[01:20:20] Hi team. I'd like to connect to this analytics replica [s3-analytics-replica.eqiad.wmnet] but I don't have credentials yet and keep getting the error "ERROR 2002 (HY000): Can't connect to server on 's3-analytics-replica.eqiad.wmnet' (115)". Please assist. [01:38:15] My username on Toolsforge is: johnmwangiwiki [02:48:44] Petscan is down again [03:22:50] I saw this last message and was transported to 2014 [03:40:16] Some things never change [09:52:42] johnmwangiwiki: I've sent you an email as you are not online. tl;dr you have to use s3.analytics.db.svc.wikimedia.cloud [10:25:09] Hi Francesco. I saw your response [10:25:10] ``` [10:25:10] You are using s3-analytics-replica.eqiad.wmnet, that is the address for Mediawiki Replicas and it's not accessible from Toolforge. [10:25:11] You have to use the Wiki Replicas address instead: s3.analytics.db.svc.wikimedia.cloud [10:25:11] For example, this command works: [10:25:12] mariadb --defaults-file=$HOME/replica.my.cnf -h s3.analytics.db.svc.wikimedia.cloud -e "SHOW DATABASES;" [10:25:12] ``` [10:25:13] Many thanks for the clarification. I've done some further digging and realised that I need to be connecting to the hosts referenced on this commit [https://gerrit.wikimedia.org/r/c/operations/puppet/+/1048385]. How do I access those MariaDB instances? [10:27:55] johnmwangiwiki: you can't connect to those instances from toolforge [10:29:10] How do I go about requesting for their access? [10:29:23] "dbstore" instances contain unredacted personal data, so the access is restricted to very few people [10:29:54] dhinus I see. Thank you! [10:29:56] the wiki replicas were created exactly to provide access to the same data, what do you need to access exactly? [10:30:13] is there a specific thing that you cannot find in the wiki replicas? [10:30:27] we welcome requests to add more data to wiki replicas, unless it's personal or private information [10:31:44] dhinus Not certain data per se but I need to an instance where binlog has been enabled. [10:32:29] This is so that I can continue with the CDC experimentation that's being mentioned on that commit. [10:35:08] I might be able to proceed if there are other analytics replicas with binlog enabled that don't contain personal data. [10:35:51] I see! I think that's a legitimate request, I would suggest you to create a Phabricator task describing briefly what you want to do, and asking to enable binlogs to the existing "wiki replicas" (e.g. s3.analytics.db.svc.wikimedia.cloud) [10:37:37] I think there are some security concerns, because binlogs would expose some private data that is currently filtered out [10:37:44] Great! Will do that shortly. [10:38:21] yeah I don't think we can publish binlogs from the WMCS wiki replicas for that reason [10:52:52] yes, unfortunately we cannot just "turn the binlogs on", but the CDC use case is very interesting, so it would be good to have a request for that tracked in Phabricator [10:57:49] I also wonder if the work on the "Event Platform" will include some publicly-accessible endpoints [10:58:27] I don't know much about the plans there, but it feels like a more natural way to access "events" without relying on wikireplicas dbs [11:13:34] dhinus I've created the ticket on Phabricator and included you as a subscriber. [18:31:58] !log lucaswerkmeister@tools-bastion-13 tools.ranker deployed 43647bd6bb (l10n updates: sr-ec) [18:32:00] Logged the message at https://wikitech.wikimedia.org/wiki/Nova_Resource:Tools.ranker/SAL