[12:23:38] merry christmas! [12:23:59] I was wondering: is there some maintanance going on currently? [12:24:25] If I try tools.texbot@tools-bastion-03:~$ qstat [12:24:35] i get [12:24:52] error: commlib error: access denied (server host resolves rdata host "tools-bastion-03" as "tools-bastion-03.tools.eqiad.wmflabs") error: unable to contact qmaster using port 6444 on host "tools-grid-master.tools.eqiad.wmflabs" [12:26:11] I do not believe there is any maintenance going on over the christmas holiday Debenben [12:27:07] that is weired. I got the same message yesterday as well, but then it worked again after a few minutes [12:27:34] Debenben, try tools-bastion-02 ? [12:27:54] or tools-bastion-05 [12:28:20] tools bastion-02 seems to work [12:28:33] (tools-trusty.wmflabs.org or dev.tools.wmflabs.org if you don't have SSH set up properly) [12:28:49] mmkay [12:29:00] will leave a ticket reporting -03 as having a problem [12:30:46] Debenben, https://phabricator.wikimedia.org/T212594 [12:31:10] thanks [12:31:33] Debenben, to clarify this affects submissions as well right? [12:32:01] yes, I get the same message in the error log for a job that i submit [12:32:06] hmm [12:32:17] on tools-bastion-02 it runs fine [12:32:25] I wonder if we can leave this be over christmas or if a tools administrator should look into it [12:32:44] zhuyifei1999_, about? [12:33:02] I think it is not an urgent issue, unless it affects other servers as well [12:33:30] and maybe we should try to get some more debug messages, but i don't know how [12:33:33] right but will other users know how to check the others/ [12:33:34] ? [12:33:59] I'm conscious that some tools users may not have the skills I'd otherwise expect of cloud vps users [12:34:15] well, it might be that it repairs itself after half an hour or so, and then nobody knows what exactly the issue was [13:50:15] Debenben, looks like it hasn't repaired itself [13:51:39] yes, the problem is still there. maybe yesterday it was there as well, and it just happened that i logged into some other server [13:51:58] and then i thought it had repaired itself [13:55:16] if some admin is there, maybe he can let it reboot [19:02:31] Krenair: yes [19:02:47] looks [19:07:52] (looks like fixed by gtir.loni) [23:20:10] hi, I think I need a technical clarification by someone who knows what actually is in the public database (e.g. Quarry) and what isn't. A tool developer told me he's training (likely a neuronal network) an AI using Huggle+Rollback edits as "reverts" and thanked edits as "non-reverts". [23:20:10] I would say this is not possible for him to do, because the Thanks log contains no information about specific entries. I think that the specific edit numbers are private information that is unlikely to be present in any public database. [23:20:10] According to [[en:WP:Thanks]] and the thanks log, it is not possible for others to see *which* edit someone has thanked for. [23:20:10] Is there any way for a "normal" user to do so, or is this information private and only ever shown to the thanked user? [23:38:21] ToBeCloud: I think it is public [23:38:55] in fact, when you thank an edit, there's a warning "you are going to publically thank this edit..." [23:42:25] Platonides: I have now even attempted to set up a quarry query, and while I may have missed something, it also contains no such information [23:43:02] I think that the warning is about the public thanks log, but not the detail [23:43:16] everyone can come to conclusions about the time of the "Thanks" [23:43:54] yet a bot can't reliably do this. Here's my quarry: https://quarry.wmflabs.org/query/32269 [23:44:20] log_page points to the page ID of the user page of the thanked user, not the thanked page [23:44:46] I found this out by using https://en.wikipedia.org/wiki/Special:Redirect , selecting "Page ID" and copying the ID there [23:45:03] Someone tried something like this a few months ago, I don't remember details but I think we determined that revision information only gets sent to Echo [23:45:09] Platonides, only the fact of the "thank" being sent is public. The revision that it was applied to isn't. cf. https://meta.wikimedia.org/wiki/Special:Log?type=thanks and https://phabricator.wikimedia.org/T51087 - the comment there at May 24 2016, 9:22 AM says "we put the post/revision ID in the event metadata but not in the public logging table" so I believe it is not public. [23:47:04] "patch author is banned" lol [23:47:23] thanks for the phabricator ticket link, that's exactly what I had been looking for [23:48:09] Someone should really write that on mw:Extension:Thanks [23:51:39] ToBeCloud, sidenote: Changing the configuration would be a bit of a nightmare, because we'd have to massively-announce the change so that /nobody/ is surprised. It'd also require a huge discussion, because if the specific diff is public then the feature could easily be bodged into being widely visible, and then we've suddenly got a "public upvote" system on our hands, which isn't what the feature was intended/designed as. [23:56:08] quiddity: Ah, thanks for the detailled consideration. I completely agree, and I believe that the current state is a good idea [23:56:28] quiddity: now the bot owner just needs to think of a way to get "good edits" for their database. :D [23:56:36] not an easy task, I think [23:56:52] Most people use humans for that directly [23:56:53] :) and agreed. [23:56:56] heh [23:57:01] :D [23:57:14] ToBeCloud: is that person aware of ORES? [23:57:23] ^ [23:57:46] yes; discussion (a hell lot if you like to read through it) is available on https://en.wikipedia.org/wiki/User_talk:ToBeFree -- search the page for "Redactyll" [23:58:03] the user is sadly not coding themselves, but rather acting as a manager [23:59:05] to me there seems to be an unmanageable lot of problems in that project, and it will not be easy to tell them kindly to completely restart and redesign from scratch [23:59:13] or to give up, even