[00:21:57] 10Cloud-VPS (Project-requests), 10Reading List Service, 10Reading-Infrastructure-Team-Backlog: Request creation of reading-lists VPS project - https://phabricator.wikimedia.org/T178110#3695567 (10Tgr) [00:27:23] 10Data-Services, 10cloud-services-team (FY2017-18), 10DBA, 10Goal: Migrate all users to new Wiki Replica cluster and decommission old hardware - https://phabricator.wikimedia.org/T142807#3695569 (10bd808) [00:33:34] 10Data-Services, 10cloud-services-team (FY2017-18), 10DBA, 10Goal: Migrate all users to new Wiki Replica cluster and decommission old hardware - https://phabricator.wikimedia.org/T142807#3695570 (10bd808) [00:39:11] bd808: There appears to be a disk problem on my instance. [00:39:26] what's the problem? [00:39:29] cyberbot-exec-iabot-01 should have 120GB of available disk space [00:39:50] When I ran GZIp on my DB dump, I got a "Out of disk space" error [00:40:11] But when I ran du on the entire VM, I see only 14.6GB of actual data. [00:40:26] Am I missing something [00:40:34] Cyberpower678: do you have the role applied that allocates the extra disk? Are you gzipping to the /srv partition where that disk is mounted? [00:40:55] No clue [00:41:03] I was gzipping to my home directory [00:41:21] https://wikitech.wikimedia.org/wiki/Help:Adding_Disk_Space [00:41:43] your $HOME is alwasy going to be on the default system disk [00:42:00] I see. [00:42:11] to get the quota allowed space beyond the system disk you need to use that Puppet role [00:44:10] bd808: I'm adding the roles. [00:44:36] bd808: mutante has helped a lot in understanding how puppet works. :D [00:44:49] he's a pretty awesome guy :) [00:46:41] :-) [00:49:03] 10Data-Services, 10cloud-services-team (FY2017-18), 10DBA, 10Goal: Migrate all users to new Wiki Replica cluster and decommission old hardware - https://phabricator.wikimedia.org/T142807#3695588 (10bd808) [00:49:04] 10Data-Services, 10cloud-services-team (Kanban), 10DBA, 10User-bd808: Create and announce timeline for shutting down labsdb100[13] - https://phabricator.wikimedia.org/T175086#3695586 (10bd808) 05Open>03Resolved Announced at https://lists.wikimedia.org/pipermail/cloud-announce/2017-October/000005.html [00:50:19] bd808: I added the role and ran puppet agent, but no new volume [00:50:28] At least none the size of 120 GB [00:50:38] what's the instance name: [00:50:40] ? [00:50:57] cyberbot-exec-iabot-01 [00:51:33] I only see a lost+found [00:51:40] /dev/mapper/vd-second--local--disk 140G 60M 132G 1% /srv [00:52:22] Oh oops. I queried with the wrong command. [00:52:26] So /srv is a 140G partition with 132G of usable space [00:52:52] is the partition call "lost+found"? [00:53:16] no, that is a unix directory that is used to track damaged files [00:53:24] the partition is /srv [00:53:29] Oh [00:53:37] Oh cool. [00:54:04] there is a /lost+found directory too. It's normal stuff [00:55:06] Okay [00:56:46] (03PS165) 10Ricordisamoa: Initial commit [labs/tools/wikidata-slicer] - 10https://gerrit.wikimedia.org/r/241296 [00:58:37] * Cyberpower678 tries gzip again [01:01:38] (03CR) 10Ricordisamoa: [C: 04-2] "PS165 uses the 'hidden' class to hide the alert div on the save page" [labs/tools/wikidata-slicer] - 10https://gerrit.wikimedia.org/r/241296 (owner: 10Ricordisamoa) [01:04:45] Cyberpower678: basically any root directory of an ext4 filesystem, be default, has a lost+found directory [01:05:13] zhuyifei1999_: so how much space do the VMs have without SRV [01:05:18] fsck can put files in it when it files lone inodes [01:05:35] by default / has around 20 gigabytes [01:05:38] when it what? [01:05:58] all other disk space are unallocated [01:06:08] https://wikitech.wikimedia.org/wiki/Help:Adding_Disk_Space [01:06:33] zhuyifei1999_: I saw that but "fsck can put files in it when it files lone inodes" [01:06:43] What's files lone inodes? [01:06:52] you can create new partitions on it and mount it (which that srv puppet role does automatically) or extend / [01:06:54] uh [01:07:07] basically files are three parts [01:07:11] 1. the data [01:07:55] 2. the inode, marking where the data are stored, the attributes (eg owner, access time, etc.) [01:08:03] 3. the directory listing [01:08:37] the file name is not saved in the inode, but in the directory listing [01:08:45] this is why you can have hard links [01:09:04] which are basically two directory listings pointing to the same inode [01:09:55] zhuyifei1999_: so how do I extend /? [01:10:03] Instead of mount srv? [01:10:07] if there is a file (an inode), but there is no directory listing pointing to the inode, no one knows what the inode's filename is [01:11:02] and the file will not be easily accessible. fsck can find such inodes and put it under lost+found directory of the corresponding filesystem [01:11:15] https://wikitech.wikimedia.org/wiki/Help:Resize_root_partition [01:11:21] ^ to extend / [01:11:45] but beware it may not be well-supported [01:12:19] there is no puppet role to do it automatically [01:12:36] uhhhh [01:12:43] I think I'll using symlinks [01:13:41] huh? [01:15:03] zhuyifei1999_: in regards to my new DB instance [01:15:13] I need it for the disk space and processor [01:16:12] I mean idk what you mean by using symlinks. do you mean moving the mariadb storage into /srv and symlink it? [01:16:23] But I would take it all of the DB stuff will safe to the root disk and not the srv mount [01:16:31] Yea [01:16:43] *save [01:16:52] yeah. that's the best thing to do [01:16:58] https://github.com/wikimedia/puppet/blob/production/modules/quarry/manifests/database.pp#L6 [01:17:12] oh nice [01:17:29] zhuyifei1999_: mutante setup my own roles, which gave me insights on how to write them in the first place. :D [01:18:09] k [01:18:10] zhuyifei1999_: Oh nice. Great idea. Thanks for the advice. [01:18:13] np [01:18:34] idk how well that 'mysql' module works with mariadb though [01:18:49] https://phabricator.wikimedia.org/T165625 [01:33:35] Cyberpower678: so, what path do you want to use as data dir? [01:33:44] /srv/mysql/data sounds ok>? [01:33:53] Yes please [01:34:23] and you want it to listen on all interfaces, right [01:34:27] Yes [01:38:42] Cyberpower678: what is the current data dir , do you know? [01:38:57] No I do not. We just installed it. [01:39:34] Documentation says it's /var/lib/mysql [01:39:50] ok, and it's owned by mysql:mysql , right [01:41:37] change coming [01:43:34] unfortunately we dont see that here [01:44:30] mutante: yes they all belong to mysql [01:44:40] Cyberpower678: https://gerrit.wikimedia.org/r/#/c/385115/2/modules/profile/manifests/cyberbot/db.pp [01:46:03] So how do these changes get applied to the already set up instance? [01:46:13] Reboot, reinstall, etc? [01:46:32] puppet agent [01:46:39] That's it? [01:46:45] yeah [01:46:50] Cool [01:47:08] the agent will apply all the changes [01:47:35] Cyberpower678: and now .. you can run it [01:48:12] bd808: since everyone has to update their URIs now anyway, how hard would it be to set up aliases like enwiki.web.db.toolforge? [01:48:21] Cyberpower678: it should create /srv/mysql/data first [01:48:31] enwiki.web.db.svc.eqiad.wmflabs is super cryptic [01:49:17] mutante: running puppet [01:49:27] Toolforge isn’t right. These dbs are for all of Cloud [01:50:39] are you going to even remove "labs" from DNS at some point in the future? [01:51:06] We talked about switching wmflabs for wmcloud but chase wasn’t ready to commit to a zone naming scheme [01:51:14] gotcha [01:52:39] We are waiting for the wildcard letsencrypt certs to introduce the Toolforge.org and wmcloud.org TLDs [01:53:31] oh, you bought those? aha [01:54:08] that makes sense for external domains [01:54:21] but the internals don't use TLS anyway, do they? [01:54:25] puppet seems to have hung up, mut [01:54:28] mutante: [01:55:16] bd808: ^ [01:55:31] Cyberpower678: what do you mean by "hung up" [01:55:48] i have never seen that happen ..like ever [01:55:49] It's not outputting a thing, but it's clearly running [01:55:54] either an error or finishing [01:56:03] Tgr: we want a sane service naming scheme. There are other things we will be using the .svc.eqiad.wmflabs domain for. [01:56:49] And there will “soon” be more zones that need differentiation [01:56:59] Cyberpower678: remind me of the instance name please [01:57:01] mutante: It's not doing anything [01:57:07] cyberbot-db-01 [01:57:17] Soon being relative to the heat death of the universe ;) [01:58:06] mutante: can I kill it? [01:58:09] Cyberpower678: cant confirm, it's running [01:58:12] and applying all the changes [01:58:18] works as expected :) [01:58:23] it just created your data dir [01:58:29] and changed your mysql config [01:58:42] Cyberpower678: no, don't kill it [01:58:43] My CLI is still blocked [01:58:53] please paste your commandline [01:58:54] I totally understand both of the why-not questions though. They are things I asked myself [01:59:20] Cyberpower678: if you were already running puppet, i wouldnt have been to able to do it :) [01:59:48] killed it for you, heh [02:00:11] Confirmed it's killed [02:00:22] Cyberpower678: see how you have /srv/mysql/data now? [02:00:44] Yes. :-) [02:00:54] Cyberpower678: and the datadir in /etc/mysql/my.cnf is set to that [02:01:05] nothing was manual [02:01:39] well, i think we have to restart it [02:01:42] and that fails :) [02:01:51] looks why [02:05:54] Cyberpower678: sigh, that existing mysql server class has an issue .. i am looking and cleaned up everyhting.. hold on [02:06:13] Depends: mariadb-server-5.5 (= 5.5.34+maria-1~precise) [02:06:15] precise :p [02:06:27] ... [02:06:36] That's ancient [02:06:53] It would probably explain ERROR 2003 (HY000): Can't connect to MySQL server on '10.68.18.167' (111) [02:07:03] Cyberpower678: that's because i removed it [02:07:07] give me a minute [02:07:10] Okay [02:10:58] damn, it's because that class installs APT sources for mariadb on precise [02:11:11] which of course conflict badly on stretch [02:11:55] maybe we should just copy the relevant parts and not use that mysql class [02:13:27] mutante: I don't get it. Why is the latest version of MariaDB relying on Precise? [02:14:01] Cyberpower678: because mysql::server::package says so [02:14:19] 6 if $package_name =~ /mariadb/ { [02:14:19] 7 apt::repository { 'wikimedia-mariadb': [02:14:19] 8 uri => 'http://apt.wikimedia.org/wikimedia', [02:14:19] 9 dist => 'precise-wikimedia', [02:14:42] hardcoded precise in there [02:14:55] that wont work anywhere nowadays [02:15:09] basically it means that the class cant be used [02:15:10] No kidding [02:15:21] or we start making that flexible of course [02:15:55] there is mariadb in normal debian repos now, we don't need to use our own [02:16:21] and then we get version 10.1 instead of 5.x [02:17:43] Cyberpower678: class mysql isn't used by anything then :) [02:17:48] in prod we use module mariadb [02:18:26] I se [02:18:28] *see [02:18:35] thinks about the easy fix for you [02:19:10] bd808: Any chance we can drop the disk limit on cyberbot-exec-iabot-01 from 120GB to maybe 40GB, and divert that to cyberbot-db-01 [02:21:59] Cyberpower678: not really, no. It would require making custom ‘flavors’ for both and the building new vms from scratch. [02:22:15] Oh [02:23:46] * bd808 is on vacation, drinking beer, and watching his Cubs. Responses may stop at any time. [02:24:54] Cyberpower678: https://gerrit.wikimedia.org/r/385119 [02:25:10] bd808: enjoy the beer. [02:26:59] mutante: awesome. I'm getting the DB dump onto that node. [02:27:22] Cyberpower678: it's not merged:) [02:27:43] Yea, but the upload will take another 30 minutes to finish [02:28:19] ;-) [02:30:51] i will change it to make it more fool-proof that it cant touch anything else [02:34:17] meh, quarry is the only thing using that to install mariadb [02:35:02] i wonder how it can work with precise sources [02:44:41] Hmmm... [02:44:55] The dump is uploaded and ready to go. [02:46:15] mutante: ^ [02:46:59] (03PS166) 10Ricordisamoa: Initial commit [labs/tools/wikidata-slicer] - 10https://gerrit.wikimedia.org/r/241296 [02:52:19] Cyberpower678: i wanna know what quarry runs on now .. [02:52:29] i really dont want to touch that in any way [02:52:39] it's the only thing using this [02:52:55] (03CR) 10Ricordisamoa: [C: 04-2] "PS166 changes save step to avoid page reload" [labs/tools/wikidata-slicer] - 10https://gerrit.wikimedia.org/r/241296 (owner: 10Ricordisamoa) [02:53:22] If it's running precise, then I'll break out Kali Linux. XD [02:53:35] o.O? [02:54:11] zhuyifei1999_: What? [02:54:34] mutante: what's the command to check mysql version? [02:54:53] Cyberpower678: how do you break precise with kali? [02:55:04] * zhuyifei1999_ is unaware of that [02:55:24] zhuyifei1999_: I was joking, since Kali Linux is primarily used for hacking. [02:55:38] zhuyifei1999_: mysql -V ? what i would really like to know is "lsb_release -a" though whereever quarry runs [02:55:57] zhuyifei1999@quarry-main-01:~$ lsb_release -a [02:55:57] No LSB modules are available. [02:55:57] Distributor ID: Ubuntu [02:55:57] Description: Ubuntu 14.04 LTS [02:55:57] Release: 14.04 [02:55:58] Codename: trusty [02:56:21] zhuyifei1999@quarry-main-01:~$ mysql -V [02:56:21] mysql Ver 15.1 Distrib 5.5.39-MariaDB, for debian-linux-gnu (x86_64) using readline 5.1 [02:56:55] interesting [02:56:58] I think the two workers are jessie but this main (db, redis, web, etc) is trusty [02:56:58] zhuyifei1999_: cat /etc/apt/sources.list.d/wikimedia-mariadb.list [02:57:03] ^ that pleaes [02:57:16] zhuyifei1999@quarry-main-01:~$ cat /etc/apt/sources.list.d/wikimedia-mariadb.list [02:57:16] deb http://apt.wikimedia.org/wikimedia precise-wikimedia mariadb [02:57:16] deb-src http://apt.wikimedia.org/wikimedia precise-wikimedia mariadb [02:57:21] how does that work :) [02:57:27] no idea [02:58:19] ok, ehm.. [02:58:39] Cyberpower678: I wan't aware of any extremely exploitable bugs in precise. that was why I asked. kali at least needs some sort of bugs to exploit [02:58:47] is torn between not using the mysql module at all ... or tryign to fix it [02:59:04] zhuyifei1999_: it was only a joke. :-) [02:59:07] k [02:59:38] * Cyberpower678 is using some of his free time to learn about ethical hacking. [03:00:15] And of course, making my own stuff secure [03:00:36] nice :) [03:04:06] mutante: https://phabricator.wikimedia.org/T165625 fyi [03:04:59] Cyberpower678: i'm fixing it by .. not using that module [03:05:10] Ok [03:05:24] zhuyifei1999_: thanks! i agree there is no future for it probably :) [03:05:25] So are we not using puppet to setup a DB? [03:05:37] zhuyifei1999_: it looks like that comes down to "make quarry stop using it" [03:05:55] Cyberpower678: oh, we are definitely using puppet, but we are not using mysql::server [03:06:02] Ah [03:06:07] well, I'm not familiar with all the puppet modules to change that [03:06:19] Because I am now getting used to Puppet [03:07:19] mutante: could you cc me to whichever patch you create to "make quarry stop using it"? [03:07:40] I do hope there's someway to test the patch before breaking everything though [03:09:42] zhuyifei1999_: so i just decided to not try and fix the mysql::server class [03:09:55] because i did _not_ want to touch quarry in any way [03:10:07] um ok [03:10:29] what i am doing now is going back to just putting the config file in place with puppet [03:10:31] well quarry still has to move to a new class, if possible [03:10:35] and installing the package [03:10:36] ok [03:10:42] because i want the fix for Cyberpower678 right now [03:10:51] ok [03:10:58] while i can guarantee i am not touching anythng else [03:11:07] k [03:13:10] * Cyberpower678 gives mutante a beer [03:13:34] * mutante cools it and is planning to open it in 5 min [03:13:36] uploads [03:13:48] * mutante is already missing the bots here [03:14:02] because with them i wouldnt have to keep saying that or manually link to gerrit [03:15:54] CC'ed you on the gerrit change [03:16:30] https://gerrit.wikimedia.org/r/#/c/385125/ [03:16:32] (03PS167) 10Ricordisamoa: Initial commit [labs/tools/wikidata-slicer] - 10https://gerrit.wikimedia.org/r/241296 [03:17:57] Cyberpower678: fixed, you can try it now [03:18:05] mariadb 10.1 (not 5) is installed now [03:18:09] puppet is happy [03:18:14] ERROR 2003 (HY000): Can't connect to MySQL server on '10.68.18.167' (111) [03:18:37] please add the commands you are running [03:18:38] From cyberbot-exec-iabot-01 [03:18:39] because: [03:18:39] mysql [03:18:39] Welcome to the MariaDB monitor. Commands end with ; or \g. [03:19:04] (03CR) 10Ricordisamoa: [C: 04-2] "PS167 clears the interval for the conflict handler upon starting the save step" [labs/tools/wikidata-slicer] - 10https://gerrit.wikimedia.org/r/241296 (owner: 10Ricordisamoa) [03:19:13] you can't expect a connection from another host to work yet [03:19:16] before adding GRANTs [03:19:35] I thought it would give me an error saying it won't allow me to connect. [03:19:48] didn't it? [03:19:57] It gave me ERROR 2003 (HY000): Can't connect to MySQL server on '10.68.18.167' (111) [03:20:13] The server apparently didn't respond at all. [03:20:22] what's the difference between "me an error saying it won't allow me to connect" and "ERROR 2003 (HY000): Can't connect to MySQL server " [03:21:16] The last time we tried to connect when it was working, it said this user isn't allowed to connect. It didn't say it can't connect. The former confirmed a response from the server [03:21:30] ERROR 2003 is a mysql error though [03:21:54] That latter implies it didn't. Unless the messages got vaguer in the newer versions [03:22:11] mutante: well wouldn't the mysql app return a mysql error. [03:22:15] "didnt respond at all" != "responds with error 2003" [03:22:41] Cyberpower678: do you still have your GRANTS line from earlier? [03:22:49] or just give me user/password :) [03:22:58] or i set one and give it to you [03:23:20] also, weren't you going to import it .. from localhost [03:23:31] and that was the whole point of uploading the .sql file [03:27:14] running the import of the dump is a step before granting access to it [03:27:30] so that a DB exists that you can give GRANTs to [03:28:39] so the name of the database is "s51059__cyberbot" with 2 underscores? [03:29:16] !log cyberbot copying mysqldump file to /srv/; gunzip'ing it; creating database s51059__cyberbot ; importing mysqldump [03:29:20] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Cyberbot/SAL [03:31:38] Cyberpower678: it's only 12G unzipped [03:31:46] didnt you say even more? [03:38:05] bd808: quick question. Isn't a large instance granted 80GB of disk space. Mine only got 60GB in the SRV [03:39:45] ok, i purged all the things , removed that precise file, let puppet install it all again .. [03:42:41] Cyberpower678: import ongoing .. sice in /srv/mysql is increasing [03:42:44] size [03:42:57] root@cyberbot-db-01:/srv/mysql# watch du -s . [03:43:00] ^ try that [03:43:10] mutante: looks like it's writing to the correct folder now. [03:43:32] Cyberpower678: last attempt i hadn't removed the remnants of the old version and restarted [03:43:39] now i did [03:43:47] That'll do it. ;p [03:44:11] Cyberpower678: try this: become root, and then run "screen -x" [03:44:29] do you see my command then ?:) [03:44:50] apparently yes :) [03:44:54] since you are hitting keys [03:44:57] dont:) [03:45:01] let it run [03:45:28] Yes, I do, but that was me scrolling on the trackpad. I was trying to scroll the terminal up. [03:46:02] How do I get out. [03:46:10] i am trying to clean that up but it adds more [03:46:15] trackpad is not gonna work [03:46:18] I don't want to accidentally kill the import [03:46:39] Cyberpower678: Ctrl+A +d [03:46:43] mutante: sorry, I'm not even try to add more. I switched between windows. [03:46:57] you hit Crtl+A, then you let go of both keys, and you hit "d" [03:47:12] Ooops [03:47:15] it should say "detached" and bring you back [03:47:22] did it? [03:47:30] I hit Crtl+A then Crtl+D [03:47:35] Yes [03:47:53] it's only a "d" for detach [03:48:03] all screen commands start with Ctrl+A [03:48:06] Apparently both work. :-) [03:48:09] and then some letter [03:48:23] so now you know how to get out of it [03:48:53] with "screen -x" you can see it again [03:49:10] this means i can now disconnect and stop working .. [03:49:17] while the import keeps running [03:49:20] and you can watch it :) heh [03:49:21] Cool [03:49:30] Well it's 23:49 here [03:49:52] you are on the US East Coast? [03:49:57] Yep [03:50:21] well, here is what you do.. just disconnect and sleep.. when you come back ssh to it and "screen -x [03:50:29] and see it finished :) [03:50:45] Or see a new error. :p [03:51:03] yea, but unlikely at this point [03:51:50] you might want to screen -d -r if nobody is attacked to that screen session [03:51:57] mutante: BTW, the previous attempt indeed wasn't working anyway. [03:51:59] ERROR 1130 (HY000): Host 'cyberbot-exec-iabot-01.cyberbot.eqiad.wmflabs' is not allowed to connect to this MariaDB server [03:52:20] That definitely confirms the server is responding to my connection attempt [03:53:11] Cyberpower678: right, it's a different error [03:53:17] but both are mysql errors [03:53:56] one is the _client_ saying i cant even connect [03:54:04] ann one is the server saying you arent allowed [03:54:08] mutante: Yes, but the other one was a complete failure to even connect, or get a response from the server. I get that error when my tunnels to the DB aren't open. [03:54:29] it was because the new package was installed but there was no restart [03:54:35] mutante: ah. [03:54:39] so it didn't have the "bind-address 0.0.0.0" part [03:54:45] and wasnt listen besides localhost [03:54:49] Yea, that was my point. :p [03:54:53] you are right then [03:55:35] * Cyberpower678 goes to bed. His eyelids weigh as much as anvils [03:56:19] mutante: good night and thanks for your help. :D [03:58:27] Cyberpower678: good night :) leaving too [07:09:49] (03CR) 10jenkins-bot: Localisation updates from https://translatewiki.net. [labs/tools/heritage] - 10https://gerrit.wikimedia.org/r/385139 (owner: 10L10n-bot) [07:17:17] 10Tool-wikiloves: Generalize wikiloves tool beyond the "Wiki Loves X" format - https://phabricator.wikimedia.org/T173005#3695693 (10JeanFred) [08:23:54] (03PS1) 10Lokal Profil: Add WLM favicon [labs/tools/heritage] - 10https://gerrit.wikimedia.org/r/385151 [08:28:39] (03CR) 10Jean-Frédéric: [C: 032] Add WLM favicon [labs/tools/heritage] - 10https://gerrit.wikimedia.org/r/385151 (owner: 10Lokal Profil) [08:30:55] (03Merged) 10jenkins-bot: Add WLM favicon [labs/tools/heritage] - 10https://gerrit.wikimedia.org/r/385151 (owner: 10Lokal Profil) [08:32:30] (03CR) 10jenkins-bot: Add WLM favicon [labs/tools/heritage] - 10https://gerrit.wikimedia.org/r/385151 (owner: 10Lokal Profil) [10:57:57] 10Cloud-VPS, 10Graphite: grafana-labs often fails to generate graphs with c.datapoints is undefined - https://phabricator.wikimedia.org/T177747#3696296 (10fgiunchedi) >>! In T177747#3675523, @hashar wrote: > Hurrah it works fine on https://grafana-labs.wikimedia.org/dashboard/db/labs-project-board?orgId=1&var-... [11:39:03] 10PAWS: Server error (500) while trying to download files from Commons from PAWS - https://phabricator.wikimedia.org/T178567#3696395 (10Chicocvenancio) [14:23:44] 10Cloud-VPS (Quota-requests): Request increased quota for mwstake Cloud VPS project - https://phabricator.wikimedia.org/T178012#3696666 (10MarkAHershberger) We've moved to WMFLabs... I'll turn off the old AWS instance tonight [15:04:40] chasemp: I need your assistance [15:09:04] Cyberpower678: chase is out for a long weekend [15:10:42] andrewbogott: I'm trying to mysql into cyberbot-db-01.cyberbot.eqiad.wmflabs from tools-login.wmflabs.org [15:10:47] It won't connect [15:12:06] * andrewbogott tries it [15:12:58] do you get an error or does it just hang? [15:14:10] It eventually times out and says it can't connect. [15:14:41] ok [15:14:55] I assume this isn't by design — will dig around a bit [15:15:29] andrewbogott: I can connect just fine from within instances in my project [15:15:46] so this is the database you made yourself, right? [15:16:01] Have you already changed your security group to allow access to things outside the project? [15:16:39] andrewbogott: mutante suggested I needed to. [15:16:46] But I have no idea what to change [15:17:03] I have a webservice on tools that needs access to this DB. [15:17:19] so by default pretty much all ports are closed to access outside of a project [15:17:40] so you'll need to open up port 3306 (probably) to 10.0.0.0 (which means everything in the cloud) [15:17:51] have you modified a security group before? Or applied one? [15:17:56] probably only one or 2 ips [15:18:09] if he has dedicated clients, too [15:18:18] andrewbogott: I have not, but I have the interface open on Horizon [15:18:32] jynus: since he's trying to connect from tools-login, I assume that he's going to access it from the grid at some point? [15:18:36] oh [15:18:38] is he? [15:18:52] Cyberpower678: I thought you had dedicated clients? [15:18:57] Cyberpower678: can you tell me what your long-term plans are for where/what is going to need to access this? [15:18:59] other VMs [15:19:27] https://tools.wmflabs.org/iabot needs access. [15:19:27] i was hoping it was possible without the "all of 10.0.0.0" part [15:19:38] Also the web workers on the grid need access too. [15:19:40] and we'd just need the right IP [15:20:03] if it's something running on the grid I don't think we can be more selective [15:20:24] for some reason i first assumed it's like this: tool -> exec node -> db [15:20:32] but then i learned he needs both [15:20:42] connections to db from exec node AND from a tool [15:20:54] Cyberpower678: is that really true? ^ [15:21:16]