[07:41:44] <elukey>	 gtirloni: thanks a lot!
[07:56:42] <elukey>	 everything works super fine, I owe you a big one!
[09:11:00] <gtirloni>	 elukey: yw :)
[12:10:13] <bawolff>	 I'm trying to figure out why my job seems to randomly fail when started via jsub from crontab (but doesn't fail when run manually, even when running the jsub line directly from the login host)
[12:10:23] <bawolff>	 I tried to use the qacct command to get info
[12:10:32] <bawolff>	 But the command just seems to hang and not return anything
[12:10:40] <bawolff>	 Is there some trick to making that command work?
[12:12:18] <bawolff>	 The command i'm trying is qacct -j clone -b 201811101139 -e 201811101150
[12:20:38] <bawolff>	 Well i guess it was oom, as increasing memory limits seemed to fix things
[12:20:45] <bawolff>	 for my job that is
[12:40:38] <bawolff>	 musikanimal: https://tools.wmflabs.org/sql-optimizer?use=enwiki_p&sql=select+rev_user%2C+rev_id+from+revision+LIMIT+30%3B is super cool. I like that it gives performance tips. Although its mistaken about how many rows are read 
[12:42:55] <bawolff>	 In that a LIMIT with no where clause, it definitely doesn't read all of them
[12:43:30] <bawolff>	 Even weirder is
[12:43:32] <bawolff>	 https://tools.wmflabs.org/sql-optimizer?use=enwiki_p&sql=select+rev_id+from+revision+order+by+rev_timestamp+desc+limit+4%3B
[12:43:41] <bawolff>	 Which the explain claims is a filesort
[12:43:47] <bawolff>	 Maybe something weird with views
[12:43:54] <bawolff>	 but the query is clearly not filesorting
[13:18:44] <bawolff>	 What'd be really cool is if it showed available indexes too
[17:45:33] <wm-bot>	 I will let you know when I see bawolff and I will deliver that message to them
[17:45:33] <zhuyifei1999_>	 @notify bawolff for qacct the ticket is https://phabricator.wikimedia.org/T182451