00:23:52 GMT <votz> How can the original ttl value be retrieved for a key? That is, if a key was stored with 'setex foo 1000 "blah"', how can the original ttl 1000s be retrieved from Redis?
06:33:14 GMT <volga> help
06:33:32 GMT <volga> hhello
06:33:34 GMT <volga> hello
06:34:05 GMT <volga> I am not able to get ops high operation per second in high end servers smp environments
06:34:24 GMT <volga> can anybody please help
07:00:28 GMT <etehtsea> hi. https://gist.github.com/etehtsea/db6715036e3f0a30d93721bc9f8775da What am I doing wrong? HSCAN ignores COUNT option
07:09:11 GMT <etehtsea> same behaviour for zscan
07:11:36 GMT <etehtsea> nevermind, found the issue https://github.com/antirez/redis/issues/1723
08:28:34 GMT <_Wise_> hi *
08:32:29 GMT <_Wise_> we are using Redis 3.0.5 and recently deleted *zillions* of keys
08:32:56 GMT <Habbie> that's a very old redis
08:33:01 GMT <Habbie> just an observation
08:33:07 GMT <_Wise_> ram usage remained unchanged because of fragmentation (expected behaviour)
08:33:27 GMT <_Wise_> is there a way to defragment *other* than to restart ?
08:33:37 GMT <_Wise_> Habbie: yes that's old, I know
08:34:49 GMT <_Wise_> used_memory_human:36.89G
08:34:49 GMT <_Wise_> used_memory_peak_human:49.12G
08:34:51 GMT <_Wise_> mem_fragmentation_ratio:1.32
08:35:27 GMT <_Wise_> fragmentation was 1.00 before the deletion
08:35:59 GMT <minus> i see, restarting that thing would take a while
08:36:20 GMT <_Wise_> yes, like 20 minutes or so
08:36:41 GMT <_Wise_> to play the aof file back
08:38:22 GMT <minus> if you have enough ram you could replicate to a new instance and switch to that. dunno if that's a great idea though
08:45:21 GMT <_Wise_> minus: hmm, that's unfortunately not an option for us right now
08:45:49 GMT <_Wise_> I think I'll restart the slave, failover and restart the former master
08:46:04 GMT <minus> sounds reasonable
08:46:29 GMT <_Wise_> that will be a 30 sec interuption
08:46:56 GMT <_Wise_> I was dreaming of a 'defrag' option
08:47:23 GMT <minus> you could maybe set the dump dir to a tmpfs location and speed it up
08:47:30 GMT <minus> otoh the bottleneck is probably CPU
08:47:45 GMT <minus> not that it matters for the failover
08:54:29 GMT <_Wise_> minus: about the tmpfs, well I'll need ram for that :)
08:54:43 GMT <_Wise_> that I don't have
08:55:11 GMT <_Wise_> thanks for the support, I'll do the failover trick
10:28:57 GMT <badboy_> there will be an active-defragmentation option in upcoming Redis
15:16:25 GMT <_Wise_> badboy_: thanks for the info !
15:16:35 GMT <_Wise_> I'll check it out when it is released
18:35:47 GMT <mfilipe> i'm using the c hiredis-0.13.3 lib and trying to use redisGetReply, but it is freezing the code. code: https://gist.github.com/mfilipe/ff1d45e09119971e351fd37fb96340ca
18:35:56 GMT <mfilipe> what is wrong there?
18:36:45 GMT <mfilipe> why is it freezing?
18:38:16 GMT <mustmodify> I'm spinning up a redis server when I start my automated tests then shutting it down afterwards. Still need to deal with it already being started. But is this config file sufficient? It seems like everything I need. https://gist.github.com/mustmodify/1f62b0c06449a1759a15e65aef63cbde
19:19:26 GMT <minus> mustmodify: why not not daemonize it, not write a pid file and instead track the processes and kill it when you're done?
22:07:17 GMT <wrkrcoop> question, i want to start out on aws with a small CPU so redis has like a gig, but if the website gets traffic ill eventually want to have a bigger machine, if i use the redis cluster, can i add a machine then shut down the smaller one without losing data? will the data form the smaller machine move to the bigger one?