[01:03:53] https://bugzilla.wikimedia.org/show_bug.cgi?id=70493 anyone suggest workaround? please? :( [01:12:15] Svetlana: there's no workaround, but i'll write a patch that fixes it [01:14:22] ta :) [01:15:38] hah [01:16:04] ohhhhhh, i forgot Svetlana was gry [01:16:11] :) [01:16:36] yes, i like my name, i should totally become more and more Svetlana-like as time goes :) [01:24:08] Svetlana: https://gerrit.wikimedia.org/r/#/c/158906/ [01:41:21] why does the "foreach( $this->mPreloadparams as $preloadparams ) {" and the rest of the loop happen twice? the same block of code [01:42:47] i more or less understand the rest :) [01:49:26] Svetlana: type=comment and type=commenttitle have 2 different code paths, so i added it to both [01:55:46] ah. and what about 'create'? preload is also available for it, too [01:56:41] https://www.mediawiki.org/wiki/svetlana_does_not_exist?action=edit&nosummary=true&preload=Manual:Creating_pages_with_preloaded_text/param_demo&preloadparams[]=first [02:00:07] Svetlana: create and comment share the same code path [02:08:39] ah. I see. thanks a lot :) [04:35:25] ori: Please can you elaborate about FOSS intership project "wekimedia performance portal" [04:38:20] exploreshaifali: link? [04:39:21] i see http://www.mediawiki.org/wiki/Mentorship_programs/Possible_projects#Wikimedia_Performance_portal [04:39:26] did you read that section? [04:39:47] jermyb: https://www.mediawiki.org/wiki/Mentorship_programs/Possible_projects#Wikimedia_Performance_portal [04:40:05] yes I read it [04:40:21] I need access to data [04:40:37] ok [04:40:53] does ori know how to contact you? [04:41:00] in case you're not here when he gets back [04:42:17] yes I have tried to contact ori many times, but as he was busy in some other stuff he replied me after long time [04:42:51] is [[user:Shailesh Singh]] you? [04:43:02] i guess not quite [04:43:21] and than I had some budy schedule since past few weeks so unable to contact him. But he said that we can discuss on IRC so I am asking right now [04:43:36] no I am not that [04:43:51] *busy schedule [04:43:54] right, but it's a weekend. which means hit or miss. he could be e.g. at a movie or otherwise off the clock [04:44:13] all right [04:44:16] and in any case timeszones [04:44:20] timezones* [04:44:28] I should try to contact him tomorrow [04:44:29] so, how should he contact you? [04:44:36] you could always comment on ori's talk [age [04:44:38] *page [04:44:45] sure, that's an option [04:45:35] ori's talk page? [04:45:50] yes [04:46:11] http://www.mediawiki.org/w/index.php?title=User_talk:Ori.livneh&action=edit§ion=new [04:46:13] but... [04:46:20] there's a note at the top of the page [04:46:32] > {{Note|'''Note''': I prefer to be contacted via e-mail or [[IRC]].}} [04:47:19] yes, fine zthanks! [04:47:32] Thanks [04:47:44] exploreshaifali: so, what's your username, email address, or some other way to find you? [04:47:51] his email is ori at wikimedia dot org [04:47:59] try emailing him? [04:48:46] yes, we had contact on mails only. There he mentioned to talk more on IRC. btw my email address is agrawalshaifali09@gmail.com [04:49:21] ok. and timezone? [04:50:01] I am in India, timezone is IST [04:50:08] ok, so +0530 [04:50:13] yes [04:50:23] great [04:51:12] so I will contact him tomkorrow [04:51:17] ok [04:51:37] i'm interested in that project too. maybe we'll talk more after one of us gets more info from ori [04:52:35] sure! what's your timezone [04:53:32] jeremyb: what is any other means to contact you? [04:53:48] America/New_York [04:53:52] i.e. -0400 [04:54:04] cool [04:54:10] jeremy@wikimedianyc.org [04:55:22] exploreshaifali: ideally (at least according to my thoughts in the last 2 mins) you'd have not just perf accounted for but also broken machines [04:56:00] so, perf change may be explained by an outage of one node even where there is redundancy [04:56:31] https://ganglia.wikimedia.org/latest/?r=day&cs=&ce=&c=Swift+eqiad&h=ms-be1005.eqiad.wmnet&tab=m&vn=&hide-hf=false&mc=2&z=medium&metric_group=ALLGROUPS [04:56:45] and https://ganglia.wikimedia.org/latest/?c=Swift%20eqiad&m=cpu_report&r=day&s=by%20name&hc=4&mc=2 [04:56:58] exploreshaifali: ^ [04:57:47] I need to know more about dataset [04:59:20] exploreshaifali: which dataset? [04:59:40] what do you think about the example in those last 2 links? [05:01:14] exploreshaifali: hrmmm, i wonder if ori might be on now actually [05:01:56] graphs are showing about performance of nodes/cluster in terms of process, loads [05:02:37] memory, network [05:02:37] right. but do you see the change in the last hours? [05:03:18] a little more than 2 hours ago there was a big change [05:07:03] No, about which change you are takling? [05:07:25] time is always "now" there in graphs [05:08:02] well see the first link above [05:08:08] that's for a specific host [05:08:18] the host which is now a problem [05:08:46] yes it is showing host overvoew [05:09:52] see e.g. https://ganglia.wikimedia.org/latest/graph.php?r=day&z=xlarge&c=Swift+eqiad&m=swift_HEAD_200_50th&s=by+name&mc=2&g=load_report [05:09:56] that's the cluster overall [05:10:12] but all of that change is from a single host [05:12:47] probably you are stating that all changes on a cluster are due to single host? [05:13:18] correct? [05:15:01] jeremyb: ^ [05:15:13] i don't understand [05:15:25] the last link above is aggregate [05:15:35] basically just a sum of all the parts [05:16:05] and there's a dramatic change in that graph a little after 01:15 [05:16:08] yes because it is of cluster [05:16:26] i'm saying all of the change at that instant was caused by a single box changing [05:18:25] sorry but about which dramatic change you are talking? [05:18:45] I am still unable to configure it [05:19:31] well as i said, 01:15 [05:19:38] the times are labeled at the bottom [05:21:57] * jeremyb is getting a better links [05:22:23] If you are talking about graph labled "Swift eqiad Cluster load last day" it contain times Sat:6, Sat:12, Sat 18 and sun:0.0 [05:22:42] right [05:22:50] 01:15? [05:22:52] so, 1:15 to the right of sun 0 [05:23:12] but there is nothing like drastic change [05:23:14] although maybe i'm misreading [05:23:48] yeah, actually it was a little after 02:45 [05:24:36] remember not to ignore the scale on the y (left) axis [05:24:39] http://ganglia.wikimedia.org/latest/?r=custom&cs=09%2F07%2F2014+02%3A15&ce=09%2F07%2F2014+03%3A15&m=load_report&s=by+name&c=Swift+eqiad&h=&host_regex=&max_graphs=0&tab=m&vn=&hide-hf=false&sh=1&z=small&hc=4 [05:25:23] anyway, the point was that a subsystem doesn't have to be totally broken in order to have an impact on overall site perf [05:25:35] could be just a single bad box [05:27:18] if we look at whole graph, I don't think there is any drastic change after sun 0:0 [05:28:31] how about http://ganglia.wikimedia.org/latest/graph.php?r=4hr&z=xlarge&h=ms-be1005.eqiad.wmnet&m=load_report&s=by+name&mc=2&g=load_report&c=Swift+eqiad ? [05:28:50] and what's your threshold for drastic?? :) [05:29:39] what this grey shaded area mean : 1-min is what? [05:30:21] yes there is change in that parameter 1-min [05:30:42] yes, that's the only one that's expected to change [05:30:48] the metrics are static [05:30:58] the number of CPUs on a box doesn't change over time [05:31:13] (well i guess procs could change though) [05:31:22] 1min is 1 minute load avg [05:31:51] http://en.wikipedia.org/wiki/Load_%28computing%29#Unix-style_load_calculation [05:32:06] so per minute load is changed drastically [05:32:58] and yes number of CPUs won't change but Procs could also change [05:38:11] yes, very drastically :) [05:38:43] :) [05:40:23] reason s of increasing load could be increase in number of processes waiting for CPU. But it is not as Procs havn't changes much [05:44:12] jeremyb: In the graph Procs means number of running processes, right? [05:44:30] so if Procs increases 1 min should also increase [05:44:51] but here procs is not changing as much as 1-min [05:45:07] so what other reasons could be for this change? [05:47:36] yes [05:48:00] well procs could be idle or active [05:48:21] i think the procs count doesn't change depending on activity [05:48:31] see also https://graphite.wikimedia.org/render/?title=top%203%20load%20average%201min%20be/fe&from=-5hour&width=1024&height=500&until=-2hour&areaMode=none&hideLegend=&target=aliasByNode%28movingAverage%28highestMax%28servers.ms-be1*.loadavg.01.value,%203%29,10%29,1%29&target=aliasByNode%28movingAverage%28highestMax%28servers.ms-fe1*.loadavg.01.value,%203%29,10%29,%201%29 [05:52:17] as per I understood: ms-be1005 is a single host [05:52:23] correct? [05:52:53] yes [05:53:13] but what is there in y axis of last graph [05:53:47] x axis is showing time? [05:54:01] yes. y is 1 minute. x is time [05:54:36] okay so this change is due to that host [05:56:37] if we will look at https://ganglia.wikimedia.org/latest/?r=day&cs=&ce=&c=Swift+eqiad&h=ms-be1005.eqiad.wmnet&tab=m&vn=&hide-hf=false&mc=2&z=medium&metric_group=ALLGROUPS [05:57:19] there are changes in load last day, CPU last day, and network last day graphs [05:58:02] and see network out is reduced to zero [05:58:09] right [06:01:00] but what's the reason of this sudden change? [06:01:24] the disk is broken. i think. or filesystem issue of some sort [06:01:39] yes it could be [06:01:54] as Procs doesn't change much [06:02:44] hey network out is reduced to zero this could also be a reason [06:03:02] you mean, network card broken? [06:03:41] yes something like so [06:04:08] nah, we had this problem before. that's why the monitoring for it is a little better now [06:04:13] because see network in is also reduced to zero [06:04:19] it's definitely disk or filesystem related [06:04:25] but filesystem could mean kernel [06:04:30] don't remember all the details [06:05:08] okay [06:05:43] at that time networks were also showing same behaviour as they are now, totally zero [06:05:52] ? [06:06:16] idk, let's see [06:07:00] :) [06:09:18] last time was ms-be1010 [06:09:19] http://ganglia.wikimedia.org/latest/?r=custom&cs=08%2F30%2F2014+12%3A15&ce=08%2F31%2F2014+08%3A15&m=network_report&s=by+name&c=Swift+eqiad&h=&host_regex=&max_graphs=0&tab=m&vn=&hide-hf=false&sh=1&z=small&hc=4 [06:09:30] exploreshaifali: [06:11:43] so last time also network was reduced to zero, so this time also disk failure could be a reason [06:13:49] anyway, there's a lot of moving parts and interdependencies. and maybe perf dash should indicate stuff like how many boxes per cluster are broken so you can see if that correlates with more general perf changes [06:14:20] see one difference in last and this time: user in CPU last day was down to zero [06:19:56] In wekimedia performance portal what we need to do, I thoung I will have cluster response time and all such data to which I have to analyse and put them on graphs [06:20:11] but it seems like this is already happening [06:21:46] well ori said we have a lot of data and interfaces to view it but not a curator [06:21:52] to decide which to display or not [06:22:09] as i read what he wrote onwiki [06:23:53] ok [06:29:11] exploreshaifali: https://ganglia.wikimedia.org/latest/?r=4hr&cs=&ce=&c=Swift+eqiad&h=ms-be1005.eqiad.wmnet&tab=m&vn=&hide-hf=false&mc=2&z=medium&metric_group=ALLGROUPS [06:31:05] and https://graphite.wikimedia.org/render/?title=top%203%20load%20average%201min%20be/fe&from=-6hour&width=1024&height=500&until=now&areaMode=none&hideLegend=&target=aliasByNode%28movingAverage%28highestMax%28servers.ms-be1*.loadavg.01.value,%203%29,10%29,1%29&target=aliasByNode%28movingAverage%28highestMax%28servers.ms-fe1*.loadavg.01.value,%203%29,10%29,%201%29 [06:34:38] sorry gtg [06:34:56] brb in 3 hrs [06:36:07] bye! [08:03:19] Interesting. Gerrit added me to the list of reviewers of a diff, because I reported a bug that the diff is fixing. [08:14:09] Svetlana: I don't think gerrit did that, probably the person who uploaded the patch added you as a reviewer. [08:17:31] Ah. I see. [08:17:46] you can figure out by looking at the email you got [08:18:23] > [08:18:24] Hello Legoktm, [08:18:24] I'd like you to do a code review. Please visit [08:18:27] ... [08:18:28] > [08:18:29] Yeah, I should learn to use getmail. Cumulatively, I'm spending a few hours a week *waiting* for an email message to load. [08:19:10] It doesn't have a signature (or something suggesting who added me as a reviewer). :-) [08:19:13] I really like thunderbird's offline mode [08:19:20] look at the From header [08:19:27] From: "[name] (Code Review)" [08:19:44] or the reply-to header I guess [08:19:57] ah. yes, that's it, thanks. -- I tried thunderbird, but the format it uses to save messages was unreadable last time I checked (easy to corrupt 1 file and lose all mail). [13:10:46] hello, [13:10:59] assume my wiki page is too long [13:11:11] i want to have a button in the end [13:11:29] which when i click n this, the page go up [13:14:25] !test [13:14:25] This is a nifty test of niftiness ... http://z0r.de/2650 << lookie, isn't that cool? :) [13:14:29] Nice. [13:16:06] : could you help? [13:16:29] sasan: [[#top|link text here]] [13:16:40] sasan: Use that code? ^ [13:16:54] where? [13:17:09] in where i shuld use this code? [13:17:12] ^ [13:17:29] Where you want the link. [13:17:50] in my wiki [13:17:56] Do you understand HTML id=""? [13:18:06] Do you understand HTML anchors and links? [13:18:08] yes [13:18:18] Okay, [13:18:24] so there's an HTML id in the page called "top". [13:18:30] Which you can link to use [[#top]] syntax. [13:18:35] [[ ]] is internal link syntax. [13:18:38] no [13:18:41] Yes. [13:19:05] i do not find html ip=top [13:19:59] Do you have name="top"? [13:20:12] https://test.wikipedia.org/wiki/A_long_page [13:20:16] That's a test page. [13:20:23] no, i think i should generate [13:20:27] please see the wiki [13:20:33] parsintelligent.com/wiki [13:20:46] Provide a usable link. [13:21:39] www.parsintelligent.com/wiki [13:21:39] I made https://test.wikipedia.org/wiki/A_long_page better. [13:22:06] http://parsintelligent.com/wiki/?uselang=en [13:22:08] Permissions error. [13:22:18] Your wiki is restricted. I can't see the content. [13:22:38] i can send you my user and pass in private page [13:22:44] ok? [13:22:48] No. [13:23:12] so what should i do? [13:23:23] Better explain your problem. [13:23:27] So that others can help you. [13:23:58] if you enter in my wiki, you could see most of my pages is too long [13:24:10] so it is better if i have a button in the end of the page [13:24:23] which if the user press that button [13:24:23] which if the user press that butto [13:24:45] it goes up, at the first of page [13:25:08] understand my question? [13:26:05] Clearly not. [13:27:29] havr you ever visited some website [13:27:39] Yep. [13:27:42] and read some intersting note there [13:27:42] long note [13:27:57] at the end of the page, you can see a button [13:28:14] when you click on that, it goes to the first part of page [13:28:18] yes? [13:28:29] Ah, you want to add a ¨Go to top¨ link. [13:28:56] yes [13:31:02] could you help me?? [13:37:03] : sorry [13:37:16] my net has problem [13:37:17] could you help me with the problem [13:37:43] i turn offline [13:37:53] and i did not get your solution [13:38:01] maybe send me [13:39:22] where are you [13:39:29] no solution? [13:40:41] Niharika: I thought the same. [13:40:55] I made a demo page at https://test.wikipedia.org/wiki/A_long_page [13:46:09] we can not have icon based [13:46:10] ? [13:46:15] it is just a sentence [13:46:24] if it can be icon [13:46:31] You can use CSS to style the link. [13:46:36] it is more effective [13:46:46] how could? [13:46:55] could you help me? [13:47:26] have any demo? [13:51:04] : ?? [13:53:35] sasan: Carmela made you a demo: https://test.wikipedia.org/wiki/A_long_page#top [13:54:11] Niharika: I think he wants a prettier link. [13:54:56] yes [13:55:01] Ah. Okay. [13:55:10] Niharika: could you help me? [13:55:37] sasan: Can you be more descriptive what exactly you want? An icon? [13:55:54] yes, and icon [13:56:09] which link me to first part of page [13:57:29] sasan: I´m not sure how you could insert an icon. Carmela, ideas? [13:57:53] Use CSS, as I said. [13:58:20] : common.css? [13:58:24] Sure. [14:00:12] yes? [14:00:34] please let me know which command i should add [14:00:34] in order to add icon [14:02:12] Niharika: https://test.wikipedia.org/wiki/A_long_page#footer [14:02:32] https://test.wikipedia.org/w/index.php?title=MediaWiki:Common.css&diff=prev&oldid=216125 [14:02:37] https://test.wikipedia.org/w/index.php?title=A_long_page&diff=prev&oldid=216126 [14:02:40] You can do something like that. [14:05:05] Carmela: Cool. :) I hadn´t inserted HTMl directly into a page before. [14:05:34] Wikimarkup is somewhat painfully limited. [14:05:40] HTML is often needed. [14:06:21] True. [15:07:42] can anyone help with this? http://pastebin.com/67qS2icz [15:08:01] I keep getting failed to bind proxy and userDN is blank [16:03:23] legoktm: it was not your fault... now better https://en.wikipedia.org/w/index.php?title=Bus_factor&diff=624551899&oldid=622362084 [16:18:21] Hi, could someone possibly help me solve a DNS error with my Wiki? Errors such as "empty label" and "not loaded due to errors" occur. [16:18:49] It's disallowing any page on th wiki to load. [16:20:02] It says I'm missing files from "/tmp", but as usual that folder is empty, [16:20:52] I have two A records for me site: 'mysite.com' and '.mysite.com' - Are one of those incorrect? The second was added because, oddly, I could access mysite without adding www. [16:21:02] *my (I'm not a pirate) [16:21:27] *I couldn't access my site without adding www. [16:24:57] Meow_: I don't know why that would be a DNS issue. If you have a DNS issue, your browser usually complains with a message similar to "Couldn't resolve domain example.com" [16:25:48] Vulpix: But the error wasn't tere last night - I haven't touched anything on the server since then, and now "The reqested URL can't be found on the server". [16:25:56] *there [16:26:05] I had updated the DNS about two days prior. [16:26:30] It's a "404 not found" error. [16:27:48] a 404 error means the server can't find the resource you requested (URL). Be sure the file/folder exists [16:28:32] Vulpix: The scenario is that I ahd several tabs open with Wiki pages, and left my computer on through the night - In the morning, upon refreshing, they can't ben found. [16:28:38] *be [16:29:36] did you double-check if those files still exist on the server? [16:29:48] did you check apache error_log? [16:30:02] (or the error log of your server if not apache) [16:31:18] Vulpix: Yes, everything's still there. Only actual files on the server will load, such as images. Anything accessed through the Wiki (example.com/Main_Page) produces a 404 error when it didn't before. [16:32:32] To be honest, I don't know where the apache error log would be. There are so many logs in so many places, at the moment I can only recall the location of the SQL error log. [16:32:52] it's usually in /var/log/apache/error_log [16:33:01] I´m trying to push a patch to gerrit and I´m getting a permission denied (publickey) error. I tried ssh-ing directly into the gerrit account, and the authentication succeeds, yet I cannot do a git-review. Here´s the trace: https://dpaste.de/sPNr#L2,3,89,102,104,132,133,136,137,139 [16:34:59] Vulpix: Found it. I dont see anything odd - Past two error reports today, I can still see several PHP errors from developing yesterday. [16:35:43] Meow_: 404 errors are usually logged there. Do you see them? What actual file is trying to access? [16:37:52] Vulpix: Nothing about 404 today, but one odd thing that stands out would be 'mpm_prefork:notice' [16:38:46] Vulpix: I'm trying to access Wiki pages, not actualy files. It keeps telling me that "the file can't be found". [16:39:01] Niharika: try "git remote -v". Be sure urls have your username on it (niharika29@gerrit.wikimedia.org:29418 ...) [16:39:14] Correction - The URL can't be found. [16:39:54] Meow_: try accessing Special:RecentChanges on your wiki (directly by the URL) [16:40:20] Vulpix: 404 [16:40:41] Wait, you mean the literal path to php file? [16:40:56] no, just the page [[Special:RecentChanges]] [16:41:03] Yep, 404. [16:41:45] Vulpix: I have the remote urls correct. [16:41:49] Meow_: maybe with your DNS change, it's pointing to a different IP, and requests aren't being handled by that server [16:42:05] Meow_: check if the IP address it resolves is correct [16:42:18] Vulpix: Anything else that might be going wrong? [16:42:30] Niharika: did you upload your public key to gerrit? [16:42:55] Vulpix: Yes. That´s how it authenticates me over SSH, I believe. (If you saw the trace) [16:43:14] Vulpix: It is correct. I am getting some DNS errors, however, would you like to see them? [16:43:26] Meow_: yes [16:43:57] are there any wiki farms that have socialprofile pre-enabled? [16:44:10] https://dpaste.de/2t3Z [16:45:46] Niharika: does "git pull" work? [16:46:01] Vulpix: Yes it does. [16:46:32] ping ashley [16:47:26] Niharika: I have no idea, then :( maybe git-review got screwed up on your machine or something, try asking also in #wikimedia-dev [16:47:39] Vulpix: Okay, thanks! :) [16:48:15] Meow_: if the name resolves to the correct IP, I don't see what could be the problem there [16:49:05] Vulpix: Are the errors out of the ordinary? I will also note that I'm using rewrite rules, but again, they worked yesterday. [16:49:27] I don't know about bind errors [16:50:38] Vulpix: Another thing to note, PHPMyAdmin still loads perfectly fine from it's URL. [16:52:40] Vulpix: If you still have time, would you like to check out the configuration for my rewrite rules? Last thing I can think of. [16:53:53] Meow_: I'd need to know also your wiki configuration and your filesystem layout [16:55:14] Vulpix: The layout is unmodified and untouched - With the exception of the fact that PHP has write access to the cache directory: https://dpaste.de/3GaY [16:55:48] i was thinking of making a social networking site with mediawiki and socialprofile extensions but unfortunately i suck at mediawiki :/ [16:56:11] so, assuming your index.php is in the root folder [16:56:11] I just use one of those "short URL generators", but it worked just fine yesterday. [16:57:06] Meow_: what if you go to the "/index.php" URL? [16:57:25] Vulpix: In the root folder as far the server's default path it directs URLs. Going to mysite.com/index.php redirects to mysite.com/Main_Page, which is a 404. [16:58:48] Meow_: well, your rewrite rules are OK, except that it's missing a rule to translate "/Article" to "/index.php?title=Article" [16:59:20] Meow_: also note that having short urls with the page name directly under the root directory is not recommended [16:59:54] Vulpix: I understand that. [17:00:35] Would you happen to have the default settings for for non-short URLs? I'd like to see if reverting to the default style will cause them to load. [17:01:46] Meow_: comment out $wgArticlePath = "/$1"; [17:02:23] Meow_: as I said, it's missing the rewrite rule to do the actual short url, that's why it's failing [17:02:36] from wherever you copied it, you missed a line [17:03:07] Vulpix: The issue is it worked yesterday - perplexing. [17:04:24] Meow_: there can be many causes: you changed LocalSettings.php but your browser still was loading it from the cache (until today), or you changed your server configuration and it wasn't restarted until today [17:05:18] I upgraded my mediawiki from 1.20.5 to 1.23.3. After that it shows name of variables instead of their value for example it shows instead of version. And it shows <pagetitle> instead of my current page title. you can see this problem here: http://www.fsug.ir/wiki/index.php?title=%D9%88%DB%8C%DA%98%D9%87:%D9%86%D8%B3%D8%AE%D9%87 [17:06:18] I searched but I didn't found any solution. [17:08:18] this is english version of that page: http://www.fsug.ir/wiki/index.php?title=%D9%88%DB%8C%DA%98%D9%87:%D9%86%D8%B3%D8%AE%D9%87&uselang=en [17:08:55] smmsadrnezh: Did you unpack new files on an empty directory when you upgraded? or did you extracted new files over the old ones? [17:09:13] Sounds like missing language file [17:09:27] all languages are missing apparently :( [17:09:49] Vulpix: I replaced over the old ones. [17:09:50] it may be the change from PHP to JSON messages [17:10:15] smmsadrnezh: Can you verify that the file languages/i18n/en.json exists on your install? [17:11:17] if it does exist, try running the command php rebuildLocalisationCache.php --force (in the maintinance directory of your install) [17:11:18] let me check [17:11:40] smmsadrnezh: not good. You should use an empty directory and then move other needed files (LocalSettings.php) over it. At least it prevents problems of remaining old files causing problems, or new files that can't be overwritten because they're protected/locked, etc [17:11:43] bawolff: It's on shared hosting :( [17:12:38] Don't worry about the rebuildLocalisationCache.php in that case [17:14:18] bawolff: I have this directory: /wiki/languages/i18n [17:14:46] That's good. Is the file en.json in that directory? [17:15:18] bawolff: no [17:16:14] Hmm, well that's likely your problem [17:16:18] bawolff: that's strange. [17:16:19] bawolff: It's in upgrade package. but it's not on my server!! [17:16:44]