[00:05:42] https://phabricator.miraheze.org/T8772 [00:07:00] Odd [00:09:37] [1/2] see wikimedia's phabricator ticket: T315737 [00:09:38] [2/2] cleanupTitles.php seems to be possible fix [00:10:20] https://phabricator.wikimedia.org/T315737#8187139 apparently [00:12:09] [1/3] oh i think i know what happened [00:12:09] [2/3] i moved pages with a namespace in the title to the main namespace and it all got deleted [00:12:09] [3/3] broken delete [00:12:52] ty [00:56:51] maybe someone will run for global sysop on either sunday or monday (or never :p) [01:22:14] Ok [01:22:34] Why don’t we create a plan for the future of miraheze, to make sure this never happens again? [01:22:42] What is going ON? And why was I not INFORMEd [01:23:04] 2115 plan because miraheze will last 100 years [01:25:28] maybe 👀 but idk if i will [01:25:40] jk though in all seriousness you’d do well [01:26:44] side note: if my Meta sysop request only has 5 votes so far, how do I expect to run for GS haha (also it’s not the right time) [01:29:25] Nothing worth putting in announcements, same situation as yesterday, talks are ongoing for a possible succession plan instead of shutdown. [01:32:50] I already did (my nom was unsuccessful, but I ran again): https://meta.miraheze.org/wiki/Requests_for_global_permissions#Tali64%C2%B3's_Request_for_Global_Sysop [01:45:02] how's the situation going? [01:54:23] <.labster#0000> Same as earlier. It's a weekend, we're talking to financial professionals about saving Miraheze. I will write something later on tonight, but I'm taking care of family business for now. [01:55:30] <.labster#0000, replying to marrybore#0000> 'Why was I not informed' is actually the more interesting question, because it means communication broke down to our wikis. [01:55:30] miraheze is back up? [01:55:43] should we add a sitenotice? [01:55:46] it was never down. [01:56:11] (FWIW just to clarify - the financial professionals are there to make sure the new entity is set up properly, not because the project is in dire financial straits) [01:56:18] <.labster#0000, replying to originalauthority#0000> If you don't count the 503s that we all know and love. [01:56:50] Miraheze needs to audit its technical infrastructure because the reality is nobody should have to put up with so many 503s. [01:57:09] I believe the SRE team is aware they just likely need more servers ? [01:57:27] They need to audit the wikis on the platform and delete the ones that aren't actually adding any substanc. [01:57:48] Also they could probably follow Cook's advice and look into CloudFlare and get rid of Varnish. [01:57:50] There is most likely no relation between inactive wikis and the technical problems you guys have [01:58:04] no but it would free up space on the servers. [01:58:17] Storage space, comparably, is cheap to compute [01:58:18] Lack of storage space is not the cause of these performance issues [01:58:35] _looks at mysql ooming all the time_, oh right. [01:58:50] <.labster#0000> wikis not being used are not taking up memory [01:58:56] OOM would be RAM not storage [01:59:40] (Or, I'm more used to OOMs from RAM not a lack of storage space on a device. That's usually a different error message to my knoweldge) [01:59:57] either way, a technical overhaul is needed. [02:00:04] <.labster#0000> That much is true. [02:00:11] No disagreement there [02:00:12] [1/2] Perhaps I'm speaking out of turn, but I'd like to suggest that Brandon's question not be overlooked about a sitenotice. Be transparent and tell the users where things are. Yeah, things aren't sure, and you can couch it in that language - but in this Discord there is some discussion about a different potential outcome for the site than what has last been stated publicly. People do [02:00:12] [2/2] n't need certainity all the time - but informing them could save a lot of constant re-discussion. [02:00:47] Constant rediscussion is unforunately standard as people rotate in and out [02:01:09] I get that some certainty would be helpful, but I think the thing they're trying to avoid is communicating too quickly before there is a clear plan in place [02:01:38] We should at least have some notice, it seems like Miraheze will survive but the current board comms is that Miraheze will wind down on August 31. [02:01:56] If there is some possibility for MH to survive, we should alert everyone that isn’t on Discord and IRC of that. [02:03:26] There, unforunately, aren't enough details to currently add to that message that won't just open more questions than the statement likely answers :/ [02:03:37] I think the current comms is more "we are trying to find alternative arrangements, please hold" [02:03:39] At the very minimum we should so they can share their ideas here and join Discord or IRC if they wish. [02:04:03] Then let’s say that. It’s a stark contrast to “we’re shutting down August 31.” [02:04:09] There is very little we can do here at this exact moment. We are waiting on the discussions by Void and co before the community can really proceed [02:04:12] The latest tweets essentially do say that [02:04:30] I mean an on wiki notice [02:04:46] Language similar to the tweet could likely join the community noticeboard [02:04:56] I would support a Meta Wiki sitenotice that links to the Board's statement, sure, but at the same time, I do not feel it is appropriate to give people false hope, particularly when we don't know who the people involved in discussions to potentially take over Miraheze outside of Labster and Zppix. It's all very tentative, at best, and the prudent thing thing to do is encourage people to look at other wiki farm alternatives, or alternatives to [02:04:56] wikis, such as a static site generator :) [02:05:31] dmehus: just to clarify as well, not sure if you saw but my IBAN from ~3 years ago has been removed [02:05:37] I don't think encouraging people to look elsewhere right now is prudent [02:05:43] a while ago from what I can recall [02:06:13] There's a sentiment currently in this Discord that all is not necessarily dire, but the Wiki's last statement is that things are definitively shutting down. That's why I bring it up. [02:06:31] I'm not really familiar where all this has been communicated to the wider public besides Twitter and Discord - is there an on-wiki place where the current messaging just ends with "we are shutting down in September" or similar? [02:06:43] If that is the case, I agree that should be changed [02:09:57] I believe there was already a CN post by Labster detailing intent [02:10:17] on wiki [02:10:38] ? [02:13:11] nothing has been communicated to solely on wiki community members [02:14:08] So there's no existing communications that need to be corrected? That feels a bit less urgent to me then [02:15:09] Wasn't there a Board statement from 06/15/23? [02:15:24] That board statement isnt 100% accurate anymore [02:16:50] but that wasn’t communicated is my point [02:16:52] If there's a current revision of a page that says "we are shutting down" then that should probably be modified [02:17:01] there is [02:18:52] there is no impact of inactive wikis, this is insignificant [02:19:04] inactive wikis tend to have the least files, the least pages, and the least wikis [02:19:26] they are in fact tied for the least wikis with all the other wikis <:Sunglaso:603351762614616084> [02:19:56] (but yeah, general point is true that pruning inactive wikis doesn't do anything meaningful for performance issues - if anything it might slightly make them worse) [02:20:05] already had edited but "nice try" ; P [02:21:21] if it were for me I would delete "drawnfeet" but I admit I am biased due to personal distaste of the meaningfulness of that "project" (which uses many thousands of images) [02:22:12] However, mainly I agree with Brandon that the Board statement of the "inevitable" shutdown timeline needs to be amended or deleted [02:23:05] Plus it contains some very odd sentences such as "Its not in the best interests of the project to continue" (what does that even mean? Its certainly in the best interests of the community) [02:25:30] [1/2] The community != the project. [02:25:30] [2/2] The latter is the legal organization that manages the services. [02:25:54] You dont have a project without the community imo [02:26:13] And what is supposed to be the interest of the nonprofit miraheze project if it is not to serve the community? [02:27:10] I'm of the opinion its a poor position for the project to agree to support something the volunteers running the project feel they are unable to functionally run. [02:27:56] This is also true. I'm not trying to argue otherwise per say [02:29:47] they just mass resigned due to internal conflicts and asked nobody to fill the gaps [02:30:06] "i don't want it so let it burn" [02:30:58] I'm going to avoid rehashing this particular topic past "the communication could have been improved. I disagree that the statement you originally pointed at is non-informative as to why the decision announced was made" [02:32:14] An actual informative statement shoudn't have said "lack of volunteers" but rather "everyone suddenly left over some internet drama that had already been over" [02:32:31] A lack of volunteers is also true [02:33:04] The volunteer staff was already very low for the size of infrastructure that is being supported [02:33:06] Several people volunteered as soon as the problem was made public [02:33:38] Both things are true about the situation -- we have been deathly low on technical volunteers for ages, the call was put out on meta and during every miraheze meeting but the few volunteers we did get didn't stick around. [02:34:06] Yeah I'm unsure how to continue this conversation Nemo given both our current statements are "true". [02:34:15] I agree the statement could have been a lot better, and even now it's probably a good idea to make sure the outward messaging is not definitively saying things will shut down. It's probably not time yet to announce an explicit commitment to continuation, since parts of it are still up in the air [02:34:23] Recently the last few active members resigned for multiple reasons, not just internal disagreements, and the board felt that the chances of getting adequate volunteers to continue the project wasn't a reasonable possibility. [02:35:05] They made the choice to instead wind down operations on a timeline vs risking sudden collapse. However, community support for an alternative plan has recently come to be, and discussions are ongoing between that successor team and the board. [02:35:29] There will be plenty of time to look back at what went wrong, but for now I think the priority should be to install steady leadership, and work on getting the fragmented and burnt-out tech volunteers back and contributing towards a common goal [02:36:28] Right now, it's very easy for a reasonable discussion of what went wrong (and wow, there's a lot of blame to go around) to just turn into a shitshow of indirectly bashing the people that you guys still probably need to keep the site running [02:36:37] The action item listed here is in progress with Void, Zppix, and others in an active and evolving discussion. They believe they have a reasonably high chance of success but its uncertain [02:36:49] I think most of the reason why there's a lot of rehash on the recent events is because it brings out passion, emotion, and feeling. Those in this Discord are here because they have feelings about MH. But I think it also perhaps comes up so much because it feels like a puzzle that some are trying to piece together with not the most information. [02:37:23] One of Labster's comments above was a general question thinking about current communication gaps. That thought is worth considering as the project moves forwards - it is a great place for improvement. [02:37:43] Yeah, I'm involved in those discussions [02:38:04] I've not seen an entire list of who's involved sorry [02:38:26] I suspect things will work out. But my recommendation is to still move slowly and judiciously [03:15:41] Did owen leave? [03:31:54] <.labster#0000> I'm getting close to releasing a statement. If you've been following Discord closely a lot of it won't be news to you, but I hope it can calm the fears of some of our community. [03:38:14] Anyone know how to import images properly if going from Miraheze to Self host? I got an image backup and I just need to import it, but I don't know the best way to do that (besides manually reuploading everything one by one) [03:41:00] Been a minute since I’ve touched MediaWiki, but I think theres a importFiles or importImages maint script [03:41:57] Labster, may I DM? [03:52:59] <.labster#0000, replying to Cobalt#3065> Of course you may. [03:53:26] see DMs as well incidentaly [03:55:41] https://www.mediawiki.org/wiki/Manual:ImportImages.php [04:01:24] [1/3] That helped a lot, though i get this error when I run that [04:01:25] [2/3] "Wikimedia\Rdbms\DBConnectionError from line 1477 of /path to wiki/u1gawiki/includes/libs/rdbms/loadbalancer/LoadBalancer.php: Cannot access the database: No database connection (unknown)" [04:01:25] [3/3] I also get that when running the update script so I got zero idea what to do there [04:20:44] you're database is not configured corectly in LocalSettings.php [04:21:40] How do I do that? Checking documentation and tried what seemed to make sense but got nowhere, but I also don't know how to really check that [04:22:41] I'd start by running `sudo service mysql status` and see if it shows mysql is running (if you're using mysql) [04:23:57] "sudo: service: command not found" [04:44:14] You mean the 503s that we all know and hate. [05:07:26] [1/4] rather odd asking about a non miraheze wiki in here, but check following: [05:07:26] [2/4] make sure your database user's password is correct [05:07:26] [3/4] make sure your database user's password doesn't contain a dash or special characters [05:07:27] [4/4] make sure you gave correct database permissions to database user [05:07:57] try sudo systemctl mysql status [05:08:30] or replace mysql with mariadb if you installed latter [05:08:42] ensure you actually installed mysql or mariadb [05:23:15] [1/5] And I'm guessing those values would be under $wgDBserver [05:23:15] [2/5] $wgDBname [05:23:15] [3/5] $wgDBuser [05:23:15] [4/5] and $wgDBpassword? [05:23:16] [5/5] Cause I set those and it either can't find anything or gives an access denied error [05:30:15] let's get this moved to DM? [05:30:20] Sure [05:30:54] I can't message you... [05:47:08] Support thread would also work [05:48:47] i'm a bit limited as unverified user [07:59:05] This is nothing to do with the Miraheze situation, but is relevant to CreateWiki. If a wiki is marked as deleted, it is moved to the deleted.json cache, right? and removed from databases.json? Is it then meant to be inaccessable? [08:00:05] <439bananas#0000> oh dear [08:00:17] <439bananas#0000> you had to put up with naleksuh? [08:00:33] bit late [08:01:02] <439bananas#0000> yeah i didn't hear anything about him until someone mentioned it in a private server [08:02:54] https://tenor.com/view/evil-defeated-lost-the-evil-is-defeated-kid-gif-5007371 [08:03:06] sorry [08:03:56] <439bananas#0000> we also had problems with him and honestly we got really fed up [08:04:48] Yes that should be the case [08:08:11] [1/3] Interesting, for some reason, this doesn't happen on Telepedia. The database json files are changed, and the wiki is marked "wiki_deleted = 1" in the database, but the wiki is still accessible. I assume that functionality is built into CW, not MH Magic, as I know there's some delete code in MH Magic (https://github.com/miraheze/MirahezeMa [08:08:12] [2/3] gic/blob/master/includes/MirahezeMagicHooks.php#L43-L66) [08:08:12] [3/3] And I note that this has happened on Miraheze before, https://phabricator.miraheze.org/T5961, but the commit that fixed it already seems to be present in the master branch (and has been changed since that ticket) [08:08:23] Not sure if its an iscolated incident on my end. [08:24:26] <.labster#0000> https://meta.miraheze.org/wiki/Miraheze_is_Not_Shutting_Down [08:25:38] <.labster#0000> There are more details to come later, but here's what I'm able to share now. Feel free to ask questions. [08:26:03] <.labster#0000> Also @Juwi change your flair already, we're not dead yet! [08:26:16] (Hopefully never) [08:26:37] <.labster#0000> It's merely a flesh wound. [08:27:13] I’m still moving 3 of my Wiki to a new host [08:27:28] I’m still moving 3 of my Wikis to a new host [08:27:52] <.labster#0000, replying to Loves Next & Misses Sandra#2978> I understand. We all have to assess risk for our own wikis. [08:28:55] <439bananas#0000> all naleksuh's fault? [08:29:12] I moved my iNext Wiki to a new host [08:29:40] No, it isn’t [08:29:50] <439bananas#0000> huh that's what i gathered [08:30:09] Theres alot of reasons we have gotten to this point [08:30:36] <439bananas#0000> i mean i saw that at least he's part of he reason [08:31:52] <.labster#0000> Now is not a time to assign blame for anything. We need to move forward. [08:32:03] ^ [08:32:32] <.labster#0000> As I said in my post, there will be plenty of time to assess what went wrong, but this is not it. [08:35:29] well said [08:40:11] well I'm no one special but I feel better. [08:40:30] I think Miraheze will live [08:45:09] Hey @.labster just saw your post. How can I help? [08:47:35] <.labster#0000, replying to Ruchir#8499> [1/2] It depends on what you have experience in. Keep editing on Miraheze is #1 to me. Make sure people are aren't worried about the future here. [08:47:35] <.labster#0000, replying to Ruchir#8499> [2/2] Beyond that it depends on what your skills are. [08:51:12] Ok. Will have a think. Do you have a posting listing any key skills gaps you need? [08:51:30] <.labster#0000> https://meta.miraheze.org/wiki/Miraheze_Volunteering_Opportunities [08:51:58] [1/2] Yes. I agree with that interpretation of the message. [08:51:58] [2/2] That's why people panicked at first. But now we have more options. [09:00:31] <:eyes_zoom:606564192035667983> really hoping this works out!!! [09:01:16] Still gonna make sure all my shit is backed up in case of emergency, but even just that little bit of hope is a huge relief. Thank you for trying to keep this place alive [09:03:26] <.labster#0000> As a software security person, please keep important data (like wikis) in 3 different copies, at least one of them in a different location. [09:42:35] Yeah that's what I've done [09:43:32] I've got like four backups; one in Google drive, 2 in two different private discord servers (one of which the other wiki mods have access to) and one on my computer itself [09:50:53] OH I've also got my shit backed up on Internet Archive [09:58:39] I am admittedly a little tech illiterate, but preservation of data is something I'll often go above and beyond with, just bc of my shitty luck lmao [10:03:11] Got your wiki backed up on Internet Archive is the most safe option [10:03:26] But that's cold storage, not something you can get in and out easily [10:03:52] i've also got snapshots of the wiki on the Archive [10:04:21] yeah,,, that's why i've also got it saved in more convenient places dsdfdsdf [10:11:05] LET’S GO [10:19:39] Eyyy I'm happy to hear there is hope!! [10:33:40] <光纪#9056> hooray! [10:37:41] And back-ups of those back-ups, right? [10:51:33] <光纪#9056> So would Miraheze taken over by another entity? [10:51:54] <光纪#9056> What about the web server and database [10:54:35] [1/2] those will stay [10:54:35] [2/2] we don't know how the legal process will be, but essencially things won't change? [10:57:57] Fingers crossed!! [10:58:34] Miraheze is the most convenient place for us rn, especially since WikiTide hasn't decided where they draw the line on NSFW and they're so new that we have no clue how viable they are [10:59:26] <光纪#9056> tbh I don't know there is nsfw wiki on MH before😂 [10:59:45] <光纪#9056> tks [11:00:51] SDFDSDFD oh there definitely are. They can vary in severity from the occasional reference to NSFW topics all the way to blatantly adult content [11:02:24] as long as they follow content policy and respect other users (as in use spoilers etc) [11:03:11] MHM OF COURSE [11:04:09] Our wiki does it by (a) being private, (b) the main page being PG13 at worst and (c) having a banner explicitly stating that it's 18+ only and that minors aren't allowed [11:05:04] like, one of the reasons we picked Miraheze is bc it allows private wikis. none of us wanna accidentally expose unwilling people to nsfw or horror content [11:08:32] Noob question regarding Image Dump: when it's done, what exactly do you get? Just a folder with all the pics uploaded to your wiki, or some kind of script to facilitate importing them again, like the ordinary wiki dump's .xml? [11:09:49] I'm asking because I got all the pics on my PC so I wouldn't need to download them again, but it could be interesting to have some script to make it easier uploading them again to another wiki, in case the worst happens [11:10:05] i believe it's the latter. at least, my image dump didnt have the pics [11:10:31] i've saved the image dump and the individual images, better safe than sorry [11:10:59] thanks! [11:11:22] np! [11:14:15] image dump is a bit wonly rn [11:14:45] I suggest using third party scripts - WikiTeam Tool (python) ot FANDOM DownloadImage (node.js) [11:17:34] is that bc of how many people are requesting them? or issues due to old volunteers leaving? [11:17:39] I mean it have to in some form have them, how else would you import them, to a completely unrelated wiki service, specially in the case that MH would have went down (which ofc by now is basically a hypothetical case, but still), but I wouldn't be surprised if it is in a form that is not viewable with normal image software [11:17:57] glad i did mine earlier though lmao [11:18:36] yeah i think that's the case, the images are technically there, we just can't really access them [11:20:51] [1/2] the dump queue is limited, as I understand [11:20:51] [2/2] couple of days ago it was real bad, a wiki w/ 40k images stopped the rest wikis in queue even for xml dumps [11:21:29] <光纪#9056> ? oh I don't even know that [11:21:41] <光纪#9056> quite terrible [11:21:42] A H [11:22:04] The WikiTeam dumpgenerator python script is out of date, (requires Python 2.7). Mediawiki Scraper is the same script rewritten for the current version of Python https://github.com/mediawiki-client-tools/mediawiki-scraper [11:22:17] Macfan had to stop it manually, I think [11:22:33] yup, I read that [11:22:52] I thought both were created by WikiTeam, will edit my pin [11:23:28] The WikiTeam dumpgenerator python script is out of date, (requires Python 2.7). Mediawiki Scraper is the same script rewritten for the current version of Python https://github.com/mediawiki-client-tools/mediawiki-scraper [11:24:48] node.js one is simpler tho [11:25:13] for images [11:25:38] Mediawiki Scraper will dump XML and images/files, etc. from Fandom [11:26:51] I specified that FANDOM script works w/ all images [11:30:42] MediaWiki Scraper is a fork from WikiTeam, not the same people ... [11:31:03] yeah, I figured now ... [11:35:06] I am having an issue with Infoboxes on my wiki [11:36:16] This should probably be added to #announcements [11:37:10] probably missing CSS, took this infobox from Wikipedia? [11:37:36] I suggest to create a thread in #support tho [11:46:19] I've done so, thanks!! [12:06:17] what's up with this [12:09:34] @antoniokf5 which wiki [12:09:47] This is the service notice board [12:10:10] @zppix can you grant confirmed please? [12:12:01] Is the captcha broken? [12:12:17] Yes [12:14:39] Ill grant it to expire in 3 weeks, should likely gain auto confirmed by then [12:15:21] Should be fine now [12:16:13] thanks [12:18:01] Thank you! [12:41:55] [1/3] Agree with you Rakkeyal. [12:41:55] [2/3] The Miraheze is not shutting down post should be shared on announcements too. [12:41:56] [3/3] Some discord server receive relayed posts on their servers. Like a webhook and maybe don't read wiki or this general channel. [12:55:39] in bot channel plz @!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [12:57:04] Sorry 😦 [12:58:06] Dumb idea: kitchen gun plan (this is a joke) [12:58:26] Delete the edit history of all wikis unless they opt out (again, a joke) [14:04:37] New status just for @.labster [14:17:20] [1/2] This is wonderful to hear and I wish the new team the very best 🙂 [14:17:20] [2/2] While my hope is those who left may at some point be welcomed back, it will be difficult consideration for _some_. Much damage was inflicted by those leaving, that saddened me greatly [14:17:27] will the wikitide fork be resolved? [14:18:38] you should ask that at Wikiforge [14:18:56] *wikitide [14:34:52] [1/2] it is probably better to first focus on getting Miraheze fully stabilized and a full postmortem, ensuring new policies and procedures are in place to avoid a repeat of this [14:34:52] [2/2] personal opinion: I am not very keen on inviting back in the resigned folks who have inflicted so much damage on their way out. They walked away, probably is not the time to be inviting back so soon...🪙 my opinion probably won't mean much. [14:38:32] Agreeds, and while for those reasons I don't really trust WT, it might in the long run be beneficial for both them and us having that actual direct competitor, giving an incentive for both to strive to provide the best possible service. Yes technically Fandom is a competitor as well, but is in a whole other league, that we can't feasibly be competing with them [14:38:54] are they really competitors tho? [14:39:13] yeah, Fandom is a massive corporation that's swallowed up other companies. a volunteer-run non-profit could never compete [14:39:41] It's not like MH gets any extra money for having more wikis or users, so the financial incentive isn't there for us, which doesn't may they may not be other inventives [14:40:06] *doesn't mean [14:40:12] it's like comparing McDonalds to the little food joint run by locals, one of them has a lot more resources than the other [14:41:21] This. [14:42:33] Apologies yesterday paladox for my tone, was not my intention to owen or yourself for choosing to walk away [14:42:57] 👍 [14:46:30] [1/5] 2 things I would note to do in postmortem [14:46:30] [2/5] 1. policy for enforced breaks and encourage informing of overwhelming stress to then step away [14:46:30] [3/5] 2. tighten wiki creation acceptance and proactively clear out any obviously dead or abandoned wikis as a 1 time thing [14:46:31] [4/5] a concept I learned from FrozenPlum is a "community of practice". I think that would be good to visit for Miraheze to implement. [14:46:31] [5/5] I had strongly advised CA to walk away, was ignored and disregarded leading to deteriorating relations [14:47:24] what do you mean with "community of practice"? [14:50:10] [1/3] Your status says: [14:50:10] [2/3] > Long Live Miraheze [14:50:10] [3/3] That reminds me Taylor Swift song named "Long Live" [14:53:25] www [dot] wenger-trayner [dot] com/introduction-to-communities-of-practice/ [14:56:46] [1/2] [14:56:46] [2/2] > italic: Communities of practice are groups of people who share a concern or a passion for something they do and learn how to do it better as they interact regularly. [16:16:14] An "enforced break policy" seems like a really bad idea to me, it would just escalate things [16:18:00] Enforced breaks do not help volunteer welfare. Telling people how to sort their welfare out is a shit idea. You kinda need to listen and understand them. [16:25:48] [1/3] In general, I like both of those ideas, with #2 being as essential. I would also add a third, as also being essential, and that is likely a rebrand. The Miraheze brand is badly tarnished. That's not to say Fandom is not tarnished, for different reasons, or that we can't operate with a tarnished brand; we absolutely can, but will be severely limiting our pote [16:25:48] [2/3] ntial audience/customer base to certain types of wikis. [16:25:48] [3/3] #1 is an interesting idea, and it's one I have thought about and agree with in general. Does that mean resigning? I don't think so, necessarily, Companies have mandatory minimum vacation leaves, and this is no different. That does not mean mandating a break without social support, though. The community, and leader, support needs to be there [16:27:46] No it's not, mandating breaks at a specific time is not a holiday [16:28:06] Frankly, forced breaks is a bad idea [16:28:13] Saying you have 25 days a year to take when it suits is very different [16:35:22] @cookmeplox@rhinosf1 that was meant in regards to being informed or realizing a team mate is overly stressed then making them take a break to avoid a public emotional meltdown. I am more geared towards temporarily revoking access for a set period. Essentially, go get rest. Go do something fun. Recharge. [16:35:35] Is Miraheze still shutting down? Why was no official announcement made in #announcements [16:36:15] I trust people to know to step back of communication is proper [16:36:16] That's a decent idea. We've never had readdition clauses, and this would be a good use case for one [16:36:31] Oh you guys should definitely post the same to #announcements [16:36:45] @danielmanfox see https://meta.miraheze.org/wiki/Miraheze_is_Not_Shutting_Down [16:37:01] I understand the idea behind it, but forced breaks I could only see the potiental of causing further issue [16:37:36] Not at a specific time, no. The volunteer should be free to take their break whenever, so long as, say, they take at least a five-day, ten-day, whatever it is, break at least once per year [16:37:48] Yeah in practice that sort of proposal tends to escalate things when people are already stressed. Trust me, not a good idea [16:38:13] yeah, well in this case. communication failed...I feel as if my efforts to stop CA in private, contributed to the meltdown [16:38:51] in terms of confidentiality, I withheld what I should have told MH staff a week ago... [16:39:15] I also think it's unfair to be blaming CA. There were a confluence of issues, some of it was volunteer burnout, sure, but in large part, Naleksuh was the reason behind Miraheze's collapse [16:40:20] not blaming CA, pointing out that if was on break the past week; maybe this could have been avoided to the extent that happened. perhaps a calmer discussion... [16:41:24] Absolutely. I didn't mean to suggest you were blaming CA; I know he can be prone to brash decisions, and sometimes private discussions help and sometimes they don't. I just meant in general, I've seen a lot of people blame CA in recent days and just happened to be a reply to your comment [16:41:41] We need stable leadership [16:42:10] No point in trying to convince them to stay [16:42:26] We had stable leadership, actually. The issue was volunteer burnout, a shortage of qualified and trusted volunteers, and active leadership [16:43:17] I have been very unhappy with CA this entire week. he seemed to have slowly started to understand... [16:44:20] ok, good to know. wasn't aware it would cause more harm [16:45:36] I think Miraheze will rebuild [16:46:41] I think you guys need to take a hard look at why the volunteers burnt out. It wasn't just due to them suddenly deciding they didn't want to do MediaWiki stuff anymore. The more relevant issue is that miraheze, institutionally, has been really bad at resolving conflicts without things escalating to the point that they threaten the entire project [16:47:13] To the point that people who weren't even involved with any of the drama were feeling like, wow, this is really just stressful and not fun and not what I signed up for [16:47:20] I think there's a way to implement both, mandatory minimum annual breaks and social supports from volunteer leaders and the community. Communication is key [16:47:54] These people are the ones that matter the most - drama is a mess, but the non-involved users are the ones that matter the most. The ones that just run their wikis and don't even get involved in Meta. [16:48:30] Agreed [16:49:20] Vokunteers means they donate their time and energy, unlike employees which are paid and forced to work [16:49:23] We need a new team in order to continue running the Wikis, plenty of people have wikis and they don’t deserve to have their hard work and valuable time go to waste. [16:50:05] And it's easy to point blame at individual people or situations but I think that's missing the point. If everyone individually is acting irrationally (which is kinda what happened),your problem is that your org needs leaders, not that everyone individually is broken [16:50:39] We don’t need blame, we need reconciliation and we need everyone to work together [16:50:53] Yes exactly, there's no inmediate shutdown risk but we still need a new team to run the service. [16:51:02] We absolutely need to work together, I couldn’t agree more. We need a new team to operate and administrate. [16:51:06] [1/2] Conflict resolution was a big part of it; there were many internal disputes within the SRE team over the years, and a lack of leadership in resolving them, but it was more than that. It was a shortage of qualified and trusted volunteers, sometimes our standards for new SRE team members was too high, and the workload was not distributed equitably amo [16:51:06] [2/2] ng the SRE team members. That is to say, of the few SRE volunteers we had, some were not seen as 'pulling their weight', so to speak, leading to some having to do most of the work, which exacerbates the volunteer burnout [16:51:34] Well; a shutdown risk could happen if this is left unchecked. We need someone new to takeover and have the torch passed onto them. [16:52:04] that's Labster and MH veterans atm [16:52:56] I agree with this - I think the vast majority of people want a safe space to run their wiki, and don't care about all the other stuff, and the initial statement from June 15th put them all in a really tough spot. I hope the follow up statement from Labster shows that they're focusing on the right things now [16:53:09] hello [16:53:14] WikiTide is also a good option, and I'm told it is nearing break-even, thanks both to subscriber donations and its premium WikiForge subscription model [16:53:22] exactly [16:53:35] Me and @Tizian run a Wiki on MiraHaze and we can both confirm it’s an amazing platform which does not deserve to be in such turmoil. [16:53:38] [1/2] I personally don't trust WT [16:53:39] [2/2] yet [16:53:54] I can confirm this [16:53:55] Exactly. Drama matters a LOT less than the people that just want to run their own damm wiki, which is what this project is for. [16:54:00] It seems unlikely to me that WikiForge has a viable long-term business model [16:54:03] There's a problem with that: ads [16:54:16] The same thing why everyone hated Fandom [16:54:22] It doesn't have ads [16:54:24] [1/2] Until things are finalized, I'm not going to be making an announcement unilaterally, but I have pinned the post in question. [16:54:24] [2/2] Current board members are free to convert to a full-blown announcement, but I think we've all had more than enough of announcements with not quite enough information. [16:54:35] *not yet [16:54:40] Enforced breaks are something I would completely oppose [16:54:53] Why? [16:54:55] I think ehat should be done is mandatory rotation [16:55:14] Is this enforced breaks for all staff or for the Board? [16:55:20] a minimum amount of ads should still be on the table just in case donations can't cover or is not enough, something to think about [16:55:28] but honestly, can we take other platforms talks to #Alternative Wiki Hosting Services thread? [16:55:45] see as cook is advising against, I am going to retract that idea [16:55:48] In my view, it should apply to technical, community, and legal/compliance roles [16:55:51] You can't tell people how to manage their life. Enforced breaks don't help welfare because in a lot of cases it can make the person that breaks are being forced feel like they can't help when they want. [16:56:02] Yeah that's reasonable. I was under the impression that the old board endorsed the new statement, but you get bonus points for moving slowly and thinking before you act, haha [16:56:06] Forced breaks only applies to JP or some weird countries [16:56:06] Ads will not gain us more money than donations (lots of people use adblockers) [16:56:08] People should be able to manage their own time, we shouldn't manage others' time for them. [16:56:18] Because I, the other day, suggested we have set terms on stuff like the Community seats of the Board [16:56:47] I have retracted the idea per cookmeplox's advice [16:57:12] I think having the various roles be under cyclical contracts (so you serve for, let's say, 3 months at a time, and renew said contract at the end of that period) would be beneficial [16:57:13] I guarantee this will not work the way you think it will. [16:57:25] Term lengths are different than forced breaks [16:57:28] As an alternative, it may be helpful to have an 'on schedule/off schedule concept', right now it's all volunteers on until they get too tired to help [16:57:30] I think you guys need significantly fewer rules, not more [16:57:35] Also to clarify on this point: The Content Policy bans ads on Miraheze. So until that's changed, this is a no. [16:57:49] I mean, a contracted set of time seems pretty simple no? [16:58:00] I'm not sure even that is a good idea [16:58:12] You sign on to work at some role as a volunteer within a set amount of time which you renew every so often [16:58:17] [1/2] I agree, ads are not the future and only annoy those that don't use adblockers, but I also think premium plans and premium services are helpful. I suggested Miraheze consider becoming a domain name reseller, and manage the registrations of domains for users with custom domain requests, for no added fee. They would just earn a commission from the registrar with which they've partn [16:58:17] [2/2] ered [16:58:28] The more rules you have, the less you are signalling to your important volunteers that you trust them to handle their shit [16:58:31] This seems fine. But a schedule of enforced breaks doesn't work. There can be days for volunteers to take time off, but scheduling when they take those breaks or listing a minimum number of days to use those breaks is not good. [16:58:41] It's not a forced break thing tho [16:58:41] Instead of flat out forcing breaks, maybe set up a calendar or something similar, so it at least gets easy to tell the other people of the same role, if someone is gonna be away for a specific time period, so that people can at least take a break when they need, and still rest easy, that noone else get surprised you are not there for a period all of a sudden [16:58:47] Personally I think that the administrative team should choose a new team to continue running the site, someone competent should be leading. [16:58:47] And your most important goal should be to make those volunteers feel valued [16:58:54] <439bananas#0000, replying to dmehus#0000> yeah i was surprised he got involved in yet another wiki [16:58:55] Agreed [16:58:56] exactly [16:59:13] Should not blame CA at all, agree here. [16:59:37] there's one already in making [16:59:43] But Naleksuh also wasn't really a big part of the problem. He jumpstarted a lot of this, but in reality was a symptom of what Miraheze's community became. [16:59:55] Can we please stop trying to assign blame. [16:59:57] [1/2] Depend on what kind of volunteer [16:59:58] [2/2] Mine as a volunteer would work extremely hard but also take really long ass breaks [16:59:58] what's up w/ you three pushing same sentences? [17:00:02] ^ Name drops still not useful [17:00:07] ok, let's drop the "enforced break" idea. already retracted and isn't a good idea [17:00:13] @cookmeplox I'm basically suggesting how a normal contract works, but it's just a guidance for how long someone can volunteer for as a basis. So let's say I wanna be a volunteer at Miraheze so I agree to be a volunteer for 3 months, at the end of those three months I can either renew that agreement or choose to leave [17:00:20] I don't know, the goal should be helping those who create their own wikis and just want to give a "house" to their community - Meta doesn't mean as much. The most important people are the ones that don't get involved in the drama. [17:00:34] I believe WMF has something similar [17:00:40] agreed [17:00:47] guys [17:00:48] but they have enough volunteers to basically sustain them for a long long time [17:00:59] <439bananas#0000, replying to zppix#0000> i'm gonna be honest, i wouldn't find it unfeasible that he played a role in its downfall; he's been a nuisance on other wikis as well [17:01:00] we don't, so if too many decide not to renew their contracts.... [17:01:04] What purpose does that serve? In actual contracts, people are paid and have an obligation to work on things [17:01:11] To prevent burnout [17:01:19] Assigning blame is still not useful for actually moving forward [17:01:20] Yes, we need to fix problems not create even more drama. [17:01:22] Burnout will probably occur one way or the other [17:01:23] People work at their discretion essentially [17:01:25] exactly [17:01:27] we are past behind talking about him, everyone here are really tired [17:01:30] Blame should not be assigned to anyone [17:01:34] How does that prevent burnout? Let people work at their discretion regardless [17:01:35] Not if you choose how long you are a volunteer for? [17:01:38] We need to put an end to the drama [17:01:48] drama is over [17:01:50] exactly [17:01:56] Ok this slow mode doesn't help tbh [17:01:57] GUYS [17:02:07] I think if anything it actually does the opposite of what you want - it creates an obligation to stick around for those three months [17:02:08] The most important thing for preventing burnout is simple, have more volunteers. If work is spread out, there will be less stress on individual volunteers. [17:02:10] Can we migrate back to #tech-community or something [17:02:12] you don't listen to us? [17:02:20] But circumstances in life change. The timeframe for burnout could shrink exponentially given some particular circumstances. [17:02:26] Couldnt be said better [17:02:43] ... [17:02:49] How can we find more volunteers then? [17:02:51] https://meta.miraheze.org/wiki/Talk:Miraheze_Volunteering_Opportunities [17:03:18] Alright then that's not viable as an idea [17:03:44] So then how do you prevent the team from just getting collectively burnt out because they have some sense of obligation [17:03:49] There are a lot of people who want to volunteer. We simply need a less bureaucratic way of applying. [17:03:57] indeed [17:04:07] Ideally you'd have more volunteers, but there needs to be a vetting process for volunteering [17:04:28] It's not an enforceable solution imo [17:04:35] Yeah, I mean, that's the ultimate goal. But doing that with virtually no money and no paid staff means you need to be creative about making the people operating the site feel valued [17:04:45] It has to be a human solution. Everyone has to decide for themselves. [17:04:56] And I think having more people could also help [17:04:59] You can't force non-burnout on people. [17:05:04] That vetting process doesn't necessarily have to be a massive bureaucratic mess. A google form, and potentially a voice interview for important roles, is enough. [17:05:10] More people spreads out the work more meaning people don't feel overworked [17:05:14] Sure but there's a limit to how much good that does. [17:05:15] exactly [17:05:17] This [17:05:28] right on [17:05:36] Vetting is an interesting tool because how far does that go? [17:05:37] While true, we're well below that level of staffing [17:05:46] I'm sorry, but can you stop w/ this please? chat is fast enough already, then you add this single words messages [17:05:50] Like the past few months have like, been one or two guys managing systems meant to be run by like a dozen [17:05:51] Very true. We need at least double. [17:06:03] Im merely agreeing [17:06:15] https://tenor.com/view/destroyed-gif-20490322 [17:06:19] Everything but some board appointments and SRE is done by the community through elections [17:06:19] emoji reactions? [17:06:20] You can use a reaction 👍 if that's all you wish to communicate in that case [17:06:29] This is pretty much mh in 2023 [17:06:53] And I think there should be like, limits for those, so you have elections every few months, but I'm jumping the gun ig [17:07:12] ehhhhhh that only works so much [17:07:15] You guys need to fight your natural inclination to add bureaucracy and process [17:07:22] I've said before that the community board seats shouldn't also be directly community elected, but selected by delegates of the community [17:07:24] Reconstruct MH in 2023... [17:07:47] Maybe instead of elections we could just have applications? We need less bureaucracy and less processing, but rather more simplicity and making it easier to become volunteers. [17:07:47] Some level of bureaucracy is needed in a company and organization sadly [17:07:53] okay also what's been discussed so far [17:08:00] Yes, but not of the type you're talking about [17:08:07] https://tenor.com/view/collapse-reverse-building-demolition-gif-23091052 [17:08:07] some good ideas have come out of this and i'd love to be able to write them down and save for later [17:08:18] The easier and simpler it is the more volunteers [17:08:20] in terms of tech application, you just need to find folks with the relevant tech skills [17:08:46] [1/2] A polite note to all involved -- I do not accept unsolicited Discord friend requests with no context or message. [17:08:46] [2/2] I appreciate the enthusiasm, but my Discord experience is hectic enough already. 🙂 [17:08:54] imo; there's nothing in this conversation I haven't seen discussed elsewhere previously. 🙂 [17:09:07] I think ppl forgot to apply Sysop role which is the easiest to apply in MH [17:09:11] Like, I run a far larger wiki organization than miraheze and we have far fewer hats and applications and elections, and it works because what that forces us to do is trust each other and be flexible [17:09:17] How about forming a miraheze political party [17:09:21] No [17:09:30] If it wins the elections we can use government funding (this is a joke) [17:09:39] Yes exactly, no need for a messy bureaucratic process. [17:09:55] The Board's community seats, however many there will be, should be decided by the community members in some form [17:10:01] And it also helps that we have a functioning community-elected board, which miraheze has never had - in practice this is the source of a lot of problems [17:10:20] exactly, It only causes drama and problems, everything would be much simpler [17:10:23] Whether it be through Delegation or direct election [17:10:32] Emergency powers? [17:11:24] I think there should be terms for said elected members of the board as well, if/when the board is stable in its composition and community engagement is active enough [17:11:38] What we do, and this works very well in practice, is the community elects a board of trustees, and our main executive person (in our case me, in yours probably Labster for now) is accountable to the board [17:12:02] And then that main executive person has wide latitude to build shit [17:12:09] [1/5] So if we're reducing bureaucracy, here are some ideas, let me know what you think: [17:12:09] [2/5] - Get rid of GIWA, assign IW rights to local bureaucrats [17:12:09] [3/5] - Remove Meta admins, make CVT sysops, but allow for Meta bureaucrat requests [17:12:10] [4/5] - Straighten up Content Policy by 20x, get rid of all NSFW wikis unless they're documenting something like LGBTQ+ history [17:12:10] [5/5] - Make board anonymous, follow Zppix's plan for board/SRE as well [17:12:18] Nope [17:12:47] oh, you decided to sneak WT rules here [17:12:50] significant guard rails need to be implemented for a community elected board, this is proven as a requirement. [17:13:04] I like it. I think NSFW wikis don't need to be on Miraheze, as has been said before during the scare, one can host wiki on their own private servers or as direct programs on PC [17:13:04] what is wrong with them though? [17:13:15] You don't need the board to be anonymous, you just need them to not have to post their personal info publicly [17:13:28] why though, instead of just a "nope". explain. [17:13:29] The most important members of the community are not the meta people who just drama 50% of the time, but rather the individual wiki creators and administrators, the ones who run their own communities and just want to be left in peace and with less interference. [17:13:42] anonymous = only username posted for who they are [17:13:54] no real name or address [17:14:03] the NSFW wikis can stay but as mywikis mentioned, move them to their own custom domain and remove the Miraheze hosting footer icon. wash your hands of being linked to it [17:14:17] I think there should actually be members of the board not just of the community, but representative of groups who use the wiki [17:14:19] The first two points are tangential, minor points. Not relevant to Miraheze's future. The fourth point is not a good idea, and may not be legally feasible, as anyone can have a right to obtain certain corporation legal filings [17:14:27] Allowing bureaucrats requests would only end in chaos and NSFW wikis are the heart and soul of miraheze [17:14:35] So you have like, a seat for wiki editors, wiki users, stuff like that [17:14:44] if incorporated in the US, the board is not required to disclose PII [17:14:48] They shouldn't be [17:14:49] It's feasible, the distinction is that the set of company directors is not the same as the community board [17:14:49] Sarcasm but pretty much no other wiki farm allows nsfw wikis [17:14:53] also funny you never thought about LGBT+ history untill I mentioned it in WT server [17:15:06] if NSFW wikis are the heart of Miraheze, then we have a major problem [17:15:19] that's a massive stain on our reputation [17:15:23] Not necessarily. While corporate filings would not be free, in most jurisdictions, you can pay to obtain an annual corporation filing [17:15:26] What you do is you have the community elect members of the nonprofit, and then they choose the (public facing) officers [17:15:30] As said before, those could just be hosted and done through private means such as personal hosting or downloading and editing MediaWiki or whatever [17:15:30] NSFW wikis by default are fine [17:15:31] Why is that a major problem? [17:15:38] How would an anonymous board be a benefit, too? [17:15:56] the massive stain on MH reputation was left by reception wikis [17:16:01] How is LGBTQIA+ history NSFW? [17:16:06] Isn't defining something as "not safe for work" a form of censorship which Miraheze has a commitment not to do? [17:16:08] Honestly, I’m pretty sure a few of the top 10 largest wikis are nsfw [17:16:20] You guys are arguing about something irrelevant because you're assuming the community is directly electing the company directors, btw [17:16:21] Well-said. [17:16:24] I think it's more about banning porn wikis [17:16:28] LGBT+ history is not NSFW [17:16:38] Don’t you think we are all getting ahead of ourselves [17:16:46] NSFW wikis have long been out of control on Miraheze and have been a reason why Stewards have so much on their workload. Stuff like feet fetish wikis have been long criticized because it's a bad look for some (and has been said as much), that their donations are funding NSFW content. [17:16:52] I agree. The NSFW wikis were not an issue. I think we just need to have a requirement to have a NSFW template on the content pages, or perhaps we could make them non-visible to those that have opted out of NSFW content [17:16:58] it is not, but for big money people who push their rules on the web it is 🤷 [17:17:06] Because a main person people haven't run for the Board recently is that their PII is everywhere and publicly available. [17:17:26] In my 3.5 years here, they've absolutely not been a problem for 99% of people [17:17:27] In the US, the only person required to disclose PII when on a board is the person submitting tax filings to the IRS. [17:17:31] I quite literally never heard about "MH reputation" until I got on this Discord - I never even knew about all this drama until I saw the goddamm announcement and saw the problems here. We're a WIKI HOSTING SERVICE and we should conduct with professionalism, not caring about petty things. [17:17:37] I don't know who you're referring to, but if it's who I think, that is not the only reason, no [17:17:55] Would invite you to look at past conversations here re NSFW because it's absolutely been mentioned by many people [17:18:00] ikr [17:18:00] Any heavy workload was self inflicted and no one approached me to say they're not donating because of it, or anyone to my knowledge [17:18:14] Brandon [17:18:15] I think we need to use this opportunity to reframe Miraheze [17:18:41] It's more of a broad statement rather than pointed at a specific person. A lot of people just don't want to have their name and address everywhere, which is perfectly reasonable. [17:18:43] I have, and I've been involved. No community proposal was ever brought or successful to ban them [17:18:44] Exactly, we just need to provide the service we're supposed to provide, without all the added shit. [17:18:50] [1/2] you sneaked this here [17:18:50] [2/2] when nobody today or in the past week in this server ever discussed nsfw wiki [17:18:51] Slow mode increased to 15s. Play nice, all. [17:18:52] Miraheze is like the only wikifarm that I am aware of that openly allows NSFW wikis...everyone else accepted its a terrible idea. not sure why its censorship or a debate. [17:19:01] It's not censorship to not allow NSFW content, there are rules in society, there are consequences to speech and action. [17:19:06] <:thistbh:764649024875528192> [17:19:23] [1/2] you sneaked this here [17:19:24] [2/2] when nobody today or in the past week in this server ever discussed nsfw wiki [17:19:27] This channel is like a bug zapper for people getting distracted by controversy haha [17:19:35] Yes; yes it is [17:19:36] Let's just give people freedom to build their wikis and stop interfering unless it's something illegal/CP/doxxing etc? [17:19:37] exactly, You can say what you want, but you cant control what happens after [17:19:40] I mean outright porn wikis should be banned but wikis about nsfw games shouldn’t be, as well as nsfw literature [17:19:40] I don't think it was the NSFW wikis, though, or even the polandball wikis. It was all the drama and Reception wikis, the listicle-style wikis, a general lack of useful content people want to read and be involved with [17:19:48] GUYS [17:19:49] Brandon, NSFW removal is not on table at present as part of restructure. First and foremost order of business is ensuring continuity of current service, then talking about changes. [17:19:52] Banning porn wikis isn't censorship, it's professionalism. If there is a significant market for porn wikis, they can either privately host them or form their own wiki farm for that purpose [17:19:53] Guarantee you you're wrong, I just had a conversation about this yesterday [17:19:56] Don’t ban them, ban obscene wikis [17:19:59] Why should legal pornography be banned? [17:20:21] exactly, it will make everything more professional and bring a better reputation [17:20:29] You guys have bigger issues than arguing about nsfw wikis. I don't know why you are all engaging on this right now [17:20:42] Doxxing are outright banned on MH, so this is why reception also got banned [17:20:42] Obscenity is not protected under free speech in the US, and I think that’s where the wiki is going to be restructured [17:20:50] where? [17:20:57] somewhere in this server [17:20:58] We are about 30 steps ahead of ourselves right now and its not productive [17:21:11] dude common [17:21:14] Exactly [17:21:16] Yep. Gonna ask for a disconverse on this for now. There's a time and venue for it and it isn't today and here. [17:21:23] there's been so many conversations it's hard to keep track 🤷‍♂️ [17:21:34] [1/2] This, I think the rules should be [17:21:35] [2/2] Wikis simply serving as a repository for pornographic content, sexual activities, or other content which is not appropriate for ages below 18 should nor be allowed. But content documenting things in a context decoupled from overt pornographic or sexually explicit nature's should be allowrd [17:21:46] Anyways, we should focus on ENSURING SERVICE CONTINUITY - we're getting ahead of ourselves. [17:21:47] I don't think it is the issue. I've never seen seen anyone associate Miraheze with 'porn wikis'. In fact, we even hosted `pornwikiwiki` and it didn't have any pornography on it, but did have useful content [17:21:56] Please stop talking about this [17:22:03] yes, please [17:22:11] This is not the time to discuss changes!!! [17:22:15] They are, and MH was supposed to have NSFW warning or something like that when you surf those wikis [17:22:27] Repeating request for disconverse [17:22:31] Let's move on. [17:22:33] Timeouts will follow [17:22:44] Thanks all. [17:23:04] Big Wikia is watching us [17:24:34] anyways, once the new organizational structure is set up. the next step is a full postmortem by independent party. Even if you think you know what all the problems are, chances are you are missing items. [17:24:39] Anyways so is Miraheze being moved to the US? [17:24:54] Yea [17:25:00] Still being discussed, but this seems to be most likely outcome. [17:25:52] 🦅 🦅 🇺🇸 [17:26:28] [1/2] Most up-to-date info can be found here, for now: [17:26:28] [2/2] https://meta.miraheze.org/wiki/Miraheze_is_Not_Shutting_Down [17:26:35] sarcastic trash posting isn't helpful... [17:26:42] k [17:26:42] Tbh I do like that miraheze allows nsfw but also as long as it’s restricted (I don’t use them to be clear just stating a point) [17:27:05] It separates itself from the competition and as long as it’s legal nsfw stuff then it should be fine [17:27:20] Agreed but moving on, please. [17:27:29] Can we please move on from this? [17:27:31] Will this discussion be had in the future [17:27:44] Ok [17:27:50] It's not the time or place to discuss this, right now the focus is on ensuring service continuity, not making changes. [17:28:28] Changes need to be discussed to keep the community sharp, fresh, and able to survive tho [17:28:39] what we were on, again? [17:28:47] Volunteering management [17:28:51] 1984 big brother wikia [17:28:51] alright, so wasnt here for most of the conversation, whats up? [17:29:24] Yes, but this is not the time and place for discussions about things like NSFW wikis. [17:29:32] Yes ik [17:29:33] Yeh I do agree [17:29:47] [1/2] The future of NSFW can and will be discussed, but today isn't the day. [17:29:47] [2/2] As mentioned, current focus is on ensuring continuity of service, then talking about how the platform changes to ensure long-term sustainability [17:29:47] GUYS [17:29:49] indeed [17:30:04] Alright so what was discussed before the NSFW thing [17:30:15] thiw? [17:30:18] software developer/engineer seems kinda interesting. [17:30:29] This [17:30:37] From my perspective on the outside, miraheze does have a reputation problem, but the problem is a reputation for organizational instability, rather than porn. And this is a much larger issue, because it severely hinders attempts to get power players (good tech people, institutional donors) to support you [17:31:28] [1/2] Miraheze's reputation problem doesn't affect the most important users, those who just want to get their wikis set up and have a place for their community. [17:31:28] [2/2] However, this massive organizational instability certainly does. I thought my wiki was going to be deleted, and so did 100s of others. [17:31:51] It’s better to be alone and free than supported but chained [17:32:02] I agree but I believe that the porn only contributed to the reputation problem [17:32:15] Can we stop talking about the porn thing for now [17:32:24] frankly, the only one i agree with is the first one [17:32:25] Hm [17:32:27] https://tenor.com/view/youknow-you-gif-19056787 [17:32:28] Guys please stop coming back to porn / nsfw. We will start issuing timeouts [17:32:43] Yes please. This is not the time or place. As I've mentioned, we need to focus on ensuring continuity of the service, not discussing about "muh porn". [17:32:46] Yep, we're done on this topic. [17:32:53] Like, prior to the previous drama in March, we were thinking about just donating to cover all of your infrastructure costs for the foreseeable future, because MH is super useful to the overall MediaWiki ecosystem. But then shit hit the fan and I was like, wow, this is radioactive and risky to be associated with, I don't see anyone in charge that I can trust [17:33:01] What if you went byond miraheze to find volunteers [17:33:04] I believe though Brandon is right, there needs to be a remaking of the roles [17:33:06] i dont see why meta sysops should be eliminated and replaced by cvt. [17:33:10] exactly, we need to be professional and focus on service [17:33:26] Having meta as an independent wiki has always been kind of an odd choice. [17:33:43] As if you had enough volunteers they could work shifts [17:33:54] It's not precisely a topic wiki, but rather a global hub to set standards for all hosted wikis. [17:34:06] i dont even think that anonymizing the board is feasible :p [17:34:21] Shouldn't there be like, a default thing? [17:34:28] but frankly i dont know much about laws regarding companies so [17:34:39] [1/2] people, I'M SORRY I HONESTLY didn't think this shit would explode, I only noted that Brandon sneaking something absolutely unrelated to ongoing discussion [17:34:39] [2/2] aight [17:34:41] When you search up Miraheze it should bring you to the stuff metawiki serves currently [17:35:01] US incorporation law is a bit different in this regard. You don't need to have everyone on publicly visible blast with name, birthdate and home address. [17:35:05] Question for all if anyone knows: Do we know if donations have been paused through all venues for the current Miraheze iteration? [17:35:33] It's just chairman or whoever files to IRS afaik [17:35:35] What does that question even ask [17:35:39] Funnily in the U.K. all the directors used @owenrb address I think? [17:35:52] It's not wrong until it is. 😄 [17:35:56] For service address [17:36:05] wasn't really trying to sneak it, just mentioned it as an idea. really didn't expect it to blow up into what it became. [17:36:21] You still record an actual address. Fossbots shows a random office in london. [17:36:22] Legroom, Brandon: Drop it please. [17:36:29] Ok [17:36:37] I think we need to kinda rework everything in the bureaucracy and administration [17:36:41] Is it a FOSSBots address or just some office [17:37:00] Cause like, at least to me, an average user, the whole thing seems so over-complicated given what it could be [17:37:09] Formation Agents office [17:37:28] What's a Syop? [17:37:39] Psychological op [17:37:45] What's everything else, that's the thing, it's all very convuluted [17:37:47] guys, can we let the group working to finalize do so in peace? [17:37:48] sysop you mean? [17:37:58] huh? [17:38:02] Yea [17:38:05] administrator [17:38:19] sysop is system operator (i think?) [17:38:21] "system operator" i believe [17:38:29] admin role bellow bureaucrat on wikis [17:38:39] but people use sysop and admin synonymously [17:38:39] That's what I'm saying though, the current administrative system and bureaucracy is so convoluted [17:39:03] A person who manages the paperwork for us [17:39:05] It's the standard structure in MediaWiki and all wikis have it, not just Miraheze [17:39:20] It's interesting haha, on Meta the technical name is `sysop` but in Special:UserRights it's called administrator [17:39:20] They also forward letters and stuff. [17:39:22] It's also a fairly simple structure to me [17:39:26] Well maybe a bit of simplification or explanation could be in order [17:39:41] @notaracham#0000, that is generally true, yes, though it may vary by state of incorporation. It is also generally true of Canadian corporation laws by province. However, the filings may be subject to freedom of information disclosure laws, meaning anyone could still obtain the corporate filing, upon payment of a fee. So it is not necessarily 'private'. It's just not on the open public web :) [17:39:45] The biggest complaint I've seen is that Miraheze has a very bad bureaucracy with many redundant roles [17:39:49] idk about yiu but I knew this since FANDOM [17:39:55] I've never even see the office fossbots is based at. I've walked round the area before but never been in or looked for it. [17:40:00] dmehus: I think the consideration was incorporating in Delaware? [17:40:13] Any additional roles would have been added by the community in the past [17:40:16] could be wrong, correct me if I am [17:40:26] The per-wiki structure (sysops/bureaucrats) is fine, but I would strongly recommend simplifying all the global stuff [17:40:34] It is not [17:40:52] So could you outline the current structure and what you propose? [17:41:01] hm. not sure where. heard that then, it was something about taxes I vaguely remember [17:41:08] I think your idea sounds interesting but I'd need to compare how it all works rn, and how you want it to work [17:41:14] You don't really need a trust and safety team, you probably don't need 6 segmentations of your tech volunteers, you probably have more complicated global steward/meta roles than are necessary [17:41:30] It was an option [17:41:43] I'm less familiar with Delaware corporate law, but likely there is some degree of public disclosure [17:41:44] Trust and Safety is needed, IMO. just not maybe how it currently is [17:42:02] I agree with the tech bit though [17:42:06] So T&S was created as ToU enforcement needed to be standardised and regulated. Tech is just two teams, not segmented too much [17:42:19] What you need is a single person that runs things day to day, plugs the holes, and is accountable to the community elected board [17:42:23] What's wrong with the current Trust and Safety structure? It's a very minimal role. Its involvement is only what is legally necessary [17:42:48] @cookmeplox could you make like a graphic or something? [17:42:49] At least from my observation, it being segmented has lessened collaboration and communication to some extent [17:43:06] It's not segmented though? [17:43:14] It's an unnecessary organization that causes a crazy percentage of your problems relative to its value [17:43:25] The T&S and SRE teams collaborate privately, via Discord, just as the board does with T&S and SRE [17:43:30] A team that focuses on MediaWiki and a team that focuses on everything else [17:43:38] That's what's been communicated to me at least - I'm not in SRE so I can't tell but from what UO, others have talked to me about, the separation of Infra/MW seems to have been damaging [17:43:42] ^ :thistbh: [17:43:55] (in reply to Own's comment) [17:44:11] Owen's* [17:44:12] I think the separation has actually been really really useful. I think @paladox can speak from experience [17:44:22] I agree [17:45:07] Before it was a distraction from SRE's duties, and sometimes deprioritized over other SRE/MWE tasks on Phabricator [17:45:15] ccing @reception123 here [17:45:34] I mean, this debate about bureaucracy vs one man in power comes from Ancient Rome, at least. Any system can work, as long as the people in charge are capable and skilled. But when you don't have any certainty about those people's skills, or a high turnover rate, organization is everything... again, as long as you have enough capable people on it [17:45:43] I'm sure he's following this thread as he has time :) [17:46:32] In recent months, I agree that split was probably not useful. But that's not because of the structure, that was because you had a lack of access to resources and skills to effectively do it [17:46:38] @owenrb if I move again, who do I tell me new address at the moment? [17:46:44] Owen was accountable to the board; it just so happened he was DT&S and Chair/Secretary of the board, which was a function of the board having only two members (a third was added when CA/UO was nominated as the Technical Team's representative) [17:47:21] RhinosF1: I'm not sure you need to update your address? I could be wrong, though [17:47:30] dmehus: NDA people do yes [17:47:31] Agree for sure. It's ahead of ourselves at this current stage, but we could consider combining sections until new volunteers are found? [17:47:38] Probably not needed now [17:47:41] Owen has my current address at the moment [17:47:53] It will change at some point in the next few months [17:48:08] Likely august [17:48:08] RhinosF1: I wasn't aware of that. I'm also not sure to what extent that is audited, either, even if it's a requirement [17:48:10] If you combine them, you don't fix a problem, minimize a problem or anything. Whether they're split or not changes nothing [17:48:25] dmehus: Owen performs 0 checks [17:48:36] RhinosF1: ah [17:48:36] But he did say I had to tell him last year [17:48:49] Owen: for clarity, I have not moved. My address remains the same :) [17:49:12] It's not explicitly said anywhere [17:49:50] though I did buy a condo presale, but will likely retain my current address as my mailing address [17:49:51] We should focus on finding more volunteers as one of our top priorities. More volunteers means more work done and less burnout. [17:50:01] Generally it's not a legal requirement (or even common practice) to update an address on an NDA [17:50:15] RhinosF1: ah, okay, yeah, that's what I thought [17:50:37] If anyone wants to reach out to people once this organizational thing is done....probably would be welcomed. [17:53:32] respectfully, you are kind of late, this has been stated as major thing days ago and primarily happens in #tech-community [17:54:05] [1/2] yea I know I'm kinda late 😭 [17:54:06] [2/2] But there was a conversation ongoing so I joined in [17:57:24] In my experience, just blasting out a call for new skilled volunteers (especially tech) is usually a net drain on resources unless you are quite careful about it. You'll get about a 2% conversion rate from people expressing interest to anyone actually contributing [17:57:54] 2% of our numbers is 0 [17:58:03] The EOIs have been low so far [17:58:18] Well, not zero-zero [17:58:24] But we didn't retain the few we got [17:58:29] So effectively zero [17:58:44] The people expressing interest are well-meaning but most of them go into it thinking "these guys just need someone to write them a python script to automate stuff", without realizing you're essentially looking for skilled long-term oncall [17:59:27] Yeh. [18:00:12] If I were you I would focus on shoring up the existing volunteers, getting them aligned on a cool project together, and start looking at what parts of their job could be made easier by some initial upfront time investment [18:01:34] Whether that's automating more stuff, getting infra that isn't constantly on fire, building more self-service tools so you're editing localsettings less often [18:02:48] You will have a far easier time getting really good external MediaWiki people (like me haha) to make one-time large contributions, rather than signing up to do oncall forever [18:03:13] Ping me if anyone needs anything or wants to talk - I'll be viewing but probably won't be intervening for a bit. Hope all remains well with the Miraheze community while I'm gone, I have the utmost confidence @.labster will pull us through this time. [18:03:23] true, self-service tools & automation would really make everything much easier for both volunteers and individual wiki creators. [18:03:48] THIS [18:04:21] But it’s easy saying that. Implementation is different. Especially when you have to consider how to do it in a way for security reasons [18:04:44] E.g for example ssl certs. You don’t want private certs being leaked [18:04:46] but as my final message: ImportDump has helped a lot, hopefully RequestSSL will soon, and for community (not SRE), there is a Automatic Interwiki adder extension request somewhere on phab [18:07:05] Silly question if anyone knows: what is the permission limit on access to Special:DataDump for a wiki? [18:07:14] Any user, or must have admin/bureaucrat rights? [18:07:52] any user [18:07:54] I think whoever the permission is given to, sysops by default to generate and users to access? [18:08:09] well actually....Special:Export is any user but can check [18:08:29] Thanks, that's the set I was thinking was the default, will dig into permissions to check particulars [18:08:56] it's any user, at least for Meta... [18:11:27] Confirmed, default permissions to view, generate and delete dump are consolidated on Administrator. [18:11:51] Average users do not have access to the Special:DataDump page [18:12:41] I stand corrected then 🙂 I think that kind of makes sense anyway, whether it was sysop or user for viewing [18:13:08] Yeah, looks like MH is the odd one out, either by dint of how it was set up in a previous era of default permissions or an intentional change [18:16:00] @notaracham#0000, default permissions are set at [[Special:ManageWikiDefaultPermissions]]. Any changes only apply to newly created wikis from that point forward [18:16:00] https://meta.miraheze.org/wiki/Special:ManageWikiDefaultPermissions [18:16:22] unless specified in LocalSettings.php for all wikis, then that would override MW DefaultPermissions [18:16:35] I believe it's by default set to `sysop` yeah [18:20:35] Yep, that's the conclusion I reached as well. [18:20:45] Back to guide-writing [18:25:43] <.labster#0000, replying to realcastle#0000> I hope we can resolve the fork, but it would also be nice if both sides succeeded. At the very least I want us to cooperate on software. [18:27:40] Well Shoutwiki isn't letting me make a new wiki [18:27:57] So this is kinda bad, I don't got a alternative on me [18:28:29] WikiTide? [18:28:32] wait theres an issue between miraheze and wikitide? [18:28:49] There isn't [18:29:43] [1/2] #Alternative Wiki Hosting Services [18:29:44] [2/2] although the latest statement is that there are more chances for Miraheze to continue running [18:30:34] Reducing slow mode, thanks all [18:31:11] kinda lame there's no "to the first message/top" button in those threads [18:31:36] Discord has room for improvement [18:34:34] <.labster#0000> Banks do enforced breaks. I think Miraheze should do them when we have as much money as a bank. [18:34:47] miraheze is unbeliveably slow at the moment [18:34:55] Normal day at Miraheze [18:35:12] its worse than usual, taking like minutes to load an article grrr [18:35:36] infra on fire again [18:36:40] <.labster#0000> Stop making backups guys (I don't actually know what is causing it) [18:37:22] I know this is perhaps a bit of an odd question for me to ask, but... SREs here are unpaid volunteers? [18:38:05] That's correct. There are no paid roles at Miraheze [18:38:11] People quit six figure jobs from Google, Facebook, Amazon, and Microsoft because they have to do on-call. Truly all the respect goes to the volunteers. [18:38:56] Amen [18:39:50] <.labster#0000, replying to !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!#7050> I'm forming a 501(c)3 not a 527. [18:40:29] Also, we took a look at WikiForge's pricing. Even if they have 1,000 paying wikis, there's no way they are going to be able to pay each of their staff a reasonable SWE wage for the amount of work they are doing. If they are a for-profit corporation, I doubt they will be making much profit other than to sustain their operating costs. [18:40:59] I guess you should have it so api.php hits different mw servers same for thumb_handler and have it so jobs aren’t run on the servers having traffic rather on other servers. And have more powerful servers hosting mw with a lot more cpu power per mw [18:41:17] What’s swe [18:41:22] SWE = Software engineer [18:41:25] sorry but keep this channel for Miraheze discussions [18:41:26] Oh [18:41:29] It is image dumps [18:41:32] Sorry! [18:41:45] Oh yeh database servers need more processing power [18:41:57] Yes, it is unlikely WF will ever be successful when WikiTide undercuts their own offerings for free. [18:41:57] Including more ram and to figure out why one frequents oom [18:42:05] people should use scripts 😔 [18:42:37] I'll move this discussion to #Alternative Wiki Hosting Services. [18:42:41] <.labster#0000, replying to talthiel#0000> Everyone agrees that there should be change, and change right now. Of course not everyone agrees on what this change should be, so it's important to go slow enough to actually capture what works best for volunteers and wiki admins. [18:43:42] I support Tizian, so do some of my friends and neutral people who are users of the site. [18:45:00] I've offered to do a deep dive for a few weeks into your infra issues once there is a stable organization btw [18:45:17] perhaps a formal statement that informs a plan is being finalized would reduce that? possibly a twitter announcement [18:45:22] Part of your problem is just that the hardware is exceptionally old, but I think there are probably some quick wins [18:45:53] One has been made. I linked you to it. [18:46:15] wouldn't 501(c)(7) be better? under impression that is the better option, less confusing [18:46:47] i'm aware, that was more intended for everyone else so they quit panicking with dumps [18:47:00] I don't think 501c7 makes sense since there aren't dues-paying members, no [18:47:06] No [18:47:11] [1/2] To give some context for what I am used to: being a full time Site Reliability Engineer usually pays >$100,000 in the U.S. tech industry and burnout is frequent. It is incredibly important to be able to sustain a service with multiple SREs and not have them burn out. Not paying them and/or not giving enough time off is a recipe for disaster. This isn't [18:47:11] [2/2] something I am saying to woo people over to MyWikis or any other place. I am saying this as a fellow human being who works in the tech industry and deals with burnout from my coworkers at FAANGMULA. [18:47:48] I don't think anything you're suggesting is prescriptive for success though [18:48:08] what is MULA? [18:48:32] Microsoft, Uber, Lyft, and I forgot what the A stands for [18:48:56] facebook, amazon, apple, netflix, google, microsoft, uber, lyft... [18:49:21] oh Airbnb [18:49:42] oh right, yes Airbnb, thank you [18:49:53] first time hearing that long one [18:50:16] would a 501c7 require paying a membership fee [due]? [18:50:21] these days people on Blind also say MAAMA (Meta Apple Alphabet Microsoft Amazon) since Netflix isn't part of the Big 5 in tech [18:50:42] which state would be chosen for this new Miraheze 501(c) incorporation? [18:51:14] a 501c7 doesn't even allow for tax-deductible donations, so it makes no sense [18:51:39] That's the primary benefit of even going in the 501c space in the first place, instead of just a C Corp or similar [18:52:23] well, that and not paying corporate income tax :p [18:53:03] There are also usually legal requirements on a c7 about a percentage of revenue needing to come from membership dues, although it has changed recently iirc [18:53:13] either way, not an idea that works [18:53:33] In the US, there's also the added confusion of certified b corp vs benefit corporation, one of which is a tax and incorporation status, the other is very much not. [18:53:52] 501(c)3 makes the most sense, and I don't think anyone would seriously contest that. [18:54:00] someone seems to have contested it haha [18:55:11] I suppose lol [18:56:42] [1/5] https://www.irs.gov/charities-non-profits/other-non-profits/exempt-purposes-code-section-501c7 [18:56:42] [2/5] `A social club must be organized for pleasure, recreation, and other similar purposes. A club will not be recognized as tax exempt if its charter, by laws, or other governing instrument, or any written policy statement provides for discrimination against any person based on race, color, or religion. ` [18:56:42] [3/5] https://www.irs.gov/charities-non-profits/charitable-organizations/exemption-requirements-501c3-organizations [18:56:43] [4/5] `To be tax-exempt under section 501(c)(3) of the Internal Revenue Code, an organization must be organized and operated exclusively for exempt purposes set forth in section 501(c)(3), and none of its earnings may inure to any private shareholder or individual. In addition, it may not be an action organization` [18:56:43] [5/5] maybe someone can clarify the difference for tax deductible donation... [18:57:16] You aren't actually being serious... right? [18:57:20] I personally like the idea of a membership-based non-profit, as the paying members the ones able to elect a board but not receive any additional permissions [18:57:50] i'm quoting IRS website, both of these are tax exempt [18:58:18] I can go fetch law.cornell.edu if you so wish [18:58:28] 501(c)3 status is likely the hardest to obtain, no? Since they can issue charitable tax receipts [18:58:50] I mean to me this is more of a social club if anything...? [18:58:58] its just digital [18:59:18] Does the Miraheze data reside in the EU or the UK? If so, what would the legal ramifications be of having an American entity control that? To establish MyWikis Europe, we essentially had to come up with a plan to airgap the EU data from anyone outside of the borders of the EU. Might be something to consider for regulatory purposes. [18:59:27] It resides in the U.K. [18:59:39] Potentially. It may be seen as a hosting service, though, so it may not qualify is what @mjl1 might be saying. But then would it qualify for 501(c)3? [18:59:40] regardless of either, you will be under a big regulatory limit on political or religious content [18:59:53] I have no idea what Brexit has done to GDPR in the UK. [19:00:00] Does our Privacy Policy even allow for a data transfer? [19:00:01] We still have the gdpr [19:00:17] and this stuff sounds like something that should be consulted with an attorney licensed in the state of incorporation [19:00:21] there's something about that in ToU, no? [19:00:22] EU keeps killing the data transfer agreements... [19:01:07] Have you seen the latest Facebook decision? [19:01:09] If the current board were to establish an agreement yes [19:01:13] From my understanding, and I am not a lawyer: that is not as simple as saying "we can take your private data outside of the EU/UK!" in the privacy policy. There is some sort of procedure that the GDPR prescribes. [19:01:22] 7.3 does allow for it [19:01:55] change the privacy policy then? [19:02:02] yep, I have my doubts on compliance [19:02:05] ugh, no. You want it to be organized as a charity. The mission is to provide free knowledge or whatever. Look at WMF and all their affiliates. [19:02:10] I think you guys are getting ahead of yourselves and talking about things you don't know about [19:02:14] some stupid ideas here lol [19:02:14] That's quite possibly true, yeah...I just confirmed it is in our Privacy Policy at 7.3 [19:02:51] Facebook have unlimited budget, fines are like sands to them [19:03:12] <.labster#0000> I have achieved 501(c)3 public charity status for a wiki farm before (and not 501(c)3 private foundation). It's definitely possible. [19:03:12] Yes doesn't mean the decision can't affect others [19:03:24] Something something called precedence [19:03:43] ^most likely don't until EU completely blocks them [19:03:44] [1/2] https://www.law.cornell.edu/uscode/text/26/501 [19:03:44] [2/2] denotes educational which is the legal shield for most wikis hosted here [19:04:25] Have the people who intend to take over the project consulted with lawyers in England or Wales and the intended state of incorporation in the U.S.? [19:04:40] Not yet [19:04:47] its holiday today so wait until tomorrow [19:04:56] Ooh, tomorrow is Juneteenth. [19:05:12] that's not entirely recognized as a day off [19:05:22] Jeff you're not helping lol [19:05:34] It's now a national holiday in the US afiak [19:05:47] just said that [19:05:49] you can say that for a lot of people here, Cook [19:06:00] Let's be polite [19:06:01] whosjeff [19:06:10] mywikis is jeff [19:06:15] The person behind this account is Jeffrey. Hi! [19:06:20] oh hey Jeff [19:06:38] oh @notaracham when you have a sec - @dmehus should be removed from GIWA role [19:07:04] ah [19:07:09] procedural bit there [19:08:15] [1/2] I am providing the bits and pieces of info I can remember from our extensive discussions between our business partner in Austria and our attorney in Austria to make MyWikis Europe possible and compliant with EU laws. I have firsthand experience with the U.S.–EU data boundary thing. Unless there is some rule violation I am making, I would prefer to talk about, whether anyone thin [19:08:15] [2/2] ks it's "helpful" or not. And I would encourage you to only make such a judgment of whether I know what I am talking about if you have been in a similar position. [19:08:54] MyWikis might be a competitor to Miraheze, but people being sued for breaking the GDPR accidentally helps nobody. The fallout to the whole MediaWiki community would be devastating. [19:09:09] I have been in a similar position, I run the RuneScape wikis [19:09:17] Oh, you are Jonathan? [19:09:26] the one and only [19:10:07] well, that settles that. MyWikis is Jeffrey and Cook is Jonathan. [19:10:53] real names only in discord usernames, go! [19:11:02] (this joke brought to you by Twitter) [19:11:05] Hard pass. 🙂 [19:11:22] Well, you and I have had a history for the past 10+ years, so I won't respond any further. I am here to provide my thoughts and to answer questions about moving in the #Alternative Wiki Hosting Services. [19:12:28] Regarding the substance of the GDPR/data boundary concerns, generally you only start to run into problems when the GDPR-relevant data is stored outside the EU/UK [19:12:37] [1/2] ok, I'll concede though still unsure if this lawfirm is accurate. law.cornell.edu has less info than IRS... [19:12:37] [2/2] https://www.wagenmakerlaw.com/blog/social-club-q-section-501c7-tax-exempt-organizations [19:12:38] And that is not the case currently, and probably won't be the case in the future [19:14:34] @.labster has already said the intention is to make it 501(c)3, and that this is possible. There are no benefits to making it a 501(c)7 rather than 501(c)3. Like, none. [19:15:18] noted [19:15:53] It's not just that there are no benefits, it's that you legally can't unless you change the funding structure [19:20:26] <.labster#0000> 501(c)7 isn't happening, guys. Talk about something productive. [19:22:01] Okay, so just to clarify, if someone exports a XML dump from Miraheze, that is generated completely from data stored at rest in the UK? [19:22:14] Now I am asking because this affects the migration procedure at MyWikis. [19:22:34] AFAIK yes [19:22:50] Thank you very much! [19:26:42] I believe all of our physical hardwere is hosted by server choice in the UK. the only data we have anywhere outside of the UK is static html cached elsewhere [19:27:07] Okay, perfect. That will affect our migration process then. [19:43:48] It will be interesting to dive into why you guys have so many CPU cores engaged [19:44:32] you're averaging about 50 CPU cores in use to serve about 8 pageviews a second. That's like 7 CPU-seconds per pageview - something doesn't add up [19:45:08] Uh thinking who could answer that [19:45:37] I don't think anyone can answer it without a deep dive [19:51:34] but, like, your mw servers are averaging like 83% utilization. that's nuts [19:51:47] first of all that's maybe a sign you should be devoting more resources to that compared to the other stuff [19:52:08] but also like...why is it THAT high? is it job queue parses? actual page serving? etc [19:55:39] similar thing with the swift object servers, why are they averaging 82% utilization? how many images are actually being served a second? [19:55:52] would be cool to get a ~1 minute chunk of the nginx logs or something [20:08:34] Well that would be the I/O [20:09:02] And nothing like the issue affecting mw servers [20:10:38] @dmehus see DMs [20:16:39] I'm not sure I buy that personally [20:17:01] Especially because even if the root cause is wildly different, the underlying solution (better caching layers) might solve both [20:17:36] I mean that’s what the graph shows. But neitherless the server is shitty as well (it’s similar to cloud12 and is a hp) [20:17:53] Dreadful cpu performance. It’s a bit better on cloud13/14 [20:18:07] As those are dells with more powerful cpus and also a bit more modern [20:18:33] sorry what I meant more generally is that most of your problems are i/o based [20:18:41] Ohhh [20:18:48] No I meant for swift [20:18:49] including the mediawiki servers. mediawiki latency is dominated by i/o latency to an extent that nobody really realizes [20:19:11] Because the object servers use hdds where as the rest of the stack is ssds [20:30:11] Iirc Miraheze job rate is 0 they run it via Cron. [20:30:48] what servers do they run on though? [20:31:54] All the mw servers [20:32:01] And the mw task server [20:32:53] https://github.com/miraheze/puppet/blob/master/manifests/site.pp [20:33:01] Probably helps figure out what is running on which server. [20:34:42] Is Miraheze using a full rack in Stevenage, [20:34:57] (Not really related just wondered) [20:35:25] Well we were given a full rack I think but we are limited on power [20:35:33] To increase it costs a lot lot lot more [20:35:51] Ahh, yeah I remember you guys had to shut some things off didn’t you because you were using over the limit? [20:36:00] Or something like that [20:36:05] Well they were very lenient [20:36:15] We still have to switch it off… (that being cloud10…) [20:36:26] It’s still running? [20:36:35] Yeh it’s got db101 and db112 [20:36:49] The disks on cloud13/14 are broke [20:36:49] so whether the job rate is 0 or not is kinda irrelevant to exploring this, right? [20:36:53] Oh neat I’d have assumed they’d have been pretty strict about that [20:37:15] When was that supposed to be done by? [20:37:34] Last year I think [20:37:55] 😳 [20:37:59] I’m not entirely sure [20:38:10] so mwtask141 is the only one that is purely dedicated to job runner stuff? [20:38:16] Cloud14 isn’t looking really good. I’ve noticed speed lagging on the disk [20:38:20] do you know how jobrunner stuff is load-balanced? [20:38:23] Compared to cloud13 [20:38:40] like would mwtask141 be doing the same % of jobs as each of the mw### servers? [20:38:41] Was wondering why my wiki was going slower [20:39:22] Yeh, it’s supposed to be where maintenance commands are ran as well [20:39:41] i'm curious if you know the answer to my most recent q [20:39:49] because https://grafana.miraheze.org/d/W9MIkA7iz/miraheze-cluster?orgId=1&var-job=node&var-node=mwtask141.miraheze.org&var-port=9100&from=now-24h&to=now-1m looks pretty not-busy [20:39:54] We have specific jobs that aren’t allowed to run on mw* but only are for mwtask [20:40:06] Like datadump jobs [20:40:29] but for something like a refreshLinks job (which I assume you guys run), where is that gonna end up? [20:40:46] I’ll check [20:41:25] Looks like all the mw* servers including mwtask [20:41:32] https://github.com/miraheze/puppet/blob/ec999b36bed3407b4b9779c409d77e3783bf0fc0/modules/mediawiki/templates/jobrunner.json.erb#L4 [20:41:47] Is for all mw* [20:41:55] https://github.com/miraheze/puppet/blob/ec999b36bed3407b4b9779c409d77e3783bf0fc0/modules/mediawiki/templates/jobrunner-hi.json.erb [20:42:01] Is for mwtask [20:42:53] Actually just mw* I think [20:44:17] Wonder if allowing mwtask to claim any job would help? [20:45:20] i guess what i'm getting at is [20:46:05] if mwtask is doing at least as many jobs as the mw### servers, then it follows that the utilization graph of https://grafana.miraheze.org/d/W9MIkA7iz/miraheze-cluster?orgId=1&var-job=node&var-node=mwtask141.miraheze.org&var-port=9100 is probably a good upper bound on how much of the mw### machines are being used for job running [20:46:14] and in general, mwtask is NOT very busy [20:46:31] which means mw### is not very busy with jobs. but it's still extremely busy with other stuff [20:46:41] and the question you'd want to answer is, wtf is that other stuff [20:47:04] Tbh that would be a good idea [20:47:39] I don’t think production mw servers with traffic should be running the jobs but it’s needed to deal with the queue [20:48:05] The good ole job issue [20:48:34] i guess my point though is more...unless one of my assumptions is wrong, your mw### load issue isn't due to jobs running on prod webservers [20:48:49] it's due to something else. maybe bad redis/db latency [20:49:07] yeah, probably not, mw122 shows the CPU utilization is concentrated on php-fpm workers [20:49:37] Is there any way to balance that anymore? [20:50:08] balance it among what? each of the 6 mw### is averaging >75% CPU [20:50:35] some of them are worse than others [20:51:34] but something is just not adding up. if you're doing 600k pageviews a day it doesn't make sense to me that you'd have basically ~30 CPU constantly engaged [20:51:45] that's 4 seconds of just pure parse time per page, which I don't think you have [20:51:48] Think jobrunner should be switched off on me121 and 122 [20:51:56] Those are two servers who struggle the most [20:52:08] Don't think that would help much [20:52:25] The problem is probably more here than anywhere else [20:52:42] Interesting, didn't know there were servers that began with "me"; I thought they all began with "mw" [20:52:58] Ohhh sorry typo! [20:53:04] do you have nginx logs (probably on the cp* servers)? [20:53:10] Yeh [20:53:24] Do they contain request latency in seconds? [20:53:30] I'm terribly sorry, just to clarify for myself, that CPU usage also was like that a month ago, when people weren't harassing dumps? [20:53:30] I changed the whole stack back to also log locally a few log files in addition to logging to a central server [20:54:16] Question, if we were to allow mwtask to take any jobs, could we then dedicate more to php-fpm where we currently are dedicate process for jobs, or would that not matter [20:54:37] if I were you I would prioritize figuring out WHY php-fpm is so damn busy [20:54:45] rather than just throwing more resources at it [20:54:56] because there is a math problem here that doesn't really make sense [20:55:26] Could it be dump requests causing in increase of php-fpm? [20:55:43] it doesn't look like the mw* load has appreciably changed in the last month [20:56:02] actually i take that back a bit [20:56:20] Its been a long time since I’ve been a sysadmin so im not entirely sure how things are setup anymore [20:56:22] it's gotten a bit higher, but was still mind-bogglingly high every day before then [20:56:29] nginx logs don't seem to have any timing on them [20:56:45] dang. that would be a good thing to add as a first investigative step [20:56:54] yeah, I can change that [20:56:56] though [20:57:27] the php-fpm slowlog is gigantic [20:57:33] https://tenor.com/view/great-scott-surprised-amazed-oh-my-shocked-gif-16490502 [20:57:59] (btw that graph is on a server with 6 cores. soooo...) [21:01:17] Is that on mw121/122 @cookmeplox [21:03:10] what has been going on while iw as away [21:03:44] Not looking anymore but yes I think so [21:03:55] Seems about right [21:03:56] Cocopuff2018, when were you away from? [21:04:25] Ahh yeh the load on there will be higher then 131 but yeh php still a problem neitherless [21:04:46] you may want to focus on finding fresh folks who are reliable rather than try to convince old problematic folks to return... [21:05:29] I think this is a bit naive and also not conducive to moving forward [21:05:35] We have php childs set at 20 per server [21:05:37] my computer for three days [21:05:48] uhh so is the sre team stepping down [21:05:54] oh yeah dmehus, if you're trying to ping someone with the tag `#0000` (or #0), just get rid of that part [21:06:00] That's not necessarily a problem, I think ours is set to 100 [21:06:01] @fiddlestix check my dms [21:06:01] its part of the new discord username update [21:06:11] Which I kinda think is low but changing it higher causes load to go up well high. (For instance 32). [21:06:46] I think the db could be playing a part and it could also be the servers may need tuning (sysctl config changes) [21:07:08] My advice is to add the latency to nginx [21:07:12] fiddlestix#0000, ah, thanks. Yeah, wasn't sure if cross-platform pings worked or not. They used to, then they didn't. Sounds like they do again, if you include the disambigatuator [21:07:21] disambiguator* [21:07:21] We only give the db 4 cores esch [21:07:29] Because one of your assumptions had to be wrong, I'm just not sure which [21:07:30] 👍 [21:07:37] I think @orduin is doing [21:08:00] I'm poking a few other things at the moment (easier to poke through what we have than to introduce new changes to logging) [21:08:04] Dough check dms [21:08:05] And then once you do that, take a look at, say, 1 minute of nginx logs [21:08:36] Cocopuff2018, yeah, Owen has resigned as Board Chair/Secretary and as DT&S. He's still a member of the board. Reception123 resigned, and the Board announced an orderly wind-down of operations and/or potential transition plan for Miraheze [21:08:59] Paladox also resigned [21:09:11] idk what can i do to help? [21:09:44] If the sum latency is, say, over 1800 seconds for that minute, then that's super odd but at least gives you something more to investigate, figure out why you either have more requests than you realize, or the average one is taking 4+ seconds [21:10:07] Stay with us [21:11:09] If the sum latency is much lower, then it means something else is eating up all the php-fpm processes, which would be much weirder but probably points to a more immediately fixable misconfiguration [21:11:45] I think it’s more than that [21:12:10] Or maybe the answer is "oh shit, we're serving a lot more traffic than we realized" [21:12:30] Which is also a possibility. I'm not sure who is monitoring those types of stats [21:13:14] I mean 50 actively engaged and not idle. I know the fleet is like 140 or something, but at any given moment it's about 60% idle [21:15:33] What are you guys using to virtualize the machines on the physical cloud* servers? [21:15:39] Proxmox [21:15:46] php-fpm used to be very good at stalling @cookmeplox [21:15:46] And kvm [21:16:19] We should have request stats. @paladox would have access. [21:16:27] I don’t have access [21:16:34] I resigned [21:16:36] What is measuring that? [21:16:40] Oh ye [21:16:46] @orduin then [21:17:00] To all users (including me tbh): I would recommend a pause on permissions requests. We don't know what Miraheze's structure will look like yet in this new format, so filing permissions requests won't help. Wait until after. [21:17:02] Best way is probably to look at Matomo [21:17:19] You might be able to pull php-fpm / nginx request logs [21:17:28] Matomo will overestimate because static too [21:17:36] Not sure if grafana has anything useful [21:17:39] From my memory we have 180-190k a day [21:17:40] static? [21:18:01] @cookmeplox files for public wikis didn't used to go via php-fpm [21:18:08] Might have changed with Swift [21:18:16] It doesn’t track static I don’t think. As we only insert the tracker on mw [21:18:23] Ah fair [21:18:34] That makes sense [21:18:36] Oh you're doing matomo clientside, ic [21:18:40] Yeh [21:18:55] Didn’t realise you could do it another way lol [21:18:57] I am thinking out loud and close to sleep, remembering everything is hard [21:19:37] I remember when we were in the low double digits (thousonds) now it’s in triple [21:19:38] Yeah honestly if one of you wants to throw me a chunk of the nginx logs, even without latency, that would go a long way to understanding the scale [21:19:56] You'd need an NDA [21:20:02] Noticed a lot of increase throughout covid I think [21:20:15] We went up pretty quick [21:20:18] Not if we take the PII OUT [21:20:26] I am hearing a lot of different numbers for daily views, it seems like you guys aren't quite on the same page [21:20:44] But yeah, happy to do an NDA, or you could probably just regex out the IPs [21:20:52] to clarify, what PII is there in nginx? [21:21:04] i don't fully understand what nginx is lol just a curious question [21:21:06] Ip addresses [21:21:10] @cookmeplox I would ask @orduin as I assume it falls to him now (cc @owenrb ) [21:21:13] ah, got it [21:21:28] IP, User Agent [21:21:41] I would want to keep the UAs in [21:21:43] There's also a lot of non public info [21:21:55] hmm looks like $request_time is used in Graylog but not in the access log on the servers themselves [21:21:58] I’m not aware of user agents being included as PII [21:22:00] Oh okay, thanks for clarifying, just trying to learn a bit moree 🙂 [21:22:18] @orduin could we just sign an NDA for @cookmeplox [21:22:20] Since one of the questions I'd want to answer is, are you guys just getting fucking hammered by search engine crawlers? [21:22:20] No one should include sensitive information in their user agent [21:22:32] That wouldn't show up in matomo [21:22:37] If they are going to be deep diving stuff, it's useful to be able to share as much as possible [21:22:46] But could conceivably be like 80% of your problems [21:22:47] Most likely, I've throttled bots before [21:22:49] That's a good question and has happened before [21:23:08] brb, building a graph in graylog [21:23:27] I tweaked bings rate once, they also used to be a weird never heard of bot that refused to abide by crawl rates [21:23:45] https://github.com/miraheze/puppet/blob/master/modules/nginx/templates/nginx.conf.erb#L86 [21:23:45] They claimed it was impossible to comply with crawl rate so I think I rate limited them [21:23:56] MJBot? [21:24:06] Bingbot is awful, they almost always have some sort of crawler bug that causes them to not respect robots.txt [21:24:26] Oh [21:25:10] Sounds familiar [21:25:12] Should the main page of new wikis be fully protected by default? [21:25:22] Poll bot failed [21:25:25] Yeah, notoriously doesn’t respect robots.txt [21:25:25] @paladox where did we used to keep the custom rate limits [21:25:46] mediawiki.conf.erb [21:25:56] And then varnish nginx config for sites [21:26:06] I'm 90% sure I heavily rate limited or banned it from Miraheze. I definitely banned it from fossbots for being utterly useless. It was crawling nonsense anyway. [21:26:44] Yes I banned it and it's still banned [21:26:52] And then there's the other crawlers that are even less useful and [21:26:54] https://github.com/RhinosF1/puppet/blob/master/modules/varnish/templates/mediawiki.conf#L26 [21:27:05] https://runescape.wiki/robots.txt here's our robots [21:29:22] Do you guys have separate nginx logs for each cp, or do they get combined into one nice place eventually? [21:29:25] @cookmeplox my memory tells me it was aggressively crawling Special:ChangeCredentials when I banned it [21:29:32] And all one log in graylog [21:29:52] Well we log locally for all servers and then we also send them centrally [21:29:56] Yeah going down an insane rabbit hole of garbage [21:29:59] We use graylog [21:30:13] We have ~40000 requests per hour from something that mentions Bytespider [21:30:23] Crawling old diffs, special pages with infinitely many parameter variations, etc [21:30:26] I decided to send it down the rabbit hole of a 403 [21:30:47] 35000 per hour from Bingbot [21:30:58] O.o [21:31:05] I tried to actually talk to them but their support was basically we don't care and will continue to ignore policy and good practice [21:31:06] Send me some excerpt of the nginx logs, happy to do an NDA [21:31:15] 20000 in the last hour from AhrefsBot [21:31:34] bytespider alledgedly ignores `robots.txt` [21:31:38] I am really good at counseling correlating nginx logs to resource usage [21:32:08] That's TikTok company's search engine [21:32:09] Okay so this is really good actually [21:32:16] Wrong message, byte spider is [21:32:23] Id ban bytespider tbh [21:32:29] Because all of those suck and you should block them [21:32:36] Well, don't block bingbot exactly [21:32:45] Figure out what it's crawling for starters [21:32:49] BingBot no but Ahrefs and Bytespider can be blocked [21:32:55] But the others? They provide nothing to you [21:33:08] And sounds like they could easily be using half your resources [21:33:19] I have no idea who owns the google workspace and bing accounts now [21:33:29] It used to be @reception123 [21:37:46] there are always these pages in my google console, always listed under "found but not indexed" [21:39:10] sorry for the pings uh [21:39:38] Idm [21:41:03] Huh, we used to block some things in varnish, not sure where that was moved to [21:42:34] <.labster#0000> Can you approve NDAs now, @orduin ? [21:42:47] Which blocks are you thinking of? [21:43:26] I think it could have been removed? [21:43:57] <.labster#0000> I’m 99% sure Cook would be useful and the remaining 1% thinks he’s ChatGPT trying to take over the world [21:44:22] only @owenrb can right [21:44:38] i may run for global syosp idk yet [21:44:46] I think what I was thinking of got replaced with robots.txt [21:45:20] Respectfully, this would get SNOW closed within the hour [21:45:23] secretary of the board can, but that position is vacant now, idk [21:45:41] it's still Owen then (right?) because he's the only one with the access to all NDAs [21:45:47] in the future.......... [21:45:54] ....still [21:46:17] no one said i could ever have a sucesfull run [21:46:36] how far in the future? [21:46:54] if it's tomorrow then my opinion stands [21:46:55] Our robots.txt looks really small [21:46:58] guys convo [21:47:07] if it's a year from now then maybe not [21:47:16] void do you have an estimate of how many requests in the last hour contained `/wiki/`? [21:47:23] we'll take this elsewhere, thanks for the reminder 🙂 [21:48:30] Let me see what I can do [21:48:33] robot.php [21:48:44] Which we serve over robot.txt [21:49:18] https://github.com/miraheze/puppet/blob/ec999b36bed3407b4b9779c409d77e3783bf0fc0/modules/mediawiki/files/robots.php#L4 [21:49:36] not tommrow a few months.... [21:49:56] please vote for me for steward thank u [21:50:15] sorry too early and, i dont think you fit for it [21:50:23] Lmao [21:50:31] dang ok, how long do i have to wait? [21:50:39] hey everyone go vote for me as steward [21:50:47] [21:50:50] oh god what have i started [21:50:53] uhh try 6 months [21:50:55] Good job [21:51:01] Coco please [21:51:24] for real though, it seems super likely to me that 80% of your mw### usage is getting eaten up by garbage scrapers [21:51:27] I don't see any issue with Cocopuff2018's comment [21:51:30] Giving a time is never effective for functionary requests. It's whenever you've developed the skills and community trust [21:51:40] ^ :thistbh: [21:51:54] and the fix to your load problem might not require a massive overhaul, but just blocking the bad scrapers [21:52:02] [1/2] oh boy, I have 5 hours to sleep [21:52:02] [2/2] good night [21:52:03] that might fix both the db usage and the web usage [21:52:18] before making the request i would look at all the advice givien [21:52:32] I think you missed the fact it was a joke [21:52:58] <.labster#0000> My Scottish clan is the Stewarts of Appin so I am qualified to be Steward by Levenshtein distance alone. [21:53:13] I would like to test that theory. It's probably not far off correct. [21:53:16] remember when i use to spell zppix's username incorrectly everytime @zppix [21:53:26] Everyone does, nothing special [21:53:29] is this voting for real? if so, where can i cast a vote [21:53:36] no, not real [21:53:38] do i get a 'i voted' pin? [21:53:40] y'all got too many hats [21:53:45] :( [21:53:49] wanna cast a vote i got a link for ya.... [21:54:00] www.miraheze.com/castmyvote [21:54:07] [1/2] `Deny all [21:54:07] [2/2] Allow DuckDuckBot` [21:54:09] I'm sure we can get one if you are every anywhere near any of SRE [21:54:12] <.labster#0000> As a programmer I only have a cowboy hat [21:54:20] Southparkfan got some merch for FOSSDEM once [21:54:22] I have some emotes of that, enjoy 🙂 [21:56:53] by short estimate, less than half [21:57:38] I don't believe we normalise w/index.php?title= to /wiki/ [21:57:55] I'm more interested in the total count, rather than a percentage [21:58:33] but also please just give me the logs lol. I guarantee I will find about a dozen interesting things in the first few minutes, and I think I could give you a regex to remove the PII if the NDA thing is not an option currently [21:59:03] in reality i am probably gonna be less interested in a day or a week or whatever. strike while the iron is hot [21:59:23] NDA is not an option at the moment, and I'm also not sure how to best export the relevant parts of the logs as they are in graylog [21:59:40] Just because you dont have the document? [22:00:10] <.labster#0000> Void you are chair of MH now. Make a judgement call [22:00:34] [1/3] Total requests in NGINX in the past hour: 414,000 [22:00:34] [2/3] Total requests matching `/w/` in last hour: 188,000 [22:00:35] [3/3] Total requests matching `/wiki/` in last hour: 92,000 [22:00:35] If its bc of the document i think i may have a copy on my other phone [22:01:07] wait holy shit [22:01:18] O.o [22:01:21] so you're serving like 6 million wiki pages per day? [22:01:33] well there's your problem [22:01:59] That would track ye [22:01:59] damn, maybe 80% was an underestimate for crawler % [22:02:04] These are the aggregate reports across all mw* servers, which may contribute to the numbers. [22:02:12] Assuming 414,000 in 24 hours about 9,936,000 [22:02:17] I swear we only just hit 1 million [22:02:32] what are the nginx requests that aren't /w/ or /wiki/? [22:02:39] 9.2 million in the past day [22:02:42] like does that include images? [22:02:50] hmm, that might be possible [22:02:55] @.labster#000, I don't think it's that simple. Void may not be as familiar with UK laws as Owen [22:03:10] The NDA is a template [22:03:21] You fill in the blanks and sign [22:03:26] It just needs void and @cookmeplox to sign [22:03:34] Yes, but where would it be stored? [22:03:42] Void's draw [22:03:48] Anywhere [22:03:54] this is like a microcosm of your organizational problems haha [22:03:55] I assume Owen stores them in an orderly fashion [22:04:11] I assume void has a filing thing [22:04:25] I have a folder full of documents, in it is a section for fossbots [22:04:42] I'm sure void can do that [22:04:48] <.labster#0000> Miraheze Limited is dissolving in less than a month. It could be stored in a gun safe or a sock drawer until then [22:04:49] Cc @owenrb [22:05:04] Probably fine, yeah, I just meant that Miraheze Limited may need to amend its records retention policy and schedules first [22:05:34] Why does cookmeplox#0000 need access to the access logs on his wiki? [22:05:42] All wikis [22:05:49] why? [22:05:54] Because he's becoming a consultant SRE [22:05:55] To determine infra problems....? [22:06:01] I think [22:06:02] And seems to know what he's talking about [22:06:12] oh [22:06:26] is this just until we get to the new organization [22:06:30] Looks like there's about 220,000 requests in the past hour that could line up as image requests [22:06:30] <.labster#0000> Bylaws are not a straitjacket [22:06:35] or should I go through and amend documentation [22:06:36] I mean, technically Owen still is on the board, so likely Void would want Owen's opinion [22:06:40] I don't think @cookmeplox said they had that much time [22:06:53] Dmehus: I think @orduin can speak [22:07:17] guys meta made a blogpost saying miraheze isnt shutting down [22:07:24] rhinos, rhinos, rhinos [22:07:25] Correct [22:07:27] Yes, they did. [22:07:32] There's not need to go change documentation at this exact moment. [22:07:53] Well if he's a permanent SRE then it's probably good to do so [22:07:59] But if not then no point [22:08:09] No, leave docs alone [22:08:12] Nah I am here to solve a specific problem for you guys [22:08:14] does NDKILLA maybe wanna rejoin sre [22:08:33] wish y'all would focus more on solving problems than on what your hat says [22:08:37] * dmehus misses @NDKilla [22:08:50] I have nothing to do right now haha which is why I was asking [22:09:03] documentation is also sort of fun to maintain 🤷‍♂️ [22:09:03] would ya miss e if i was gone @dmehus [22:09:10] Stop asking people if they want to join SRE [22:09:12] i'm weird like that [22:09:21] Yes. Please stop. [22:09:27] Only if its being maintained correctly; which rushing to make changes can result in it being poorly maintained. [22:09:32] Cocopuff2018#6664, of course. You have a unique place in the Miraheze community :) [22:09:46] Yeah, um, I'm excellent at maintaining documentation imo [22:09:51] cookmeplox#0000, well said, and thank you for wanting to help as a consultant to SRE :) [22:10:28] Maintaining the documentation as its exists and recognizing when that documentation appropriately needs an immediate update are two different skills 🙂 [22:10:58] Maintaining is the first, maintaining isn't the second [22:11:22] Completely changing the documentation is definitely different [22:11:35] Anyway this is a tangent, let's let SRE folks get back to their stugg [22:11:45] Sure [22:13:18] i dont think ur ready to be a board member @bwm0 [22:14:04] What on Earth does this have to do with me updating the [[Tech:SRE Volunteers]] list, respectfully [22:14:04] https://meta.miraheze.org/wiki/Tech:SRE_Volunteers [22:14:11] [22:14:31] Like, literally nothing [22:14:46] I'm heading off for dinner [22:15:10] cool it all with your arguments over who should be in what roles. and the docs can sit for a day or two until we get things figured out [22:15:31] Cocopuff2018#6664 and bwm0#0000, I think it's probably best if you just disperse / drop the argument. The two of you are like oil and water :) [22:16:14] @cookmeplox mind if we steal your robots.txt? I'm thinking I want to work a good bit of that into what we've got [22:16:40] nah do whatever you want [22:17:00] although you'll probably be better off having it be responsive to the actual UAs that are hitting you [22:17:20] does graylog let you do something like list UAs by count? [22:17:24] Is that what proxmox [22:17:27] i've not used it before [22:17:35] <.labster#0000> Lots of the roles will change, so arguing about a specific hat is likely counterproductive. [22:17:38] _was scrolled back as fuck_ 🫠 [22:17:56] like the most useful thing is whatever the equivalent of `select useragent, COUNT(1) FROM ...` is [22:18:02] <.labster#0000> NDKilla is back, we’re all saved! [22:18:04] [1/2] also, for your consideration if you still want to look at nginx logs, we have two versions [22:18:04] [2/2] One is sent directly to graylog, and I don't think I'd be able get you a copy. The other is stored in the access.log files on the server (cp and mw), and I can possibly awk you a copy that doesn't include remote_addr or remote_user [22:18:38] hmm what is remote_user? [22:18:46] I know remote_addr but i'm not familiar with the other [22:18:53] dunno, it's empty in all the samples I've seen [22:19:12] "remote_user" obviously means a sentient TV remote with Internet access [22:19:27] but yeah, whatever is easiest for you [22:19:41] ideally i'd get all of the different servers separately [22:19:49] I'll DM a screenshot of what I've got (UAs without further information should be safe to share) [22:19:58] since sampling off one of the specific cps might bias [22:20:11] and getting nginx logs for mw* specifically is also useful to see how much varnish is shielding you [22:20:36] probably isn't since we don't seem to be blocking much [22:21:09] i guess i mean less blocking and more like...is able to serve from varnish [22:21:16] rather than having to go to the mw backend [22:21:41] true, but blocking might be helpful while we wait for robots.txt to propogate through [22:23:12] ah, looks like cp* access logs aren't in graylog [22:23:43] Hiya! [22:24:27] oh wow, that wasn't even cp, that was just mw? [22:24:33] so that's not even counting anything that varnish caches [22:24:50] yeah [22:43:40] Just to note: For wikis with Flow/Structured Discussions enabled, you can now generate dumps with that content [22:45:38] That is a new feature? How old is that? [22:46:24] The script already existed, but DataDump wasn't configured to use it. [22:48:03] That's great news for people with that kind of wikis [22:50:42] is miraheze shutting down or no? [22:51:17] great question, I wonder if there are any pinned messages answering it [22:52:07] if miraheze did shut down i wouldnt be soo sad tbhy [22:52:52] <.labster#0000> Should I just give up then? [22:53:08] up to you tbh [22:53:10] well, the only reason you were planning to save it was to make Coco happy, right? [22:53:17] and if Coco doesn't care, what's the point? [22:53:30] <.labster#0000> Pretty much, right? [22:53:41] 👀 [22:53:43] i care but if it shuts down i have no reason to be sad [22:53:49] [1/2] Miraheze is not shutting down article. [22:53:50] [2/2] I don't know the topic of this conversation [22:55:46] we will all end up at self hosted wikis [22:56:04] Please stop. [22:57:42] I think we need to rebrand Miraheze as M̃iraheze to show people we are serious about improving the reputation [22:57:52] It's like Miraheze except with a fake M̃ustache [22:58:06] w the hell you sexpect rebranding to fix things' [22:58:33] Why did you thistbh that post coco [22:58:35] https://tenor.com/view/jhb-team-missed-didnt-get-it-joke-gif-16924038 [22:58:41] It was aimed at you [22:59:08] im sorry to say i have soo little faith in this project at this time [22:59:18] ....and? [22:59:27] what does that have to do with either of these things [22:59:38] I for one have complete faith in Miraheze's ability to get through this. [22:59:47] but it also requires community support [23:00:01] <.labster#0000> Feel free to help or leave if you’re out of faith. We’ll be here when you come back. [23:01:01] ya know what i will put my faith into this project [23:01:13] @Cocopuff2018 why did you thistbh this as well.....it was aimed at you again [23:01:36] this [23:01:45] I....okay [23:11:17] [1/3] Cocopuff, I need you to stop making unnecessary dramatic statements and instigating unnecessary fights or leave this server. [23:11:17] [2/3] This has been a perennial issue well before the last few weeks on which you have gotten quite a bit of leeway, folks are trying to do real work to salvage the project. [23:11:18] [3/3] I suggest you take Labster's advice and consider a more constructive approach going forward should you choose to remain. [23:34:16] sorry about all this [23:43:50] Do want to ask, if this backup plan fails, when is the final month [23:44:00] Because I got mixed messages between September and October [23:44:22] Sometime in September [23:45:00] Damn [23:46:18] ^ this. Wikis will be read only from 31st July 2023, then at some point between 31 August 2023 and 1st October 2023, with seven days advance notice, Miraheze will scale back services to hosting XML and image dumps of wikis, which can be requested by bureaucrats and/or sysops [23:46:50] probably worth clarifying that the thing you're quoting is from before the new update [23:47:11] That's the latest official update from Miraheze Limited [23:47:22] and if the transition plan does not proceed [23:48:01] Well, I'm gonna need to find a way to data/image dump my stuff [23:48:07] just worth pointing out for anyone coming into the conversation that that is not what is expected to happen [23:48:09] Not sure how big of a file that'll be [23:48:18] because people are already misinterpreting what you said, like this guy [23:48:42] Yeah I am confused [23:48:55] So is Miraheze truly fallen or is it a 50/50 chance [23:49:01] @Corykun#0000, if it's small enough, you can generate them at [[Special:DataDump]]. Otherwise, [[Stewards' noticeboard]] or [[Phabricator]] [23:49:01] https://meta.miraheze.org/wiki/Special:DataDump https://meta.miraheze.org/wiki/Stewards%27_noticeboard https://meta.miraheze.org/wiki/Phabricator [23:49:03] So when is the next update of mediawiki [23:49:16] Corykun#0000, 50/50 chance is about accurate [23:49:16] What if its big [23:49:25] Like over 500 pages and over 500 images [23:49:28] I think 50/50 is badly underselling it [23:49:33] Corykun#0000, it should tell you, but then [[Phabricator]] [23:49:33] https://meta.miraheze.org/wiki/Phabricator [23:49:36] there is no immediate risk of service shutdown, and no cause to leave immediately. [23:49:46] It seems exceedingly likely that it will continue to exist for the foreseeable future [23:49:57] only Image dumps need to be requested [23:50:02] There are competent people running the transition [23:50:09] I'm gonna need to request the image dump then [23:50:10] for XML I would say try it yourself first [23:50:16] I got quite a lot [23:50:24] Cookmeplox#0000, yeah, though I don't feel anyone is misinterpreting what I said. It's good to be prudent and have a backup plan in place :) [23:50:38] I disagree fully [23:51:02] wwhes the next mediawii update? [23:51:03] You disagree it's not prudent to backup your wiki and have a contingency plan in place? [23:51:25] No i disagree there was misinterpretation [23:51:34] I'm just prepared for the worst is all personally. I have 5 years of pages and editing on the line [23:52:01] You should feel confident that regardless of how things proceed, your content will not be lost [23:52:02] As fun as it was here, if this place goes, I don't want my work to go down with it [23:52:07] It's not a BAD idea to keep your own personal copy [23:52:25] Of course but I don’t forsee Miraheze going down [23:52:30] Also a good idea to have back-ups of that. [23:52:31] But there would have to be about 5 different levels of cascading failures for the content and backups to actually get lost, even if the company and service ceases to exist [23:52:35] only issue is you cant transfer the images in a dump [23:52:36] Corykun#0000, that's fair and prudent. As MacFan4000 stated, you should be able to request an XML dump at [[Special:DataDump]], but if you run into issues, then file a task on Phabricator. For image dumps, those need to be requested on Phabricator [23:52:36] https://meta.miraheze.org/wiki/Special:DataDump [23:52:36] (which to be clear, is not likely) [23:52:50] I'll try and request it there then [23:52:50] Thats why we do image dumps [23:53:01] And where can I access phanricator again [23:53:10] [[phab:]] [23:53:10] https://meta.miraheze.org/wiki/phab: [23:53:11] https://phabricator.miraheze.org [23:53:11] [23:53:54] Thank you [23:55:41] np [23:59:35] When importing miraheze wiki to a non-miraheze wiki, anyone know what interwiki prefix to use