06:34:53 GMT <badboy_> AsenLenBeyraklar: that's purely client-side
06:35:07 GMT <badboy_> you can pass it around, but you can't reset it, so re-using is ... not really possible
06:35:16 GMT <badboy_> (why didn't we add a way to reset it? o_O)
11:20:58 GMT <aerendil> Hey guys :) I'm running an experiment on Redis using it as a message queue where I'd need BLPOP to unlock ms (300-400ms for ex). I candidly copied the BLPOP command in a custom one and added the unit UNIT_MILLISECONDS to the timeout function (which already supports ms) and changed the HZ of the server to HZ 100 (as I have around 1-2k clients connected). However I have lower perf than expected (unblocking happens in 600ms to 1s). Any
11:21:48 GMT <aerendil> Or maybe have an idea of how I could improve the precision there (I admit I did not dive to deep yet in the server cron method and the unblock clients logic - so I'd appreciate pointers if any)
15:39:12 GMT <willvarfar> howdy all!  I am using the rpoplpush 'reliable queue' trick.  I have a lua script to rpoplpush upto N items and return them, and this takes no time at all.  And I have another script that I use to call lrem on all the items later....
15:39:33 GMT <willvarfar> however, the lrem is taking a really really long time.  I am typically rpoplpushing 100K items at a time
15:40:24 GMT <willvarfar> I'm guessing the problem is the complexity of lrem is O(n), so lremming M items takes O(n*m / 2) average?
15:54:30 GMT <minus> where are you lreming them from?
15:54:35 GMT <minus> can't you just drop the whole key?
16:01:13 GMT <willvarfar> I have several workers sharing a 'rollback' queue
16:01:15 GMT <willvarfar> hmm
16:15:21 GMT <willvarfar> minus: will see if I can give each worker its own progress list so it can be cheaply deleted instead
16:35:15 GMT <DVS[1]> Hi, I'm using redis 3.2.3 in a cluster and keep getting an error Partial "resynchronization not possible (no cached master)". help would be very much appreciated
16:37:11 GMT <DVS[1]> "MASTER aborted replication with an error: ERR Can't SYNC while not connected with my master"
21:34:52 GMT <harej> I am planning on using Redis to store a bunch of key-value pairs in the form of appname:filename:yyyymmdd. I am not worried about the values – the values are integers. I am worried about the keys, since many of them are repetitive. Is there anything I can do to more intelligently name my keys, or should I trust that Redis will work its optimization magic?
21:37:01 GMT <harej> One strategy I could consider is lopping off the date part and store all the values for a certain date within a hash table. Would that get me less memory usage? Would it cost me in terms of performance?
21:37:06 GMT <harej> err
21:37:16 GMT <harej> store all the values for a certain *file* in a hash table