[07:59:00] <joakino>	 morning
[08:04:09] <phuedx>	 o/
[08:04:15] <phuedx>	 hey joakino
[08:04:43] <joakino>	 hi phuedx!
[08:04:49] <joakino>	 i'm breakfasting
[11:09:55] <phuedx>	 out for lunch
[11:10:03] <phuedx>	 joakino: jealous of your internet connection right now
[11:10:10] <phuedx>	 i'm on 55/10 right now
[11:11:47] <joakino>	 that's good enough
[11:12:00] <joakino>	 from the 300 i loose 220 over wifi lol
[11:51:15] <joakino>	 out for lunch
[14:56:45] <hashar>	 125 MBytes/sec / 25 MBytes/sec
[14:56:48] <hashar>	 but that is really useless
[14:57:21] <hashar>	 wireless does not even reach that and there is most probably an Ethernet port (~10Mbytes/sec) on the way
[14:57:26] <hashar>	 optical fiber is overrated
[14:58:05] <hashar>	 I think the use case is being able to max out wifi bandwidth on a few devices and be able to watch 4K videos on a couple TV set
[15:52:50] <phuedx>	 ^ that
[16:00:37] <joakino>	 and downloading bluray movies in 4 minutes too
[16:00:43] <joakino>	 but yeah agreed
[16:01:01] <joakino>	 my options were either 30/3 or 300/30
[17:30:29] <bmansurov>	 phuedx: hey, about keeping the module
[17:30:38] <bmansurov>	 phuedx: please see the email titled [Reading-web-team] Pushing to production / reverting / lessons learned 
[17:30:54] <bmansurov>	 phuedx: that was the reasoning, but i'm happy to change the patch if you think otherwise
[17:31:07] <joakino>	 i'm off for the day, see you tomorrow!
[17:31:53] <bmansurov>	 later joakino
[17:34:44] <jhobs>	 cya joaquin
[18:15:57] <phuedx>	 bmansurov: ah i had to re-read roan's messages
[18:16:10] <bmansurov>	 kool
[18:16:17] <phuedx>	 so if the module hash doesn't change bad foo happens?
[18:16:36] <bmansurov>	 something like that
[18:16:40] <phuedx>	 assuming that we're reverting in between train deploys (i.e. SWAT)
[18:16:45] <bmansurov>	 yes
[18:17:03] <bmansurov>	 i read it as something something don't remove files something something
[18:18:31] <phuedx>	 bmansurov: i've asked for clarification on the wiki page for that reason :)
[18:18:45] <bmansurov>	 ok
[19:10:35] <phuedx>	 g'night y'all
[19:10:45] <phuedx>	 (i am full of pizza)
[19:37:52] <bmansurov>	 later, phuedx
[20:21:49] <dr0ptp4kt>	 mdholloway: bearND coreyfloyd i moved our regular meeting to tomorrow
[20:22:17] <mdholloway>	 dr0ptp4kt: cool
[20:23:40] <bearND>	 dr0ptp4kt: sounds good
[20:33:06] <mhurd>	 niedzielski: thx again for the twn updates! :) it was a bit weird seeing my name in the headers for some of the languages this go-round, but i guess removing unsupported markup from a couple strings counts as an edit
[21:02:09] <Pchelolo>	 mdholloway: re https://phabricator.wikimedia.org/T131554 the article of the day in mobile app is parsed from the Main_Page, right?
[21:02:20] <mdholloway>	 Pchelolo: yep
[21:02:32] <Pchelolo>	 mdholloway: okey, lemme investigate
[21:12:31] <mdholloway>	 Pchelolo: Seems like another redirect-related issue: https://de.wikipedia.org/api/rest_v1/page/mobile-sections/Wikipedia:Hauptseite gives the old content, but https://de.wikipedia.org/api/rest_v1/page/mobile-sections/Main_Page provides the current content
[21:13:43] <Pchelolo>	 mdholloway: no, I don't think that's the issue. https://de.wikipedia.org/api/rest_v1/page/html/Wikipedia:Hauptseite gives new content, so the update reached RESTBase
[21:14:18] <Pchelolo>	 but got lost somewhere in dependency rerendering. 
[21:26:00] <Pchelolo>	 mdholloway: we've fixed the problem for now by manually making a rerender, will investigate what caused it
[21:26:17] <mdholloway>	 Pchelolo: sounds good, thank you!
[21:46:16] <Pchelolo>	 mdholloway: okey, we found the reason... Remember we've limited the mobile pregeneration to main namespace only? 
[21:46:38] <mdholloway>	 Pchelolo: aha, yes i do
[21:47:02] <Pchelolo>	 mdholloway: Wikipedia:Hauptseite is in the project namespace..
[21:47:09] <mdholloway>	 Pchelolo: right.
[21:47:11] <Pchelolo>	 so the update was simply skipped
[21:47:19] <Pchelolo>	 will fix soon
[21:48:13] <mdholloway>	 Pchelolo: great, thanks again. 
[21:49:54] <Pchelolo>	 mdholloway: which leads us to a question, which namespaces should we pregenerate? Main and Project are a given. File?
[21:53:31] <mdholloway>	 Pchelolo: I would think just Main and Project.  In the back of my mind, I'm wondering about the future of the beta definitions feature, which was the reason for the restriction in the first place.  But that's more of a product question.
[21:54:00] <mdholloway>	 bearND: do you think any other namespaces should be pregenerated?
[21:55:12] <Pchelolo>	 mdholloway: okey, I'll create a PR to make Project pregenerated. This issue affects some big wikis like french and german, I'll run a script to quick-fix it for now until next restbase deploy
[21:56:26] <mdholloway>	 Pchelolo: Yes, especially since we're now up to 50% of users using content service loading in the stable app.  Thanks again for your help on this.  
[21:56:52] <bearND>	 Pchelolo: mdholloway: main and project sound good. File might be good as well
[22:01:10] <gwicke>	 mdholloway: we were just wondering if it's worth excluding any namespaces for projects other than wiktionary
[22:02:44] <gwicke>	 the main namespace that I could see contributing some load would be User:, from dashboard-like pages
[22:03:06] <gwicke>	 pretty sure that most others are low volume
[22:04:03] <mdholloway>	 gwicke: bearND: Pchelolo: That makes sense.  In general I'd say it would be best to apply the namespace restriction only for wiktionary, and pregenerate most or all namespaces for the others.  
[22:04:59] <mdholloway>	 gwicke: I know mobrovac has some strong feelings about getting into too much special-casing based on project/namespace :)
[22:05:04] <gwicke>	 all would be easiest for us, as that would also avoid these cache update issues
[22:06:09] <Pchelolo>	 I've run a script across all wikipedia sites to update mobile content for main pages
[22:06:10] <mdholloway>	 gwicke: Yes, I'd vote for pregenerating all.  
[22:06:26] <bearND>	 gwicke: do i understand this correctly that if it isn't in the pre-generation list then it won't get timely updates?
[22:06:27] <mdholloway>	 gwicke: except for wiktionary.  
[22:06:54] <gwicke>	 bearND: yes, the way it's currently implemented those two are coupled
[22:06:57] <mdholloway>	 bearND: gwicke: I was a bit confused about that myself -- why not being pre-generated implies updates wouldn't get picked up.
[22:07:15] <bearND>	 gwicke: that's just a temporary bug, though,right?
[22:07:39] <gwicke>	 they don't need to be coupled, but we'd have to implement special purge behavior to avoid pre-generating
[22:08:26] <gwicke>	 bearND: do you think pre-generating all namespaces would cause too much load?
[22:08:51] <gwicke>	 in terms of requests, the difference should be fairly small
[22:09:53] <bearND>	 gwicke: you would know better than me since I don't have too much insight into even how long pre-generation runs
[22:10:49] <bearND>	 when some major (hypothetical) bug in MCS get fixed we might ask for a fresh pre-generation run again
[22:11:00] <mdholloway>	 gwicke: bearND: Other namespaces are very seldom requested in the app.  After the pregeneration pass I'd expect the impact to be minimal
[22:11:10] <mdholloway>	 gwicke: bearND: And after all that was the original behavior, right?
[22:11:33] <gwicke>	 yes, I don't think things were falling over from load before
[22:12:22] <gwicke>	 so, my proposal is to go with all for now, and then look into possibly moving to purge-only for some namespaces later
[22:12:54] <gwicke>	 rather than just skipping updates to some namespaces
[22:13:11] <mdholloway>	 gwicke: Except for wiktionary, right?
[22:13:15] <gwicke>	 yes
[22:13:25] <mdholloway>	 gwicke: that sounds good to me
[22:16:24] <bearND>	 sounds good
[22:18:18] <gwicke>	 great, thanks!
[23:10:19] <coreyfloyd>	 brion: Hey you got a few minutes… need some guidance on hosting a file on multiple domains