[06:02:12] hello. with WDQS, how do I select for something to be 'NOT'? I want to check for something that doesn't have a particular thing? [08:55:40] Josve05a: sorry for the spam, but at least I found the source of the problem. [09:29:38] Lydia_WMDE: Hi, I would love we have a meeting with the other Wikidata mentors. What do you think? So we can work on the various triggers needed. [09:30:43] I have sent you an email with the other mentors with the triggers i have come up with. I will still send more but I was thiking if the mentors can together validate and approve these triggers, then that will be fine. [09:49:25] d3r1ck: hey [09:49:27] yeah sounds good [09:49:29] sorry for not being very responsive [09:49:36] lots of things to do at the moment [09:49:43] but we should have a meeting with steven and maybe bene [10:32:15] Lydia_WMDE: Ok, great. But the problem here is time zones :) [10:32:31] Lydia_WMDE: Stephen is 9 hours behind you and 8 hours behind me :) [10:32:54] d3r1ck: yeah a call in the evening our time will have to do i guess [10:33:45] Ok, that sounds great. So when do we put the meeting? [10:35:55] Any suggestions? I am thinking 9pm. That will be 10pm in Berlin and 2pm for Stephen (pacific time) [10:54:31] how about 2 hours earliner? [10:54:51] 10 pm is definitely too late for me to properly think :D [10:56:24] Ok by 2 hours earlier you mean 8pm [10:56:33] Ok 8pm is fine by me. [10:57:08] That is 8pm my time, 7pm time in Berlin and 11am pacific time for Stephen. [10:57:12] ok can you suggest it in an email to stephen? [10:57:35] Yes i can do that. When are we having the meeting? Today or should we fix a day? [10:58:34] Also, should the meeting be IRC based or Hangout on Google or a Skype call? Which should we use? [10:59:13] think audio is better if possible so skype or hangout or something else [10:59:20] let me check for a day [10:59:48] the coming week tuesday and friday works for me [11:00:51] Hmmm, ok. I have a schedule for all my mentors except Sam Tarling. Can you help me his email? Let me also check if they are free on Friday or Tuesday so we can schedule the meeting. [11:01:11] i don't know him and don't have an email for him [11:01:32] i am not aware of any involvement by him in wikidata so let's concentrate on getting a call with stephen and us set up [11:01:34] for now [11:01:54] ok [11:02:06] Weekly schedule for the other mentors are [11:02:08] Bene * - Fri, Sat, Sun, Mon. [11:02:08] Slaporte - Thurs, Mon[next] (10pm - 11pm) [11:02:09] Hoo - Everyday (Code Review + Emails) [11:03:19] So, which this, I can plead with Stephen to borrow me 2 hours of his time on Friday. [11:03:25] :) [11:03:39] which -> *with [11:03:55] cool [11:04:02] I think we can have a meeting on Friday. Friday can work [11:04:19] :) [11:04:37] One more thing, how long will the meeting take. I am taking everything into consideration :) [11:04:37] Lydia_WMDE: I'm about to send an email that makes you very happy [11:05:06] d3r1ck: depends let's go for 1 hour and max 1.5h? [11:05:10] Amir1: :D [11:05:14] i like those! [11:06:04] Lydia_WMDE: Ok, 1hr - 1.5hr. [11:06:29] I am still thinking the best Audio means we can use to have the meeting. [11:06:50] I think skype is the best since it suites its purpose (VoIP) :) [11:06:53] what do you think? [11:10:09] d3r1ck: i don't mind to be honest. whatever works for the others [11:10:11] and you [11:12:05] Lydia_WMDE: Ok. I will send an email to Stephen right now. When he replies after accepting to borrow me 1.5hrs on friday, then I will write to all the mentors. [11:12:14] perfect :) [11:12:18] Informing them about the meeting :) [11:14:19] Lydia_WMDE: Sent [11:14:28] looking [11:14:29] Sorry it's sooooo long [11:14:58] not arrived yet [11:15:09] ah there [11:15:10] i also love those mails [11:15:30] it is more accurate yes [11:15:48] i doesn't mark that annoying IPv6 user as harming anymore [11:16:42] Amir1: wohoooooooooooooo! :D [11:16:46] :) [11:16:50] sjoerddebruin: indeed :) [11:16:57] \o/ [11:17:11] <- does a little happy dance [11:17:13] :P [11:17:32] Actually we did it yesterday, but I passed out afterwards [11:18:17] Lydia_WMDE: Can you please pass the information when you meet Hoo and Bene* tomorrow at the office? :) [11:18:21] They are not online now [11:18:40] d3r1ck: neither of them work in the office - they both work from home and only come to the office sometimes [11:18:48] so i won't see them tomorrow [11:18:56] Ok [11:19:10] Then I will inform them today when they come online [11:19:23] cool [11:19:23] I have emailed Stephen, I will be waiting for his response. [11:21:23] Lydia_WMDE: you might want to check this out :) https://www.mediawiki.org/wiki/User:Alangi_derick/IFTTT_GSoC_Report_2016 [11:21:43] will do :) [11:21:53] ok [11:41:40] sjoerddebruin: I ran some stats on who labeled the most [11:41:45] you were one of them [11:41:47] thanks :) [11:41:55] nice :) [11:55:30] hmm, it seems this entity is in a conflicted state about 'what' it is: [11:55:31] https://www.wikidata.org/w/index.php?title=Q4585033 [11:55:36] a place or a movie... [11:57:20] revert to 28 nov, thedj [12:18:27] how do i get an entity deleted ? [12:18:32] https://www.wikidata.org/wiki/Q23888349 [12:18:41] vanity entry [12:19:31] thedj: the problem is that it still contains a "valid" sitelink [12:19:34] ah of course i should have known that there is a WD:DELETE shortcut :) [12:20:34] sjoerddebruin: which too is up for deletion [12:21:00] thedj: yes, but most request will be stalled upon deletion on the other project... [12:21:27] how to distinguish "instance of " from "sub-class" proprieties ? [16:02:28] meh what FUD https://en.wikipedia.org/w/index.php?title=Template_talk:Infobox_software&diff=708571215&oldid=708508469 [16:06:28] Nemo_bis: nice references are possible with LUA... [16:08:13] sjoerddebruin: yeah that's a typical example of "my article are just perfect now nobody touch them" [16:10:09] yep, I was pretty annoyed by this answer too [16:30:11] MichaelSchoenitz: ah there you are :) I arrived to that talk page while trying to understand why people still have to do things like https://en.wikipedia.org/wiki/Special:Contributions/Tortipouss [16:32:19] jup, migration of software-versions to wikidata would make live much easier. [16:32:57] I'm glad to see that in German wikipedia it's slowly starting [16:36:58] and in German WP we have sighted versions – so all edits upgrading the version-number wich are made by IPs or new users have to bee reviewed by someone with the right to sight. I'm reviewing ~20 edits containing nothing else than version numbers per day. [16:44:39] MichaelSchoenitz: sssshhh, that additional review work is a feature for some people [16:45:16] I think one could gradually just eliminate all the {{LSR}} subpages on en.wiki without anyone noticing [16:45:17] 10[1] 04https://www.wikidata.org/wiki/Template:LSR [16:45:54] (checking that the Wikidata items actually contain the latest release of course) [17:11:59] Do you think that deadline for queries should be extended? [17:12:02] https://query.wikidata.org/#%23Whose%20birthday%20is%20today%3F%0ASELECT%20%3Fentity%20%28year%28%3Fdate%29%20as%20%3Fyear%29%20%0AWHERE%20%0A{%09%0A%09%3FentityS%20wdt%3AP569%20%3Fdate%20.%09%0A%09SERVICE%20wikibase%3Alabel%20{%0A%09%09bd%3AserviceParam [17:12:06] %20wikibase%3Alanguage%20%22en%22%20.%0A%09%09%3FentityS%20rdfs%3Alabel%20%3Fentity%0A%09}%20%0A%09FILTER%20%28datatype%28%3Fdate%29%20%3D%20xsd%3AdateTime%29%09%0A%09FILTER%20%28month%28%3Fdate%29%20%3D%20month%28now%28%29%29%29%0A%09FILTER%20%28day%28% [17:12:10] 3Fdate%29%20%3D%20day%28now%28%29%29%29%0A}%0ALIMIT%2010 [17:12:34] Too long URL :) [17:13:14] Anyway, the query "Whose birthday is today?" will never be completed [19:34:19] PROBLEM - WDQS SPARQL on wdqs1001 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 416 bytes in 0.043 second response time [19:35:08] PROBLEM - WDQS HTTP on wdqs1001 is CRITICAL: HTTP CRITICAL: HTTP/1.1 503 Service Temporarily Unavailable - 416 bytes in 0.006 second response time [21:11:49] SMalyshev: are you reloading with the april 11 dump? [21:14:34] aude: already mostly done, yes from April 11 [21:14:49] I've reloaded one server and now doing the second one [21:15:29] which is live? [21:15:33] wdqs1002 [21:15:54] it's still about 1.5 days behind but I assume will catch up by the end of the day [21:16:16] is the coordinate order switch supposed to be live? [21:16:40] aude: yes [21:16:44] ok [21:16:52] then i'll make a patch to fix the map [21:17:18] SMalyshev: Are you making a task on how to prevent/mitigate this in the future? [21:17:42] This corrupted data is quite disruptive now already and the impact will only get bigger over time [21:18:16] multichill: I know. The problem is it was kind of bad conincidence of dumps going bad and us reloading and not discovering it. [21:18:37] I guess we need somehow to watch the dump logs for exceptions and alert if there's some [21:19:00] Or have some basic sanity checks like: Is this dump larger than the previous one? [21:19:20] multichill: well, strictly speaking it doesn't have to be larger. especially with compression involved [21:19:54] Can you give me one example where the Wikidata dump became smaller and it was correct? ;-) [21:20:15] it just doesn't have to be like 1G smaller. The problem is if it were half as big, I'd notice something. But it was only slightly smaller, so I didn't think it was something weird [21:20:46] Dumps will always be bigger unless something changed [21:20:58] multichill: not currently, but I don't know sizes of all dumps now and in the futire :) I think checking for exceptions is much better [21:21:14] that way we'd have caught it way earlier [21:21:35] we don't need dumps too often, it's just a bad coincidence that dumps gone bad while we needed to reload everything [21:21:39] SMalyshev: You know you're going to get a "I told you so" if this ever happens again? :P [21:21:56] this wouldn't happen again. Something else will :) [21:22:45] Are the wdqs* servers vm's or real hardware? [21:23:04] And how much disk space are we talking about? [21:24:01] I would normally take a snapshot of a VM before major changes. Do everything, test it, and have the snapshot age in the normal backup strategy [21:24:04] multichill: hardware, VMs can't handle it. The DB is slightly over 100G now [21:25:26] 100GB, that's not too much. So you could keep the old db around when you load the new one? [21:26:28] multichill: making copy of the DB is possible, but given the servers were reimaged, you need somewhere to copy and hold at least 30-40G (DB is somewhat compressable). It is not a trivial task to find such place. And usually just having 2 servers is enough. The problem in this case was we didn't know the dump is bad until we've done both servers. If we knew it, we'd just left [21:26:29] the second one alone and fixed the first, but we learned about it when we already reloaded both servers. [21:28:31] SMalyshev: I'm just used to always having a rollback option when doing a change. You basically killed your rollback option when you started on the second server. [21:28:58] Not having a rollback sounds like a risk to me I would want to mitigate [21:29:58] multichill: right, because we thought everything is fine. I'm still not sure how we could detected the dump is bad (except the size, but the dump could be same size and still bad) [21:33:28] I get that, but that's just a source of problems. You would alway import the dump to each server one, by one. Now we have two servers, maybe we'll have 5 in a couple of years. Not being able to go back to a previous working state is scary [21:35:05] nn [21:41:40] well, before this reimaging we didn't have the diskspace to keep the copy locally. Now we do on one server, other one will be upgraded soon. So we could rollback at least when we're not reimaging [21:46:14] I've created T134131 to watch the dumps