[00:03:41] Thanks bd808 and tgr for helping me think through this. I may come back with more, hopefully better formulated, questions. [05:56:53] Can we enable mod rewrite for a project? [05:58:22] gyan: you mean, for a cloud VPS project or toolforge project? [05:58:34] toolforge [06:02:10] toolforge doesn't run apache. you are developing under lighttpd right? [06:03:21] hmm I see it's also called mod_rewrite in lighttpd https://redmine.lighttpd.net/projects/lighttpd/wiki/docs_modrewrite [06:03:41] anyways, yes you can [06:04:40] see https://wikitech.wikimedia.org/wiki/Help:Toolforge/Web/Lighttpd#URL_rewrite for some examples + help [06:05:22] Thanks for the informaton. [06:05:52] np [06:06:34] I have gone through the docs but I am wondering how can I enable it. I am running a laravel app and the routs showing 404 [06:06:45] but working on local with apache. [06:07:05] what's the tool's name? [06:08:06] Well I am just starting stuff [06:08:13] https://tools.wmflabs.org/gyan/ [06:08:16] my first guess is that the url handlers don't handle // in front of every url [06:10:11] Is anyone set up a laravel app on toolforge here [06:10:57] there are quite a few PHP tool authors here, but idk what frameworks they use. [06:11:25] I myself don't write PHP :( [06:12:42] ah [06:12:45] gyan: https://tools.wmflabs.org/gyan/index.php/mass-message [06:12:54] Yes [06:13:09] How it worked? [06:13:25] I guess what you want is index.php/ in every url [06:13:43] maybe mod_rewrite can help here [06:14:00] We can do it using htaccess I gues [06:14:08] (to omit index.php) [06:14:57] ok. whatever works for you, go for it :) [06:16:02] for htaccess we need to enable modwrite which works on apache. I am not sure about lighttpd. This is first time I am working on lighttpd [06:18:12] zhuyifei199: Thank You [06:18:17] np [06:59:54] Niharika: your userpage 'This page intentionally left not blank' reminds me of 'There is only one line of text on this page' in some places [07:45:42] Can we change url.rewrite-once = rule for toolforge project [08:01:16] Where shall I put it https://wikitech.wikimedia.org/wiki/Help:Toolforge/Web/Lighttpd#URL_rewrite [08:34:27] gyan: did you find your answer yet ? [08:35:59] No [08:41:33] https://github.com/daylerees/laravel-website-configs/blob/master/lighttpd.conf#L7 [08:42:33] but instead of the lines starting with "^/ start them with "^/gyan/ [08:42:58] and i think that ought to work... [08:43:21] Where shall I put it? [08:43:30] The web server reads any configuration in $HOME/.lighttpd.conf [08:44:26] Ahh thank you. That was my question. Will try and update. Thanks a lot for sparing time for my question. [08:46:48] gyan: ps, once you get it working, consider documenting it on that page for others ;) [08:50:50] Sure [09:48:03] Does global block have a table [10:20:09] Hauskatze: it has two [10:23:12] Reedy: thanks, the one I was looking for is in 'centralauth' not on the individual wikis. Manuel helped in in databases [13:45:36] !log snuggle migrating project to eqiad1-r [13:45:37] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Snuggle/SAL [14:11:31] !log services migrating project to eqiad1-r [14:11:32] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Services/SAL [15:03:14] Hi all :D [15:03:35] do web requests to tools.wmflabs.org end up anywhere cool for analysis, such as hadoop? or not? :D [15:04:44] addshore: no hadoop or the like to the best of my knowledge [15:09:12] thanks [15:09:53] addshore: there is a table on labsdb with this information, but it's of dubious quality [15:10:03] oooh? [15:10:14] bd808 would know more about it [15:10:15] is that documented at all? [15:15:17] !log commonsarchive migrating project to eqiad1-r [15:15:18] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Commonsarchive/SAL [15:15:42] !log wikitextexp migrating wikitextexp-expt-1002 and wikitextexp-base-1002 to eqiad1-r [15:15:43] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Wikitextexp/SAL [15:21:19] addshore, they probably go into an access.log file on the web proxy [15:21:38] nothing fancy like prod's setup though [15:40:16] hello, I've set up a new instance maps-warper3 to run Stretch on to replace maps-warper2. Like what we saw a couple of days ago, the /home directory isn't mapped to the project home directory [15:41:01] you're trying to add NFS to a new instance? [15:42:04] chippy, as I recall that project has not yet responded to confirm that NFS is required [15:42:20] the previous instance had it's /home point to home -> /mnt/nfs/labstore1003-maps/home I'm assuming that's what I should expect for new instances [15:42:33] If you are absolutely sure you need it. [15:42:38] absolutely [15:42:44] what makes you think it's needed? [15:43:02] it's where all the data is kept for warper.wmflabs.org [15:43:13] so one instance writes to it and another one reads from it? [15:43:42] oh, no. we are turning off maps-warper2 because it runs ubuntu precise. Im making a new instance to run Stretch on it [15:43:55] yes... [15:44:03] addshore: the stats table for toolforge requests exists, but it is broken. The data collector behind it needs to be changed so that it is hooked into the nginx log rotation. Right now it only gets data from a portion of the requests that happen each day because we rotate the access log based on size rather than time [15:44:03] but I'm saying what is it that requires NFS? [15:44:33] like at an application level, what is running that requires /home be mapped to NFS on this instance? [15:44:50] I know your old server had it mapped, but was that deliberate or just because historically everything did? [15:45:42] well it requires more disk space than the instance has [15:46:29] what is big that gets stored there? [15:46:30] originally on the first instance I think there was a special role that added extra space to /mnt and we used that [15:46:38] ok [15:46:41] was that tried for this one? [15:46:44] images, converted to tiffs [15:47:02] so you're using /home to store the application's data across servers? [15:47:12] then when labs moved to horizon, we used the nfs [15:47:24] its only using one server [15:47:25] chippy: if you're working on updating the maps project, it might be good for you to respond to this email thread: https://lists.wikimedia.org/pipermail/wikitech-l/2018-December/091216.html [15:47:58] well i'm only in charge of the warper which was kind of added into that project. Im not in charge of the actual project [15:48:04] ok [15:48:23] so I dont have anything to do with the old osm work , nor do I know who to approach regarding it [15:48:42] aude might know better [15:49:03] aude is long gone, as far as I know [15:49:07] ahh [15:49:47] Looking at maps-warper2 it has the scratch, dumps, and maps NFS shares mounted, but no /home. [15:49:55] so the instance uses /home/warperdata to store the converted / georeferenced images from commons which warper.wmflabs.org uses [15:50:19] bd808, yeah I couldnt find it in mtab, but a ls / -lah shows home -> /mnt/nfs/labstore1003-maps/home [15:50:33] a couple of days ago it was added when the instance moved [15:50:36] re-added [15:50:40] bd808: there are two issues: 1) Maps requires a hard-coded IP in puppet to mount NFS for /home 2) Maybe that VM doesn't need to use NFS anyway [15:50:53] Trying to resolve 2) first [15:51:22] well, Krenair is [15:52:02] 227G /home/warperdata [15:52:03] its using 230G on the NFS just for the maps (not including the things in /home) [15:52:09] :) [15:52:14] so how is that thing mapped on the existing server when the NFS server config has no home set up for maps? [15:53:37] All I know is https://gerrit.wikimedia.org/r/#/c/operations/puppet/+/479017/ [15:54:42] Krenair: "labstore1003.eqiad.wmnet:/maps on /mnt/nfs/labstore1003-maps" that's the mapping that the /home points into. That export volume has the /home and /project directories in it [15:55:15] So it's easy for me to make a similar patch for maps-warper3 if mounting /home on NFS is the right solution [15:55:43] And, actually, I'm not sure we can make this do-we-need-NFS decision on a per-instance basis… I think it has to be done per project [15:55:49] which might mean we don't have a choice here. [15:57:05] andrewbogott: there is some tantalizing hint in the "# This is a temporary change for T147657 until the misc server exports are managed by nfs-exportd" comment, but I don't know what it really means. Maybe chasemp or bstorm_ can shed some light on that. [15:57:06] T147657: Move maps share to labstore1003 - https://phabricator.wikimedia.org/T147657 [15:57:48] bd808: yeah, I saw that when I wrote that initial patch but opted for getting chippy unblocked rather than diving into the rabbit hole :) [15:58:06] I do know that the stuff from labstore1003 is due to move "soon" but the target servers are the ones that keep having that raid battery alert [15:58:24] bd808: andrewbogott I think that comment is a bit convusing but in essence maps moves ot labstore1008/9 and can be managed via nfs-exportd? wheenver [15:58:34] labvirt1019/1020? I thought they were for databases, not NFS? [15:58:35] who knew those two would be stalled for 6months+ :) [15:58:43] * andrewbogott lost [15:59:33] andrewbogott: oh! you're right. 19/20 are the bad raid battery. cloudstore1008/9 had some other blocker for a long time (10G?) [16:00:23] They were just blocked on network. They are 1g no matter what now [16:00:32] so just waiting on T209527 [16:00:32] T209527: Set up scratch and maps NFS services on cloudstore1008/9 - https://phabricator.wikimedia.org/T209527 [16:00:40] For now ...gave up on 10g [16:00:42] anyway… https://gerrit.wikimedia.org/r/#/c/operations/puppet/+/479464/ [16:04:17] chippy: putting all that data from an active server in /home is silly, but the mounts on that VM should be back to the way you'd expect them [16:05:00] thanks! [16:17:42] Looking back in the tickets T112641 was an issue with the labs::lvm::srv added disk. The system was storing less data at that time generally though, and from what I remember, I've always seen /home as being project home. I'll be open to setting up the server in a better way in the future, hopefully when / if the other, older instances get sorted [16:17:43] T112641: maps-warper /mnt vbd partition errored, turned read only and went missing after reboot - https://phabricator.wikimedia.org/T112641 [16:44:30] bd808: is it your understanding from T204566 that I can now delete mw-expt and mw-base ? [16:44:30] T204566: cloudvps: wikitextexp project trusty deprecation - https://phabricator.wikimedia.org/T204566 [16:45:15] andrewbogott: yes! I think we can cleanup all the legacy region bits for the wikitextexp project [16:45:27] ok [16:45:39] including the wildcard DNS stuff that I think only we can see/change in Horizon [16:45:58] we == "admin" project members I think [16:46:00] hm, ok — I'm not sure I understand that but I will look [16:46:17] !log wikitextexp deleting mw-expt and mw-base VMs [16:46:19] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Wikitextexp/SAL [18:13:10] !log tools.wikiloves deleted local crontab on tools-bastion-03 (please see https://wikitech.wikimedia.org/wiki/Help:Toolforge/Grid#Creating_a_crontab) [18:13:11] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.wikiloves/SAL [23:07:38] Anybody here using java on the Toolforge Kubernetes cluster? T210827 has some questions about how to get tomcat running there. [23:07:39] T210827: Help to run Java8 web-app - https://phabricator.wikimedia.org/T210827