[08:59:38] (03PS3) 10Urbanecm: Add some introduction text [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/466868 (https://phabricator.wikimedia.org/T206810) [09:00:20] (03PS4) 10Urbanecm: Add some introduction text [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/466868 (https://phabricator.wikimedia.org/T206810) [09:01:22] (03PS5) 10Urbanecm: Add some introduction text [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/466868 (https://phabricator.wikimedia.org/T206810) [09:02:11] (03PS6) 10Urbanecm: Add some introduction text [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/466868 (https://phabricator.wikimedia.org/T206810) [09:03:19] (03CR) 10Urbanecm: [C: 032] Add some introduction text [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/466868 (https://phabricator.wikimedia.org/T206810) (owner: 10Urbanecm) [09:03:36] (03Merged) 10jenkins-bot: Add some introduction text [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/466868 (https://phabricator.wikimedia.org/T206810) (owner: 10Urbanecm) [09:10:55] (03PS1) 10Urbanecm: [bugfix] Missing link in the element in login message [labs/tools/commons-mass-description] - 10https://gerrit.wikimedia.org/r/467113 [09:11:20] (03CR) 10Urbanecm: [C: 032] [bugfix] Missing link in the element in login message [labs/tools/commons-mass-description] - 10https://gerrit.wikimedia.org/r/467113 (owner: 10Urbanecm) [09:11:36] (03Merged) 10jenkins-bot: [bugfix] Missing link in the element in login message [labs/tools/commons-mass-description] - 10https://gerrit.wikimedia.org/r/467113 (owner: 10Urbanecm) [11:03:05] (03PS1) 10Zoranzoki21: Add translations for Serbian [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/467115 [14:17:25] (03CR) 10Urbanecm: [C: 032] "Fine from code-side." [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/467115 (owner: 10Zoranzoki21) [14:17:41] (03Merged) 10jenkins-bot: Add translations for Serbian [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/467115 (owner: 10Zoranzoki21) [17:57:47] !help I'm getting "MySQL server has gone away again". It's for a bot I'm rewriting, and I notice that the original bot has a comment, '/* SLOW_OK */' in the query it uses. I'm using an ORM though, should I try to insert that comment in my generated SQL? [17:57:47] audiodude: If you don't get a response in 15-30 minutes, please create a phabricator task -- https://phabricator.wikimedia.org/maniphest/task/edit/form/1/?projects=wmcs-team [18:00:20] audiodude: I am not sure if that comment is curently respected by scripts. I'd have to check. That has been used to signal to a long-runnkng query killer to leave it be. If you can pass that through your ORM, it might work [18:00:50] bstorm_: I figured it was something like that, thanks [18:06:12] audiodude: note that for long-running queries you should use the *.analytics.db.svc.eqiad.wmflabs database servers. Those have longer allowed running times than the *.web.db.etc ones [18:07:22] valhallasw`cloud: I am using the analytics.db.sv.eqiad.wmflabs database servers yes [18:22:07] audiodude: do you have the exact query that you are encountering the error? [18:24:53] I believe so [18:27:21] https://gist.github.com/audiodude/af4fee5bd4d9e86aa74357d9fee518c4 [18:28:22] zhuyifei1999_: ^ [18:28:50] and what is this %s substituted with? [18:29:05] which database is it executing on? [18:29:05] 'Portal-Class_Africa_articles' [18:29:07] ok [18:29:41] it's executing on: enwiki.analytics.db.svc.eqiad.wmflabs enwiki_p [18:29:50] k [18:29:52] * zhuyifei1999_ tests [18:30:34] actually that's the query that it reported died with "MySQL has gone away" so it was probably the query before that, that actually timed out [18:31:00] that should finish in like 0.14 secs [18:35:27] 'Category-Class_Africa_articles' [18:36:39] 59705 rows in set (11.95 sec) [18:36:47] that's similar to what I got yes [18:36:51] that shouldn't time out [18:38:17] so I guess I'm not sure what's happening here [18:38:43] because I'm consistently getting the "MySQL Server has gone away". The python program hangs for like 10 minutes and then returns that exception. [18:39:04] this time I *do* believe I am using connection pooling [18:40:10] well, then don't pool connections [18:40:23] idle connections get killed [18:40:37] I think that's the default behavior of SQLAlchemy [18:40:46] gotcha [18:49:51] I'm not sure that connection pooling is the issue, because it always runs for a few minutes and then hangs at the exact same project: 'Africa' -- https://gist.github.com/audiodude/d1352ae93db2589034dec4f08f83b41f [18:53:34] it hangs for 10 minutes and then dies with the "server has gone away" exception [19:22:49] audiodude: does it hang in the query or is the script doing processing? [19:23:05] in the latter case, you could try retrieving all rows in one go, then doing the processing [19:23:39] because iirc slowly pulling in rows from a resultset can also give that error [19:23:45] (not 100% sure about that, though) [19:27:36] interesting [19:27:50] I can't tell if it's hung in the query or in processing, I'll add more logging [19:28:16] I'm pretty sure SQLAlchemy is setup to iterate over the resultset though [19:31:16] because the Query object I'm using is a generator, not a list [20:58:13] (03PS1) 10Zoranzoki21: Use tabs instead of spaces in code [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/467234 [21:04:26] (03PS2) 10Zoranzoki21: Use tabs instead of spaces in code [labs/tools/map-of-monuments] - 10https://gerrit.wikimedia.org/r/467234