain and the access rights from lars [19:26:15] i will add it to my browser plugins [19:26:24] than i'll have an eye on it [19:26:42] I don't mind to have sms notifications but for critical issues only [19:26:59] Ok. Then I let this configuration and concentrate me only on plugins & Co. [19:27:38] i don't think we have a phone-gateway? like sms on this server [19:27:39] just fyi [19:27:46] critical issues to your phone? [19:28:01] Ada_Lovelace: monitor.opensuse.org is in salt-key -L output, but it is not in pillar/id/monitor_opensuse_org.sls, add it please [19:28:34] * thomic has to leave, i will read the backlog. have a nice meeting. bye. [19:28:34] Ok. [19:28:55] how do we log in? [19:29:20] Lars did the setup and I should do the configuration. I want to do that. [19:29:36] okay I'll wait then [19:29:44] We have got a user icingaadmin by Lars. [19:29:57] tampakrap: I remember some username and password mentioned in a ticket ;-) [19:30:10] Yes it is there. [19:30:15] ah okay [19:30:39] https://progress.opensuse.org/issues/15260 [19:31:02] I'll check it [19:31:06] thanks for working on this [19:31:30] anything else or we can move on? [19:31:50] The status page is missing. Do you know the status there? [19:32:32] apart from the mails that lars sent at the heroes mailing list, I don't know anything more [19:32:52] Ok. Then let's go to the next topic. [19:33:21] Next topic: Ticket triage status and Due Dates [19:34:57] Martin and Gerhard worked faster than darix. :-) [19:34:58] 133 open tickets yey! [19:35:23] I'd prefer 30 open tickets, but 133 is already a big improvement :-) [19:35:47] Adrin has got an eye on the ticket queue, too. (Rich told us) [19:36:02] *Adrian* [19:36:12] also darix did a "my assigned stuff" page [19:36:18] also very useful [19:36:46] I'm not if the URL is the same for everybody, to me it shows https://progress.opensuse.org/projects/opensuse-admin/issues?query_id=40 [19:37:03] The result: We have to assign him tickets with his stuff... [19:37:03] I see the same [19:37:30] ditto [19:37:40] redmine supports querying for "assigned to: << me >>" which is a nice feature [19:37:54] Yes. That's great! [19:38:02] and the query seems to include tickets assigned to groups where you are member of the group [19:38:13] * cboltz wonders who added him to opensuse-admin-provo [19:38:36] I can remove you [19:38:46] How do we want to do it in the future? Do we need communiy people for assigning them to special people? Or is that the job by Max now? [19:38:54] it was not meant as a complaint, so keep me in ;-) [19:39:56] Ada_Lovelace: I don't get the question sorry [19:40:30] How do we want handle it with assigning issues to special persons? [19:40:37] so fun fact: in opensuse-admin-provo there is noone from MF-IT. there are two suse people that are in provo though, robert and craig [19:40:48] what do you mean by special? [19:40:55] Do we need community people or is that the job by Max Maher now? [19:41:41] darix is doing only issues which are assigned to him and doesn't look after the topics in the queue... [19:42:10] cboltz and I assigned all issues to darix and Martin until now. ^^ [19:42:35] darix is using the view "assigned to me". [19:42:35] ah, so your question is "who does assign a ticket to the person/group who has to handle it"? [19:42:37] i look over issues every now and then, and reassign [19:42:50] Yes! :) [19:43:08] And we have got Max as a new admin for us, too. [19:43:15] feel free to assign I'd say, and if you are in doubt you can join this channel and ask [19:43:32] at least that's what I do [19:43:40] and nobody complained so far :) [19:44:51] is that clear, or I still didn't get the question? [19:45:04] Do we need Due Dates? Lars said "Use it!", Rich said "Don't use it!" [19:45:18] Yes. It is clear. [19:45:30] I vote for "use them" [19:45:52] I say use them when it makes sense. It won't always. [19:45:54] it is a good reminder for important or long standing issues [19:46:00] sure, not always [19:46:22] 1 month like in the last meeting? [19:46:46] IIRC we decided to set it to two weeks ;-) [19:46:47] depends on the case, but 1 month in general sounds good to me [19:47:02] a default due date? [19:47:55] Yes. A default due date for tickets with special assignees (admins). ^^ [19:48:07] okay] [19:48:31] i guess over-due tickets will be item #1 on the next meeting agenda :-) [19:48:33] We used it for mirror and BuildService tickets until now. [19:48:42] so I would say: if the ticket is straightforward (add a mirror, add an account, add an ml) sure set the due date [19:48:58] yup, that makes sense [19:49:03] if the ticket is for something more complex that needs time/planning, then set it to the far future, or don't set it at all [19:49:18] I have a few salt related tickets for example that can't have a due date [19:49:25] Ok. [19:49:41] would it make sense to let the admin@o.o mail handler script the due date? IMHO > 90% of the mails to admin@o.o are "straightforward" [19:50:09] It can't set the special admin... [19:50:22] I think it doesn't make sense, I would prefer the guys that are doing the wrangling to decide if it needs to be set [19:50:41] as the wrangler needs also to decide about the group and the assignee, and the privacy of the ticket [19:51:44] Other questions about our ticket queue? [19:51:59] question - is it sufficient to set the category when assigning? [19:52:50] I believe "yes", because special people are looking after the category for fixing issues... [19:53:32] the category is important yes [19:53:40] the problem of only setting the category (but no assignee) is that those people won't get a mail [19:53:52] in my todo is to create cached search links based on categories [19:54:00] I usually only set category and never bother with the assignee [19:54:01] and put them on our wiki page [19:54:37] cboltz: sure, that's why I want to have the list of tickets based on category as well :) [19:54:43] but yes, you raise a good point [19:54:58] @pjessen: mirror to mcaj, creating of obs projects to darix and the rest to the special admin group [19:55:07] there's already a "Tasks by Category" query in progress [19:55:17] also, if people have set their progress.o.o settings to get all mails, they can filter the ticket notifications based on the categories as well [19:55:40] will be away for 10mins, gotta go pickup my wife [19:56:04] the opensuse-admin group is something that I would kill tbh, it serves no purpose [19:56:29] it is also confusing, because people think that the ticket is assigned while actually noone may be working on it [19:56:35] I believe Lars wants to have it... [19:56:54] that group will send mails to all of us, so it might still make sense [19:57:05] I would prefer people to fix their mail notification setup instead, on their progress profile [19:57:28] cboltz: that's my point, if you fix your notification settings you will get the mails [19:57:52] I get *all* the mails, but not everybody might like that flood ;-) [19:58:01] filter them then based on category [19:58:04] That's it! :) [19:58:10] (redmine vs. mlmmj fights are especially interesting ;-) [19:58:44] I believe that's a different problem, and yes it needs to be fixed [19:58:50] I reconfigured my mail configuration on github after joining owncloud, too (with this reason) :-D [19:59:32] this topic is taking long, can we please move it to the mailing list and proceed now? [19:59:37] BTW: Can we have a list of the people in all opensuse-admin* groups? [19:59:50] the redmine page listing group members goes 403 for me [19:59:59] cboltz: remind me on the mailing list and I'll give them to you [20:00:01] Oups... [20:00:22] Next topic: new wiki [20:00:40] @cboltz: Show us elasticsearch! :D [20:01:02] as you already heard before the meeting, Sarah and I did the search setup [20:01:25] I also changed the default search to search in all namespaces (except talk pages) [20:01:59] and assigned weights to every namespace - for example, pages in the "Archive:" namespace are less likely to show up as top result [20:02:41] one detail I don't like to much is that MediaWiki 1.27 (which is the current LTS version) still requires ElasticSearch 1.7.x [20:02:51] because that series is EOL since a few weeks [20:03:21] I commented on the (already existing) upstream bug and hope they'll backport support for ElasticSearch 2.x [20:03:45] or 5.x any time... [20:04:11] that would be even better, but I somehow doubt 1.27 will get support for 5.x [20:04:13] 5.x won't be supported at the moment [20:04:44] on completely unrelated news: [20:04:57] I'm still waiting for the google maps API key for the wiki [20:05:24] I already asked several people in Provo, but didn't receive it since months [20:06:14] Perhaps somebody from SUSE-IT can create a new API key as the last solution? [20:06:22] (and was slightly annoyed when I heard "now that you are a board member, we should be able to give it to you" - hey, I need it as *admin*...) [20:07:10] I'd prefer to make this a test for Adrian - let's see how good he is at kicking people ;-) [20:07:36] I mean - getting the wiki database (including mail addresses) was quite easy [20:07:38] Anybody from SUSE has to use his Google account and the Google Developer tools. [20:07:46] so why is getting this API key so hard? [20:08:27] That's created with an openSUSE Google account? [20:08:28] how about to file a ticket and assign it to Max? [20:08:36] so that you have it written down as well for reference [20:09:28] indeed, a written reminder is always good [20:09:38] So next step: Creating a ticket for Max... [20:09:46] and set due date HA HA HA [20:10:10] no need to mention that ;-))) [20:10:11] After the Due Date an escalation via Adrian... [20:10:42] okay good, thanks for your work on implementing elasticsearch [20:10:56] am back [20:10:58] SSO is missing now. [20:11:11] How can we integrate it? [20:11:38] what is SSO? [20:11:39] Or do we need a live system for that? [20:11:46] single sign on [20:11:47] Single-Sign-On [20:11:55] first, I need to have the wiki VM behind an Access Manager proxy [20:12:21] We want to give access to all openSUSE users [20:12:28] writing the needed mediawiki extension shouldn't be too hard (at least I hope so) [20:13:51] @tampakrap: Do you know the authentification infrastructure? [20:14:35] not much, but maybe I could help [20:14:54] That would be great! :) [20:15:10] but you'll need to explain me in detail what you need, because what you said so far is only keywords for me [20:15:12] :) [20:15:26] so question from my side: we have ETA for the new wiki to get implemented? [20:15:32] or we have blockers? [20:15:40] and if we do have blockers, we have tickets for them? [20:15:45] actually I'd like to propose a deadline [20:16:09] Sarah and I want to have at least one production wiki running on the new infrastructure at the openSUSE conference [20:16:35] You know that all openSUSE users can assign into our tools via Novell/MF accounts. That has got the name SSO and is using an authentification server. [20:17:34] We want to implement that for our wiki, that everybody can open his openSUSE account for our wiki and can edit it then. [20:17:46] the Access Manager proxy basically means that we'll get HTTP headers which contain the username, mail address etc. [20:18:18] so you'll need to find out how to put the wiki test VM (and later the production wikis) behind the Access Manager proxy [20:18:19] We need the connection between our test wiki and the authentification server.... [20:18:19] okay can we talk about it at some other point and not during the meeting? [20:18:40] Ok. [20:18:43] sounds like a good idea [20:18:58] the best answer I can give you now is "I don't know sorry, need to investigate" :) [20:19:02] next topic: Salt / Gitlab [20:19:40] gitlab is fine [20:19:42] salt is not [20:19:56] I added some stuff, lars also sent a few merge requests [20:20:07] but I expect more from everyone! [20:20:15] get your hands dirty don't be afraid [20:20:28] What is missing? [20:20:28] I'm always willing to guide/help [20:20:42] I was too busy to push the wiki stuff, but I'll do it ;-) [20:21:00] (at the moment, I use a local salt master on the test wiki VM) [20:21:19] Oh. [20:21:21] Ada_Lovelace: I am working on common stuff that go to all machines (with very few conditionals) but role-specific code is missing [20:21:41] and I am not going to add in salt all of our services, because I don't know how all of our services are set up [20:21:57] I have created test machines in atreju btw, ask me for access [20:22:01] so that you can play around [20:22:12] in case you are lacking test environments [20:22:43] Yes. I thought in our wiki situation about role splitting, too. We want to have it all on different systems. That would be our next step... [20:22:47] so my next plans are the rsyslog configs, the postfix configs, user management etc [20:24:13] also, two people that know salt a lot agreed to join, I plan to talk to them about complex stuff [20:24:19] like testing and monitoring [20:24:26] I'll keep you posted [20:24:42] apart from that, not much is happening so nothing else to discuss unless there are questions [20:25:11] I don't have any questions. [20:25:18] will you also setup a MySQL VM, or should I create one that "only" handles the wiki databases? [20:25:34] you can set up mysql on those testing machines [20:25:41] (doing that is easy, but I won't complain if I don't have to handle backups etc. myself) [20:26:28] I know I can ;-) - the question is if we want a) one MySQL VM for each service or b) a "big" MySQL VM used by all services [20:27:11] my vote - one "big" one unless we have some really heavy uses. [20:27:13] At first for one service and at the end for many services... [20:27:22] so right now we have one big mysql [20:27:33] That's a test system at the moment. [20:27:41] and when we'll have more needs on one service we can discuss our options [20:27:48] same for postgresql [20:28:00] because they are both with HA [20:28:11] and they are PITA to maintain, so we avoid second instances [20:28:23] but we can discuss again if the needs grow [20:28:26] makes sense? [20:28:32] Yes. [20:29:13] Other questions? [20:29:42] Next topic: Mailing lists [20:30:12] pjessen: pick a day for maintenance and update to 42.2 [20:30:21] I'll need you present [20:30:22] @pjessen: You wrote us (Board Members) about your plans. [20:30:38] Next Saturday afternoon? [20:31:02] I can't next weekend, any weekday afternoon? [20:31:10] it won't take long hopefully [20:31:20] Uh, Wednesday would be good then [20:31:21] and if it does, we revert it to the working snapshot [20:32:04] @ada - yes, sometime last year I think I mentioned to Lars that I wanted to migrate away from mlmmj. [20:32:04] okay, before that is there any chance you can see the if the needed packages are in openSUSE:infrastructure or at the distro itself? [20:32:41] I'll do it as well, but I'd like to have another set of eyes [20:33:10] I'll double check tomorrow. I think 42.2 should be good though. [20:33:21] oh regarding mailman-test, I have created a machine for you to play, did I give you access? I don't remember sorry :( [20:34:11] no prob. I guess you need my key? [20:34:23] yes [20:34:35] I'll send it by email. [20:34:43] okay [20:35:04] wrt mailman - we have a choice, mm2 or mm3. [20:35:05] I'll send downtime announcement for wednesday, what time? [20:35:16] So we'll add the status of mailman to our next topic list? [20:35:26] let's say 1400 CET [20:35:36] perfect [20:36:01] mailman - sure, go ahead and add it, but I can't promise anything. [20:36:20] No problem. You are a volunteer. :) [20:36:39] That's the reson for asking you. [20:37:06] I'd love to get it over and done with asap, but it's a big move. [20:38:29] Migrations need some time... [20:39:28] Other mailing list issues or questions? [20:39:41] I'm done. [20:39:50] nope [20:39:57] Next topic: openSUSE Cloud in Provo [20:40:10] Is any Cloud Admin available? [20:41:11] Or Theo: Do you know something about the status? [20:41:46] no I didn't get any news for quite some time [20:42:07] best thing I can do is to ask tomorrow at work to give a status update at the ML [20:42:59] Our lst news were that the DNS would be the next issue in Provo and nothing with the Cloud. [20:43:09] Good idea. Do it! [20:43:43] Next topic: mirrors [20:43:56] I'm looking forward for a status update - the cloud setup is already taking toooooooo long ;-) [20:44:13] there seems to have been a flurry of mirror activity in the last week [20:44:32] We had a lot of issues with morrors in the last weeks [20:44:41] yes, and also martin wrote documentation in the progress wiki [20:44:54] rsync wasn't working right and so on... [20:44:56] new mirrors coming up, old ones disappearing, did somebody kick the hornet's nest? [20:45:30] btw, rsync.o.o is still missing an IPv6 address [20:45:55] widehat.opensuse.org has IPv6 address 2a01:138:a004:0:230:48ff:fe80:9a32 [20:46:01] probably Richard's mail on February 21st to the mirrors kicked several mirror admins ;-) [20:46:05] this is what host rsync.o.o gives me [20:46:06] Neither Gerhard nor Martin are available now... [20:46:23] yes, the DNS is fine, but the host doesn't respond [20:46:47] I think the IPv6 address keeps dorpping off whenever the server is rebooted. [20:47:06] file a ticket I'd say, assign it to gerhard [20:47:18] willdo [20:47:19] Martin asked for help, too... [20:47:36] Because the rsync script. [20:48:06] We are looking forward to seeing improvements in the mirror part... [20:48:31] on the bright side, a lot of tickets were closed due to people having time because of hackweek, martin wrote documentation, and the BW has been increased [20:48:50] we also already had some great improvement wrt ipv6 geolocation [20:49:24] it's still on my todo to move scanner to atreju [20:49:29] Yes. [20:49:39] so that non-suse people can finally do mirror tickets [20:49:58] I'd be happy to help out [20:50:03] currently we have 34 mirror tickets open - that's a big improvement, but it's still our biggest ticket category [20:50:47] Yes. And many mirror customers had problems in the last week. But it is better than before. ^^ [20:51:24] Any questions? [20:51:58] it would be great if the mirror list hd some more attention [20:52:19] for general mirror issues [20:52:26] Do you mean Adrian should help there? [20:53:07] I don't know who it should be, but the list is the main contact point for mirror operators. [20:53:17] and some users [20:53:41] mirror ticket lis or mirror mailing list? [20:53:50] okay I'll bring it up on my team meeting at work [20:53:52] mirror mailing list. [20:54:06] so the problem is that mails from mirror operators on the ML don't get answered? [20:54:45] no the problem is that there is no responsible person to reply asap [20:54:51] exactly [20:54:55] so they get delayed most of the times [20:55:34] sounds like the usual ENOTMYJOB ;-) [20:55:37] That's the job of the SUSE-IT. [20:56:10] So a good topic for the next team meeting. :-) [20:56:19] our mirrors are very important to us, we need someone to give them some TLC. [20:56:24] Or Adrian will be sent by us. ^^ [20:57:19] Adrian isn't organized only for Provo Admins. He should do that with SUSE Admins now, too. [20:57:34] yeah, hopefully Adrian will even take a video of his "visits" ;-) [20:58:26] Our SUSE Admins will get this topic in their next team meeting. If we don't have improvements, Adrian has got his first visit. ^^ [20:59:18] @pjessen: Thanks for the information! [21:00:00] Next/ Last topic: Naming pattern [21:01:00]