02:32:36 GMT <forgot_> i have a set which contains a large number of ids. how to keep only 1000 highest ones?
02:33:21 GMT <forgot_> im trying to avoid ZSET which is O(logN)
02:33:33 GMT <forgot_> and the members themselves are already sorted
07:00:57 GMT <mrtAkdeniz> hey there
07:36:56 GMT <cloudbud> how to set a parameter in redis config file at run time in redis0cli
07:39:32 GMT <cloudbud> config set parameter databases 256 is giving me error
07:39:56 GMT <minus> you can't change the number of databases during runtime, cloudbud
07:40:19 GMT <minus> also, it'd be CONFIG SET databases 256
07:40:36 GMT <cloudbud> okay
12:31:23 GMT <mightyspaj> I'm currently building an app that uses both Redis/PostgreSQL, but I'm unfamiliar with how the two will work together
12:32:14 GMT <mightyspaj> There are certain things I need/want to store in PostgreSQL—I only want to use Redis where it's necessary for the performance
12:34:25 GMT <mightyspaj> In this case, would Redis be used as a layer in front of psql? i.e you'd fetch data from psql for each table, then store that in redis?
12:35:31 GMT <cyberde> mightyspaj : You can do different things but for example to increase the performance of select jobs, you can just cache the queries and their result. But you need to make sure you can clear them once if you do an update.
12:36:39 GMT <mightyspaj> Yep, I see
12:37:01 GMT <mightyspaj> so basically, what approach you take depends on the design of your app?
12:38:59 GMT <cyberde> mightyspaj yes
12:39:41 GMT <cyberde> I use it in combination with AWS DynamoDB, basically if I do an update, I wipe the cache. But I understand that it is a bit different with a relational database.
12:41:48 GMT <mightyspaj> cyberde: alright, thanks. I think I've got it sorted out
13:28:56 GMT <willvarfar> we have a redis master and two slaves.  Here's a test run of our program: https://imgur.com/a/iGrrc  When we do a test run, sometimes memory is flat at 20% on all nodes.  Other times, the master just linearly increases ram until the test stops or we hit swap and die.  As soon as the test stops hammering the master, everything calms straight down.  The slaves don't seem to fall behind.  There are no expiries.  What can be going on?
22:11:31 GMT <zpojqwfejwfhiunz> I added memory benchmarks to my PR as requested if anyone's interested
22:11:32 GMT <zpojqwfejwfhiunz> https://github.com/antirez/redis/pull/3889
22:12:19 GMT <zpojqwfejwfhiunz> I hope to hear back soon about it, get it closed one way or another