[01:43:07] Coren: Ping [01:44:21] or any Labs dev [01:56:38] Coren: Ping [01:56:40] or any labs dev [03:09:55] Any labs devs around? [05:12:57] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 0 minutes) [05:26:26] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 13 minutes) [05:39:59] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 27 minutes) [05:53:19] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 40 minutes) [06:06:43] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 53 minutes) [06:20:11] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 67 minutes) [06:33:36] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 80 minutes) [06:47:06] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 94 minutes) [07:00:35] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 107 minutes) [07:14:04] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 121 minutes) [07:27:32] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 134 minutes) [07:41:02] Warning: There is 1 user waiting for shell: Wangxuan8331800 (waiting 148 minutes) [07:54:28] Warning: There are 2 users waiting for shell, displaying last 2: Wangxuan8331800 (waiting 161 minutes) Muthumani2 (waiting 8 minutes) [08:07:57] Warning: There are 2 users waiting for shell, displaying last 2: Wangxuan8331800 (waiting 175 minutes) Muthumani2 (waiting 21 minutes) [08:21:18] Warning: There are 2 users waiting for shell, displaying last 2: Wangxuan8331800 (waiting 188 minutes) Muthumani2 (waiting 35 minutes) [08:34:47] Warning: There are 2 users waiting for shell, displaying last 2: Wangxuan8331800 (waiting 202 minutes) Muthumani2 (waiting 48 minutes) [08:48:16] Warning: There are 2 users waiting for shell, displaying last 2: Wangxuan8331800 (waiting 215 minutes) Muthumani2 (waiting 62 minutes) [09:01:45] Warning: There are 2 users waiting for shell, displaying last 2: Wangxuan8331800 (waiting 228 minutes) Muthumani2 (waiting 75 minutes) [09:09:03] !rq Muthumani2 [09:09:03] https://wikitech.wikimedia.org/wiki/Shell_Request/Muthumani2?action=edit https://wikitech.wikimedia.org/wiki/User_talk:Muthumani2?action=edit§ion=new&preload=Template:ShellGranted https://wikitech.wikimedia.org/wiki/Special:UserRights/Muthumani2 [09:34:08] /msg ChanServ DROP ##otr 634bf508:eea2518c [09:34:27] o.O [09:34:33] Hi Petan [09:34:37] heya [09:34:51] Any chance of migrating Catscan to Labs [09:34:57] I use it a LOT [09:36:21] why not [09:36:24] do that? :P [09:38:09] I lack the technical comptecne [09:38:55] Qcoder00: who runs catscan? [09:39:22] Magnus Manske [09:39:43] Whose also involed in the Migration process already [09:40:35] oh he should get to it soon I suppose [09:40:39] IIRC he loves playing with new tech? [09:43:03] !log tools petrb: syncing packages on exec nodes to avoid troubles with missing libs on some etc [09:43:08] Logged the message, Master [10:38:38] valhallasw: ! [10:42:17] addshore: see -bag [11:17:32] addshore: 'sup [11:21:13] :> [11:21:19] free much today? :P [11:21:55] no, not really :-) [11:22:04] :< [11:22:06] oh well :P [11:22:11] but I'm available for short questions :-) [11:30:05] valhallasw: was thinking of getting the dump scanner working :P [11:37:54] Coren: http://tools.wmflabs.org/admin/packages.html I synced all packages on exec nodes [11:38:01] except for grub and kernel [11:38:17] valhallasw: ? :< [11:43:30] Ill put the json here for you too see ;p [11:43:30] {"dump":"enwiki","nsinclude":["0"],"titlecontains":"","titlenotcontains":"\/","textcontains":"\\{\\{orphan(s|article)?\\}\\}","textnotcontains":"","textregex":"true"} [11:55:58] addshore: do you think you could get the filtering to work? not necessarily including reading the json, but just reading every step in the dump? [12:01:19] not sure what you mean :p [12:07:50] BTW is there a tools reqwuest feature? [12:13:01] Qcoder00 access request feature :P [12:16:25] OK [12:22:41] legoktm ping [12:34:39] @seenrx yuvip [12:34:40] petan: Last time I saw yuvipand1 they were quitting the network with reason: no reason was given at 10/31/2012 7:04:47 PM (214.17:29:51.9045920 ago) (multiple results were found: yuvipanda}brb, yuvipsnds, yuvipamda, yuvipanda__, yuvipanda___) [12:34:52] @seenrx yuvipanda [12:34:52] petan: Last time I saw yuvipanda}brb they were changing the nickname to , but is no longer in channel #wikimedia-office at 12/5/2012 6:31:33 PM (179.18:03:18.7644100 ago) (multiple results were found: yuvipanda__, yuvipanda___) [12:35:02] o.O [12:35:25] @seenrx panda [12:35:25] petan: Last time I saw yuvipanda}brb they were changing the nickname to , but is no longer in channel #wikimedia-office at 12/5/2012 6:31:33 PM (179.18:03:51.5630360 ago) (multiple results were found: Yuvipandan, pandalch, yuvipanda__, yuvipanda___) [12:35:35] wtf [12:35:57] @seen yuvipanda [12:35:57] petan: Last time I saw yuvipanda they were changing the nickname to zz_YuviPanda and zz_YuviPanda is still in the channel #pywikipediabot at 6/3/2013 11:00:42 AM (01:35:15.2618870 ago) [12:36:04] here we go [12:36:09] @seenrx YuviPanda [12:36:10] petan: Last time I saw YuviPanda they were changing the nickname to zz_YuviPanda and zz_YuviPanda is still in the channel #pywikipediabot at 6/3/2013 11:00:42 AM (01:35:27.5819830 ago) (multiple results were found: YuviPanda_, YuviPanda|Prep, YuviPanda|Storm, YuviPanda|sleeee, YuviPanda|sleep and 1 more results) [12:36:14] aha [12:36:21] case sensitive [13:06:45] addshore: the function that determines whether a page fits the criteria [13:07:06] basically something that says "if page.namespace in " etc [13:07:35] hi manybubbles [13:07:56] hi! [13:08:52] manybubbles: welcome to the Foundation! [13:11:57] sumanah: thanks! I'm happy to be here! [13:12:19] I have so much to read today! [13:12:41] and so much stuff to save for when my laptop comes [13:12:51] :) There is a long but exciting ramp-up period. [13:18:18] yeah - I started with two wiki pages and by the time I'd gone through them I had 17 other pages waiting for me to read. not to mention the handbook. [13:19:19] yep! [13:20:54] So when does the lab "Turn ON THE MAIN REACTOR!" [13:23:56] Qcoder00 we already did :P [13:24:21] Oh right [13:32:33] Qcoder00: list of tools currently running on labs [13:32:34] http://tools.wmflabs.org/ [13:34:07] btw Qcoder00 isn't this what you are looking for http://tools.wmflabs.org/catscan2/catscan2.php [13:34:18] Thanks [13:35:32] whats the interwiki for labds ? [13:35:37] ? [13:36:06] as in w: is wikipedia s: WQikisource etc... [13:36:30] none I know of [13:36:46] Hmm [13:47:08] I'd like to have wikipulse running on tool labs: https://github.com/edsu/wikipulse... Is this possible? [13:48:25] lbenedix depends [13:48:32] how complex you think it is? :o [13:48:58] it looks to me like some kind of server application, which probably should live on separate labs project [13:51:20] !log bots petrb: killing gluster daemon on bots-bnr1 it eats 6gb of ram [13:51:22] Logged the message, Master [14:04:02] Qcoder00: Now witness the firepower of this fully armed and operational Tool Labs! [14:08:28] Coren: You've not built an 'Ultimate Killer Invincible Doomeday Bot' have you? [14:08:30] XD [14:09:08] No, but we do have a door over our thermal exhaust ports. :-) [14:10:19] XD [14:10:38] "Tool LAbs - Max-i-mum Po-wer!!" [14:10:40] XD [14:10:52] Coren: can we do cross db joins yet? [14:11:37] !log tools petrb: removing /etc/logrotate.d/glusterlogs on all servers to fix logrotate daemon [14:11:39] Logged the message, Master [14:11:56] Coren, i had the catgraph test wikis running over the weekend, and everything went smooth so far [14:11:59] http://tools-dev.wmflabs.org:8090/list-graphs [14:12:11] no enwiki yet - too large [14:12:13] Betacommand: Yes, albeit not between wikis and commons or wikidata yet (well, unless they are on the commons shard) [14:12:23] JohannesK_WMDE: Yeay! [14:12:24] so i guess it is time to start testing with a separate vm [14:12:41] Betacommand: Cross wiki ? [14:12:42] Coren: then tools not at max :P [14:12:58] As in things like checking for images on Commons stil on enwiki and so on? [14:13:15] Coren do we need popularity contest? [14:13:27] petan: Why in blazes for? [14:13:30] it is borked anyway and produces annoying error logs [14:13:56] I wanted to remove it :> [14:13:57] petan: I'd say "No." [14:14:01] petan: wikipulse is written in nodejs and needs redis [14:14:20] does it mean "no I like it don't remove it pls" or "no I don't care you can remove it" [14:14:47] lbenedix I still don't know how it works [14:14:56] lbenedix it seems to me that it is a server that listen on some port [14:15:05] lbenedix that doesn't sound like a tool nor bot :P [14:15:14] hence it should be in own project [14:15:23] petan: It means I have no use for it. [14:15:36] its watching the irc-channels and returns the number of edits per time [14:15:38] !log tools petrb: removing popularity contest [14:15:39] Logged the message, Master [14:16:12] Coren so i guess it is time to start testing with a separate vm [14:16:39] can you create one Coren? i can do it if you tell me how ;) [14:16:52] JohannesK_WMDE: It may be. Do you think we can have a chat about it in ~45 mins when I have time to dedicate to you? [14:17:07] ok [14:18:16] right now it runs on heroku... here is the number of edits/min for wikidata: http://wikipulse.herokuapp.com/stats/wikidata-wikipedia/60000.json [14:19:46] Coren if I delete a mail from my inbox, is it deleted from your inbox as well? I mean, is root@tools.wmflabs.org shared inbox? [14:19:51] or it's a list? [14:20:11] root is an alias that explode to all members of local-admin [14:20:20] ok [14:20:27] So you can delete yours without concerns. [14:20:53] Coren, how's the researcher flag coming along? [14:21:50] Cyberpower678: No any faster than it was less than half a non-work day ago when you last asked. :-P It wont be for at least a week or two, at the very least. [14:22:32] Coren, I didn't ask yesterday. [14:22:51] I'll ask next week then. :p [14:25:30] Cyberpower678: Ah, no sorry you're right -- that was TParis. [14:26:09] Are we there? [14:26:14] no. [14:26:21] How about now? [14:26:25] Coren if someone needed to raise the limit for jobs, how can I do that? [14:27:06] petan: You should be, but I'd need a pretty darn good justification to go over 16. [14:27:37] I said "how can I do that" not "can I do that" :P [14:31:51] petan: You're a gridengine manager, so you can edit most config. [14:32:24] ok, but in which specific config is this set? and how to change per user? I couldn't find it :( [14:32:32] also is there some good documentation page for grid? [14:42:21] petan: Not really, though Oracle tends to have fair documentation on SGE 99% of which is applicable here. [14:50:50] is php-curl installed on tool labs? [14:53:32] php5-curl [14:53:38] that is [14:58:22] http://commons.wikimedia.org/wiki/Commons:Village_pump#File_moves_and_inactive_Commons_Delinker - Erm Has delinker migrasted? [15:23:16] become Coren ? [15:23:22] urgh [15:23:26] Coren, ? [15:23:29] you there? [15:23:40] Yes? [15:23:57] Coren, what's up with wikidatawiki_p? [15:24:05] It seems to be incomplete. [15:24:31] It is, there are odd things with its schema that need manual handling. on my TODO for today. [15:24:46] YES! [15:25:21] The edit counter seems to be handling it well but keeps complaining of missing tables. [15:36:30] MaxSem: Hi [15:37:36] May I ask what the status of maps-ceph1 is? [15:38:15] are you managing that through puppet, or just installing things straight out, as it is for test purposes only? [15:40:51] Coren, found another broken database. Reported by a user. [15:40:58] rowiki_p [15:41:10] is incomplete as well and has empty tables. [15:41:27] MariaDB [rowiki_p]> SELECT * FROM user; [15:41:27] Empty set (0.03 sec) [15:45:02] The actual replication for this one seems to have failed. [15:45:43] Cyberpower678: Will report to Asher. [15:45:59] Coren, thank you. [15:50:54] hi apmon [15:51:06] it's not managed by puppet [15:51:39] Do you mind if I just edit the ceph config then to get it up and running? [15:51:48] no objections from me [15:52:16] MaxSem: are you seriously setting up ceph? [15:52:32] I seriously attempted to:) [15:53:07] I'd like to try and see if I can get a proof of principle running for next weeks SotM-US, in case scaling of renderd comes up in discussions. [15:55:10] paravoid: It was pretty easy to get something / ceph running on my laptop, so it shouldn't be too difficult to set it up on labs either? [16:02:25] apmon: It shouldn't be. [16:02:31] legoktm: Around? [16:03:26] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 0 minutes) [16:16:54] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 13 minutes) [16:30:26] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 27 minutes) [16:43:55] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 40 minutes) [16:54:56] Coren, petan: hi [16:55:25] legoktm: Heya [16:55:31] legoktm have a look at /shared/memtest [16:55:41] legoktm there is a python script that works with new memcached [16:55:45] yuvipanda made it [16:55:53] ? [16:56:00] that memtest [16:56:03] it works with it [16:56:04] :D [16:56:07] nice :D [16:56:11] legoktm script does not [16:56:19] so maybe you can inspire by it [16:56:29] legoktm: Do you think you could spend a moment to revise your memory allocations? You seem to have been... overly generous. Like ufaa at 196M/2G [16:57:11] And the pywikibot stuff at 129M/2G and 192M/2G (peak 319M) [16:57:20] Coren: Ah sorry. Yeah sure. [16:57:27] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 54 minutes) [16:57:42] I was testing it but never reduced it once I finished testing [16:57:53] legoktm: It's not critical, but it would be the Right Thing to do. :-) [16:58:33] is there an easy way to see the peak memory a job is using/used? [17:02:31] legoktm: It's shown separately if it's different from the current high on the status page. [17:02:40] legoktm: http://tools.wmflabs.org/?status [17:03:46] ah thanks. lemme try and reduce it [17:03:55] legoktm: It also shows as maxvmem on the usage line of a 'qstat -j ' [17:03:59] ok [17:04:09] lol VMEM: 760M/762M [17:04:18] this is what I call optimization XD [17:04:56] petan: Actually, a well behaved program can run indefinitely at its limit if it has proper error handling for memory allocation. [17:05:32] in that case my bot is well behaving [17:05:46] Java is a good example, it's rather the glutton but once it allocated its arena it won't bust beyond it. [17:10:56] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 67 minutes) [17:14:33] New patchset: coren; "Tool Labs: Grant select on %_p on local DB as well" [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/66563 [17:15:15] Change merged: coren; [labs/toollabs] (master) - https://gerrit.wikimedia.org/r/66563 [17:15:38] Oh yeah Coren, is there an eta for the archive table being available? and is there a bug tracking it? [17:16:35] legoktm: I don't think there's a bug tracking it, and it's a couple weeks before I have a definitive answer. [17:17:40] Hm ok. I'll go file one then :) [17:18:50] Out of curiosity, is it a legal issue or a technical one thats holding it back? [17:19:29] legoktm: I think legal. [17:21:32] legoktm: Legal. [17:21:37] :/ [17:21:43] thanks though [17:22:18] I hope they say yes. Otherwise you can kiss the deleted edits count goodbye. [17:23:01] legoktm: I can tell you offhand that, if it is going to be okayed at all, it will be on a per-case basis and likely require approval with a process similar to that of getting the researcher right. [17:23:23] CP678|iPad: I still don't get why you think that. Edit count - not deleted edits = deleted edits. [17:23:40] RIP Toolserver :( [17:23:58] TS went down? [17:24:23] Coren, deleted edits = access to archive [17:24:26] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 81 minutes) [17:24:44] edit count = revision [17:24:55] Oh no lol, I was just making a bad joke [17:28:04] Filed https://bugzilla.wikimedia.org/show_bug.cgi?id=49088 [17:28:12] bbl lunch [17:28:13] legoktm, I saw/ [17:28:18] You CC'd me. [17:28:25] Cyberpower678: Nope. select user_name, user_editcount-(select count(rev_id) from revision_userindex where rev_user_text='CorenSearchBot') as deleted_edits from user where user_name='CorenSearchBot'; [17:28:55] user_name: CorenSearchBot [17:28:56] deleted_edits: 42701 [17:29:02] See? [17:29:34] And, honestly, if you're doing an edit counter you already have the counts of undeleted edits handy. :-) [17:30:00] Hey guys. I'm looking to get a public IP address for a project I'm managing. Can anyone help me out? [17:30:17] halfak_: sure. what project and what do you need it for? [17:30:48] Coren, could we still get access to the archives though? [17:31:09] "snuggle". I'm working on building an ecosystem of tools like this: snuggle.instance-proxy.wmflabs.org. I'm just starting to gain users. [17:31:20] So, it will be a web based wiki tool. [17:31:38] CP678|iPad: Like I said, if it is possible at all, it will require researcher-like access. It's not going to be by default, and it will require a case-by-case okay. [17:32:09] halfak_: were you looking to have a better dns name? [17:32:14] Ok. Who would be the ones to okay it? [17:32:30] Yes. That too. [17:32:43] I was hoping to set up a few dns names eventually. [17:32:50] oj [17:32:52] err [17:32:53] ok [17:33:06] snuggle as in STiki and Huggle? [17:33:30] :) Similar kind of tool, yes. However, this one is to find good newcomers, not revert the bad ones. [17:34:05] halfak_, so you're promoting vandalism? [17:34:13] no [17:34:15] what? [17:34:24] not revert the bad ones? [17:34:37] :\ The tool is for finding good newcomers. [17:34:46] Huggle and STiki are for reverting bad newcomers. [17:34:56] halfak_: upped your floating IP quota to 1 [17:35:23] we have good newcomers? I thought only well established editors were ok? ;) [17:35:28] CP678|iPad: I don't know; that's the kind of thing Legal will have to decide. I would expect it'll end up similar to http://en.wikipedia.org/wiki/Wikipedia_talk:FAQ/Research [17:35:45] Excellent. Do I just allocate an IP now? [17:35:56] halfak_: I like the idea of a tool that finds good newcomers, rather than just reverts people :) [17:35:59] halfak_: yep [17:36:24] Awesome. Thanks for your help! I'm back to hacking. [17:36:34] halfak_: hm [17:36:45] it looks like this takes the wiki user's username and password? [17:37:01] Yes. This is something I struggled with. Is there a better way? [17:37:16] not yet, but soon, yes [17:37:28] that said, you're not allowed to do this via the labs ToS [17:37:29] As soon as it's ready, I'll be on board. [17:37:34] :\ [17:37:41] That's a pretty big deal breaker. [17:37:49] you need to wait till oAuth or OpenID are available [17:37:55] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 94 minutes) [17:37:56] halfak_: The alternative, OTOH, will be okay [17:37:57] those are allowed via the ToS [17:38:14] So Huggle couldn't run on Labs? [17:38:34] not unless it uses OAuth, or OpenID for authentication [17:38:53] OAuth is targetted for this quarter [17:38:57] OpenID should as well [17:39:01] *should be [17:39:08] Wow. That's not OK. I guess I'll need to move off of labs then. [17:39:08] halfak_: Huggle is a bad comparison, though, because it works on the /user's/ equipment [17:39:09] *sigh* [17:39:15] halfak_: this isn't just a labs thing [17:39:46] halfak_: if we see applications asking for credentials for users, they'll likely be asked to stop [17:39:51] Coren, Huggle has a server backend. Most everything for Snuggle exists in the user's browser. [17:39:54] anywhere on the internet [17:40:09] What do you propose that I do, Ryan? [17:40:31] does this app do anything on a user's behalf? [17:40:40] like edit, or give barnstars, etc? [17:40:44] Only what the user commands it to do. [17:40:47] * Ryan_Lane nods [17:40:55] you'll need to add OAuth support [17:41:06] But there's no OAuth available. [17:41:11] halfak_: Nevertheless, the idea is that the credentials live on a server that is not under the user's control. That's a very big no-no. [17:41:17] hm. no csteipp in the channel [17:41:49] Coren, the credentials are not stored on the server. [17:41:49] halfak_: I'm asking someone to join the channel [17:42:03] Ryan_Lane: what's up? [17:42:16] csteipp: halfak_ is a great guinea pig user of OAuth: http://snuggle.instance-proxy.wmflabs.org/ [17:42:25] halfak_: Wait, the edit doesn't come from the tool, but from the *end user's* client? [17:42:59] Oh the edit comes from the tool. The user shares cookies with the server to make the edit. [17:43:09] The server can see the cookies which act like an auth token. [17:43:35] halfak_: Which domain's cookies are shared? [17:44:01] Coren, not a viable otpion [17:44:03] halfak_: snuggle uses the username/password combo to log into wikipedia, then stores the cookie, correct? [17:44:06] Sorry about that. [17:44:14] Yes Ryan, that's right. [17:44:37] Cyberpower678: What isn't? For the edit counter, you mean? [17:44:39] the issue here is that the username/password combo are passing through snuggle [17:44:39] yes. [17:44:51] because snuggle could use it to steal project credentials [17:44:57] Ah, yes. This is coming back to me :) [17:45:02] Yes. Technically, it could. [17:45:08] Cyberpower678: Could you elaborate? [17:45:13] csteipp: this is the original use-case I was mentioning months ago :) [17:45:13] Yep, so yeah, this is perfect use for oauth [17:45:41] scfc_de, I'm talking about retrieving deleted edits. [17:45:47] Is Oauth ready? I was hoping to release version 1.0 today. [17:45:55] halfak_: for this to continue operating right now we're going to need to get legal onboard with this [17:45:58] Cyberpower678: I'm afraid that, short of a researcher-like okay, archive is not going to be generally available. TS should never have had it in the first place, and any tool that relies on it will have to either get blessing or find some other MO [17:46:16] halfak_: No, not yet. We're targeting to have at least an initial version out by the end of the month... but it's not up yet [17:46:19] halfak_: at mininum, anyone with root level access to the snuggle project will need to sign an NDA [17:46:34] OK. That's cool because it is me and I have. [17:46:45] I'll not allow any others into the project until this is resolved. [17:46:54] Coren, which is why I will be patiently waiting to hear from legal. :-) [17:46:55] ok. let me get legal onboard otherwise, so they know what's up [17:47:15] we can likely make an exception to the policy for this until OAuth is ready [17:47:17] What's the project? [17:47:18] Cyberpower678: Why doesn't Coren's query work for you? [17:47:20] OK. Thanks for your help Ryan. [17:47:44] scfc_de: There's a bug in it. :-) [17:47:52] Incoming little flood. [17:47:56] scfc_de, [17:47:58] MariaDB [enwiki_p]> select user_name, user_editcount-(select count(rev_id) from revision_userindex where rev_user_text='MBisanz') as deleted_edits from user where user_name='MBisanz'; [17:47:59] +-----------+---------------+ [17:47:59] | user_name | deleted_edits | [17:47:59] +-----------+---------------+ [17:47:59] | MBisanz | -16364 | [17:48:01] +-----------+---------------+ [17:48:24] scfc_de, that's why [17:49:21] * Coren tries to figure out how that is possible. [17:50:31] Cyberpower678: Thanks for explaining. I think it should be possible to provide a view/snapshot of archive that only contains the aggregated information that is needed for editcounters. [17:51:13] Coren: http://www.mediawiki.org/wiki/Manual:User_table#user_editcount [17:51:27] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 108 minutes) [17:51:51] scfc_de: Ah, I was aware, but not that it could deviate /that/ much. [17:52:45] So someone just needs to run initEditCount.php on the cluster. Can't take long :-). [17:53:42] Actually, no, initEditCount.php only counts revision, not archive. [17:54:34] Ryan_Lane: Do you know if new files are added atomically to /public/datasets/public? Otherwise, who's the one to ask? [17:55:07] As long as I get an accurat AND realistic number. [17:56:34] scfc_de: I doubt they are added atomically [17:56:41] CP678|iPad: Actually, that should really be pushed to the API [17:56:59] scfc_de: afaik they are copied directly into place [17:57:14] maybe it's using rsync and doing a move to the permanent spot? [17:57:22] That would mean hundreds of API queries. [17:57:27] ariel would be the person to ask [17:57:36] apergos on irc [17:57:44] Ryan_Lane: There's some rsync logs there, and rsync does atomic I believe. I look for Ariel, thanks. [17:57:51] Cyberpower678: FYI: The bug with rowiki is that rowiki is on S7 and thus does not exist yet. :-) [17:58:04] Ok. [17:59:13] CP678|iPad: How do you figure? It'd just be a query/auprop [18:00:51] Moving everything to the API for hundreds of user, means hundreds of queries. [18:01:32] CP678|iPad: The API has generators and often you can specify more than one argument. [18:02:33] Crap. The siren in my area is going off. [18:04:55] Warning: There is 1 user waiting for shell: Alvaro Lopez Tijeras (waiting 121 minutes) [18:06:29] Really?!? [18:06:44] CP678|iPad: are you safe? [18:06:55] A thunderstorm is going through my area and they decide to do the siren tests NOW? [18:07:00] WTF? [18:07:42] stay safe, CP678|iPad [18:07:45] maybe it isn't a test [18:08:00] The siren is off now. [18:09:08] I plan to stay safe though. [18:18:28]