[00:00:01] e.g. the user may do delete but not truncate? [00:00:02] seems silly [00:00:06] if that is true [00:00:08] I guess it's possible [00:00:16] I dunno tho [00:00:28] * Krinkle is triaging the new workboard on https://phabricator.wikimedia.org/project/board/1329/ [00:00:42] Yep [00:00:43] The DROP privilege enables you to drop (remove) existing databases, tables, and views. The DROP privilege is required in order to use the statement ALTER TABLE ... DROP PARTITION on a partitioned table. The DROP privilege is also required for TRUNCATE TABLE. If you grant the DROP privilege for the mysql database to a user, that user can drop the database in [00:00:43] which the MySQL access privileges are stored. [00:00:45] Mysql ^ [00:01:02] Key words: "The DROP privilege is also required for TRUNCATE TABLE" [00:01:36] Granted, if we're in the updater doing cache purges, I suppose we could assume $wgDBadminuser has the permissions [00:04:10] Krinkle: Totally unrelated, but with all the work I've been doing on old-branch purging from prod, if everything's properly cleaned out (no unused branches, old branches only have static assets), the result is ~10s off the average sync* command now. [00:04:20] Touching less files -> rsync goes faster [00:04:56] RainbowSprinkles: Hm... define 'average sync' you mean syncing a directory within a single php-* subdir would also be faster? [00:05:12] I would expect the other files to not matter [00:05:16] I mean, that'll be slow because you gotta lint hella stuff [00:05:30] But not having to stat files for rsync'ing them (when they're totally unused) adds up [00:05:40] When you're talking 7000 files per branch [00:05:45] RainbowSprinkles: Right, you mean full scap will be faster because it includes all old branches [00:06:12] but sync php-123/extensions/Foo or wmf-config or docroot/noc/ or dblists/ or php-123/includes/cache/ etc.shouldn't be faster (or slower before) [00:06:15] Full scap is faster, but so are a sync-file because co-master sync is down to 7s when there's zero diff [00:06:24] Interesting [00:06:31] the rynsc masters always do scap pull? [00:06:34] (essentially) [00:06:43] No, co-master sync isn't quite a scap pull [00:06:49] It's syncing the staging dirs [00:07:08] Right [00:10:40] Put another way: [00:10:44] 393M php-1.29.0-wmf.14 [00:10:44] 13:11 394M php-1.29.0-wmf.15 [00:10:44] 13:11 395M php-1.29.0-wmf.16 [00:10:44] 13:11 4.4G php-1.29.0-wmf.17 [00:10:44] 13:11 4.4G php-1.29.0-wmf.18 [00:10:44] 13:11 4.4G php-1.29.0-wmf.19 [00:10:58] (sorry about timestamps) [00:12:17] Oh, I also fixed it so the git info cache gets done on all syncs, not just a full scap. so swats and stuff will update it properly