[00:14:33] I'd be amazed anyone dates to use such lol 😆 [00:16:34] There is always ASCII art, if you need images without images 😄 [00:21:45] I hope Cargo is not being forgotten due to this new issue [00:22:19] Cargo is still the second thing on my priority list [00:23:45] Although, for where it is now, someone from the MediaWiki engineering team could take over the next two steps [00:24:41] Does Miraheze periodically run `maintenance/purgeParserCache.php`? I'm seeing parser cache (what was keeping our Cargo-reliant wiki partially working) disappearing despite $wgReadOnly [00:25:03] normal requests can't delete things from CACHE_DB when $wgReadOnly is on [00:25:58] ?action=purge and alike won't work in "read only" mode either, so it can't be caused by that [00:26:56] could be the varnish cache expiring, not really sure [00:27:27] as long as the page is in parser cache, it'll show normally even without Cargo [00:27:40] it's being cleaned by something [00:29:35] it seems we do run it [00:29:41] It invalidates the cache every 10 days [00:31:22] well, nothing we can do about it now, it's cleaned already [00:56:26] [1/2] > From the looks of things, swiftac is going to need to be rebuilt. Some of the data can be scavenged, however much of the VM was corrupted beyond the point of usability. I also believe that proxmox will need to be reinstalled on cloud11, as the previous install is likely corrupted in a similar manner to swiftac. [00:56:27] [2/2] how long does a rebuild take [00:56:27] I'd definitely think this current issue puts cargo on the backburner until it resolves [00:58:12] Does swiftac include image data or is it for everything else? Asking on behalf of AVID staff [00:59:27] swiftac contains metadata, not images [00:59:38] This metadata routes requests to the right servers [01:04:29] An issue affecting all wikis would need to take precedence over an issue that doesn't. [01:04:38] Without swiftac, the data on the object servers that hosts contains the images is inaccessible. [01:04:54] Particularly if it's a significant issue. [01:05:31] We're working on recovering any data we can from the borked partition [01:05:56] Is there archive for real life villains wiki? [01:06:02] There is not [01:06:30] Link? [01:06:48] to what [01:06:51] he said there is not [01:07:19] OK. [01:09:02] At least, that'd be my guess. [01:09:11] It would, yes [01:09:24] Well, also, no patch for Cargo has been published so there's that [01:11:31] Unfortunately, that's often the case with Cargo, just from previous observation. I think the author is quite pressed for time. [01:18:10] How about even evil has standards page? [01:19:42] I'd be comfortable running cargo with my changes to database access. Most of what it needs though is a review of the config, a merge of the changes, and then moving all the cargo tables to their own database. [01:24:06] How about even evil has standards page? [01:28:54] At the very least, we'll hopefully be able to start reinstalling cloud11 by tomorrow [01:32:22] just to check how long does it take before the smw import for my wiki is done? an eta would be great [01:33:00] How long have you seen that error for? [01:33:16] uhhh it got approved yesterday or so [01:33:47] 31ish hours ago [01:33:50] You shouldn't see that error for more than 5 minutes max [01:33:55] well! [01:33:56] someone forgot to run a script [01:34:00] cc @Reception123 ^ [01:34:43] 🙃 [01:58:54] Late but, small note since there was mass panic over a metadata loss in another community I'm in a few months ago due to the layman not knowing what metadata is-- [01:59:11] Metadata is data about the data, not the data itself [02:00:43] Think like a tag on a shirt. If you lose the tag, you still have the shirt [02:04:34] That's an extremely good description. It's a list of properties the important object has, not the important object itself. [02:43:39]

How much is the premium and is it permanent? [02:48:16] It's a proposal so TBD [03:10:57] I'm the sole bureaucrat and sole user of a wiki I created. I go to Special:Upload and I cannot upload files. How do I give myself the permissions to upload files? What is even going on? [03:10:59] looks like we have consensus for the RfS [03:11:09] ooop [03:11:20] ransElectronics: we're currently encountering technical issues [03:11:24] all files are down right now [03:11:44] this will provide a better explanation than I can type haha: https://meta.miraheze.org/wiki/Tech:SREnoticeboard#Cloud11andSwiftissues [03:12:12] To clarify: files themselves are not lost, but there's a hiccup on the piece of hardware that's responsible for authenticating users and connecting requests to files that SRE is working to fix. [03:12:52] what NotAracham said [03:13:06] well the link you provided seems to not be working, but thanks for the info. I'm not crazy [03:13:38] by the way apologies to all if I come off as vague or confusing, been a stressful past few hours so I'm a bit stretched as is [03:13:51] They can not be authenticated without swiftac. It is completely not able to build or access object storage. swiftac is important, if rings are lost things become more difficult, and very hard tk rebuild to access from my understanding. [03:13:53] ransElectronics: are you by any chance encountering a 502 or 503? [03:14:08] no it just straight up says I don't have the permissions to upload [03:14:21] Yep, that's to be expected at present [03:14:21] ...on the SRE noticeboard? [03:14:46] I think hexchat mangled the URL you linked, I put a space into it in the URL bar on firefox and it went through [03:14:59] ah that's why [03:15:29] It is because of the relay bot converting underscores to italics IRC side, I'll fix later. [03:15:49] ah [03:18:27] @Orduin by the way (see you're typing haha) this is eligible to be closed right? [03:18:31] I believe rings are stored in puppet private, the problem is swiftac is the container server and user server [03:18:48] TBH, probably the lowest thing on my priority list rn [03:19:01] of course [03:19:18] let's get cloud11 fixed first [03:19:19] If rings are safe, then it should be able to recover I think? [03:19:38] Hmm, maybe, but we would be missing the database of containers [03:20:03] Can't that be rebuilt? [03:20:17] No idea, I've never worked with swift before [03:20:50] out of curiosity, what prompted the change from gluster to swift? [03:21:01] Performance [03:21:31] ah [03:22:16] Object storage is a lot different than file storage. [03:22:41] When gluster gets overwhelmed, wikis go down. When Swift gets overwhelmed (or a butler server goes offline like this), only images don't load. [03:22:57] (Though calling it a butler server's absolutely an understatement of its duties) [03:23:46] If this happened with Gluster, we would be 100% down, unless things were unmounted during the outage. [04:25:40] For mw configs, where can we find the default values? Curious to know what the values of $smwgQMaxInlineLimit, $smwStrictComparators, $smwgDVFeatures are [04:26:24] https://github.com/SemanticMediaWiki/SemanticMediaWiki/blob/master/extension.json [04:27:12] thanks! [04:27:32] oh actually no, these settings aren't registered that like they normally are for other extensions [04:27:33] Wait, I mean on miraheze [04:27:42] it's the extension default [04:27:46] we don't change the defaults [04:27:53] Ah ok [05:07:25] About fandom c'est une toilette. [05:23:46] If it was, you probably wouldn't be banned [05:28:00] True [06:22:49] please post FANDOM (which is bad, as we all here know) complaints in #offtopic [08:07:40] What ? [09:47:50] https://meta.miraheze.org/wiki/RequestsforComment/Premiumofferings(in-house_proposal) [09:48:47] I have created a parallel proposal to Premium Offerings, with the aim of it being hosted and developed within the Miraheze organization itself. As per the [[Request for Comments]] page, this is allowed. Please take the time to review the alternative proposal. Thanks very much. [09:48:47] https://meta.miraheze.org/wiki/Request_for_Comments [09:48:48]

[09:49:06] [[Request for Comment]] [09:49:06] https://meta.miraheze.org/wiki/Request_for_Comment [09:49:07]

[09:49:35] [1/2] well I don't know what this bot's problem is [09:49:35] [2/2] https://meta.miraheze.org/wiki/RequestsforComment [09:50:30] You forgot to add the "s" onto "Requests" [09:55:27] Ah 😆 well, human fallibity [09:56:07] It do be like that [10:02:37] @Nemo I don’t think it’s worth running 2 at the same time [10:02:56] Make your arguments and if the discussion is supportive of it that can can go later [10:03:13] Running too much at once causes issues [10:13:21] Im "asking for comments". The comment section in the existing proposal is for commenting the existing proposal. However, my proposal is comopletely different. i cannot post the whole proposal in a comment to the other proposal, nor it can be discussed there. Unless you authorize me to do so? [10:14:02] @Nemo you can make the suggestions in the comments section [10:14:03] I think this is rather more conductive than just putting the whole proposal in the other page [10:14:14] You don’t need a full proposal [10:14:26] Why? [10:14:58] @Nemo because that will let you know whether it’s needed [10:15:08] And it keeps opinions in both places [10:15:18] The proposals can’t both happen [10:15:53] If you post a comment saying would it better to just launch Miraheze premium and everyone says no [10:16:08] That answers whether it needs a seperate RfC if this fails [10:16:33] I said it and nobody said "no", furthermore several people had the same complaints as I did (the external reliance) [10:16:57] If my proposal can be listed in parallel with the existing one, the separate page can be taken down, until then, the separate page is needed [10:17:52] In other RfC, the alternative proposals were listed immediate after the original one, but doing such an edit myself would be overstepping [10:19:56] As the "notes" part imply, I'm in favor of merging my proposal as an alternative presented in the original page so that they can be voted and discussed together. [10:22:24] It should be listed on 1 page [10:22:52] You can add an alternatives subheader [10:23:04] I caution against too much information though [10:23:38] The consideration of the RfC is simply whether the board should consider and draft a formal proposal [10:23:58] Large RfCs have generally performed poorly and been slow to resolve [10:28:17] [1/3] My alternative proposal is much less text-heavy than the original one, even including the comparisons. Without the comparisons, it would be like 1/4th the length [10:28:17] [2/3] I welcome any meta admin to merge it in full, but I will not put it in "under one paragraph" or such, since that damages the whole argument. The community should have access to the most complete information possible in order to make a decision. [10:28:17] [3/3] Compare https://meta.miraheze.org/wiki/RequestsforComment/ChangestotheDormancyPolicy_(2) [10:28:18] <⭐Eri#6583> omg somehow a singular image is loading on my wiki [10:28:27] which has eight different proposals, not counting variations [10:29:02] My proposal would be one such alternative [10:31:45] @Agent @CosmicAlpha [10:44:04] is there an issue with data dump? it says failed when i try to make a backup [10:44:41] datadump doesn't work because it actually stores dumps on the same system used to store uploaded files (Swift) [10:44:52] oh i see [11:11:54] @Nemo WikiForge isn't an external organization [11:12:01] It's a project but it isn't incorproated [11:12:51] It would "merge" into Miraheze in the sense that the community would give it's okay for the project to be built under the oversight of Miraheze Ltd. [11:13:10] so just like your proposal, the original proposal would bring WF into Miraheze Ltd. [11:14:55] but as I amended the original proposal, any new project should be a subsidiary company in order to ensure financial independence from Miraheze so that any ail afflicting it only affects it and not MH proper [11:18:35] That's an interesting point. In this way, as you state it, it seems that the only way to ensure that a payment system does not harm Miraheze is to establish it as an independent project. [11:19:07] That's the intent, yes, so that Miraheze proper isn't harmed [11:20:03] The idea that Nemo / NimoStar suggests is that people could pay for that system in the shape of a "recurring donation to MH" which sounds like something that would be indistinguishable from if they were donating directly [12:39:29] is it from wikimedia? [12:43:46] <⭐Eri#6583> I don't think so [12:44:14] <⭐Eri#6583> also its only loading on the page its on, if i go on the file page itself it doesn't load [13:10:50] iirc, that is not an actual image but something more alike to an emoji [13:11:18] it is like some sort of letter character, is it not? [13:12:23] (or it is an image file of that symbol perhaps, idk, it is a screenshot after all) [13:12:53] It's probably added via styling, which can take the image outside of miraheze via an url [13:13:26] Or it's from mediawiki [13:16:24] <⭐Eri#6583> ah yeah i think its through styling [13:16:48] <⭐Eri#6583> it was like one of the first things i did on my wiki so i kinda forgot what i did lmao [13:24:19] Regraxc to match all content within a template? [13:24:41] {{.*?}}) did not work [13:24:41] https://meta.miraheze.org/wiki/Template:.*%3f [13:24:41]

[13:29:15] An image using by Wiki has been not displayed, how does an image come to be displayed again? [13:29:42] <⭐Eri#6583, replying to nisekotu#0178> images are currently down unfortunately [13:30:30] It's been 3 days since they were down. [13:32:56] As the others have said, images have been down for the last 3 days across all Miraheze wikis. If I am informed correctly, there's been an issue with an SSD on the Cloud11 server which has caused a glitch with Swift, making all files inaccessible. Please correct me if I'm wrong [13:33:38] I think that's what one of them said. Agent also had me reassured that the issue would be fixed soon. [13:34:43] Yeah, good on the Miraheze team for jumping on it and actually working to fix the problem [13:35:04] Yeah, they're not incompetent compared to FANDOM. [13:35:35] They know how to do their job, I can guarantee that. [13:36:05] You think Fandom's incompetent? Should've seen wikifoundry while they were still up [13:36:32] Isn't Wikifoundry practically obscure now? [13:37:10] [1/2] Literally shut down [13:37:10] [2/2] Not long after the wiki I'm with (CLG, now AVID) left for Miraheze lmao. [13:37:21] It cannot be repeated enough. [13:37:37] We were like, the only good thing left about Wikifoundry [13:37:44] Hi everyone Miraheze Community! [13:37:52] How are you today? [13:38:33] Good. Just waiting for the issues to be fixed. Meanwhile, I can just edit my own wiki without uploading an image at the moment. [13:38:40] We're all at least a tad bit stressed about the issues with Swift, but other than that things are pretty good [13:39:27] Not many people are active on AVID atm, mostly just staff and one guy who thinks that being blatantly self-centred isn't gonna hinder his chances of being promoted [13:39:39] We can't do without the images. In any case, I can't. [13:41:16] AVID's lucky that it's got its YT embeds right now. If not, we'd be screwed [13:55:02] Might be an image from Wikimedia Commons, they still load. [13:55:41] Yeah anything hosted externally will still load [14:01:50] Start up: start replacing images with clouded ones [14:13:01] AVID staff are doing a data dump as soon as the situation is resolved. I'd advise other wiki staff teams to do the same [14:21:51] Image backups aren't available through DataDump [14:29:29] Pretty sure there's already plans in place to deal with this [15:17:56] Goodnight! [15:36:41] Hello! I know there is no ETA or anything for the image situation but some advice could be helpful - if I were to make an announcement for all the users of our wiki (to share various non-wiki resources they can still use), how likely do you think it is to be almost completely useless because the issue would be fixed shortly after? [15:56:28] From what I heard, the issue is nowhere near being fixed...so very unlikely [15:58:24] I wouldn't say it's no where near [15:58:27] Yeah it's slow progress afaik, the announcement will be fine [15:58:49] Void is trying to extract the borked contents of the partition so we'll have to wait and see [15:59:03] Some data does exist, we just don't know which [16:00:54] Ahh, I did hear some contrasting points of view so I wasn't sure. Thanks for clarifying, I suppose the announcement is more likely to be helpful than not. [16:04:01] What's the case with the images? [16:05:21] The server which routes image requests to the proper place doesn't work [16:05:24] so that's being worked on [16:05:44] Images themselves are technically fine but without the server to direct requests, they can't be accessed [16:05:54] How did this even happen? Out of curiosity [16:06:36] The SSD that hosted the server died randomly [16:06:45]