05:54:00 GMT <nadermx> I got a quick question, I have a key in redis that expires at end of UTC day.  I want to store this key for 24 more hours as a different key and have it expired the next day
05:54:21 GMT <nadermx> since the key gets reset during the day, but still expires end of day (think donations to a timer)
05:54:37 GMT <nadermx> how would I go about storing this key as say goal_history
05:54:54 GMT <nadermx> so it updates each time the main key does and when it expires the next key also named the same thing doesn't update this key
06:13:35 GMT <nadermx> nvm set the expiration name as the date, worked out fine, thanks all
09:26:29 GMT <shmoon_> badboy_: aah so a worker in a queuing mechanism won't have to hit the list constantly, it'll hit once with a blocking time so the load on redis is less
09:26:38 GMT <badboy_> yes
09:26:43 GMT <shmoon_> but this means if i have 100 different workers, redis will need to block 100 different clients
09:26:48 GMT <shmoon_> for lets say 5s of timeout
09:26:57 GMT <shmoon_> would't that be intensive?
09:27:00 GMT <badboy_> no
09:27:24 GMT <badboy_> network i/o is async in Redis anyway, so they are just put onto an internal queue and get data once there is some
09:27:26 GMT <shmoon_> I see, nice. now i get it thanks :)
09:27:29 GMT <shmoon_> redis is cool man :D
09:27:41 GMT <badboy_> if there is no data (and the timeout is not run out) it will simply do nothing
09:32:53 GMT <moteus> hello. Is it possible pass NULL value to Lua script as argument from client?
09:33:12 GMT <badboy_> in theory it might be
09:33:20 GMT <moteus> i try $-1
09:33:29 GMT <badboy_> and?
09:33:29 GMT <moteus> but got `invalid bulk length`
09:33:48 GMT <badboy_> then it is not possible^^
09:33:59 GMT <moteus> why protocol suppotr only strings as command?
09:34:19 GMT <badboy_> not sure what you mean?
09:35:51 GMT <moteus> May be I am wrong. But seems it not possible sent e.g. numbers like `:123`
09:36:33 GMT <badboy_> as the command?
09:36:39 GMT <moteus> yes
09:37:07 GMT <badboy_> well, because commands are strings^^
09:37:10 GMT <moteus> I play with basic eval "returng ARGV" command
09:37:21 GMT <moteus> but arguments not strings
09:37:39 GMT <badboy_> I am still not sure what you are trying to do
09:38:32 GMT <moteus> In fact I work on my own implementation of client. And try figureout what I can and can not do
09:39:05 GMT <badboy_> you can do everything, you just need to handle errors^^
09:39:30 GMT <moteus> also I want use redis to use job queue
09:40:12 GMT <moteus> I try qless. and it works via Lua scripts
09:41:39 GMT <moteus> but in api thay seems allows pass optional args like `evalsha ... arg1 arg2 arg3`
09:41:54 GMT <badboy_> so?
09:41:56 GMT <moteus> and in api they test `if arg2 then`
09:42:16 GMT <moteus> but I have no idea how i can pass arg3 and not pass arg2
09:44:33 GMT <moteus> but thanks. seems its not possible.
09:44:35 GMT <badboy_> probably not at all, but you can pass only arg1
16:15:51 GMT <indiancomet> redis doesn’t allow expiry at field level of a hash for performance reasons; is there a pattern i could use to achieve a similar result?
16:41:38 GMT <minus> use keys instead of a hash
18:22:13 GMT <bironran> I'm sure this has been asked before (in fact, I know it, but I haven't found any concrete answer) - Cluster with tune-able replication. I have several classes of data I wish to store in Redis. Some of them (e.g. counters) can suffer losing some writes, others (e.g. config) can suffer losing writes acks (employ retry mechanism). However, a third kind (fronting a db query) should be consistent but may not be available.
18:22:57 GMT <bironran> I really like Cluster managing active/standby (master/slave for replicated data) and would love to use it for the 3rd kind of data as well - but I can't figure out how to swing that.
18:24:42 GMT <bironran> I saw suggestions and replies, but not a concrete example of how it can be done. My current thinking is to attach the redis node id to the key as namespace for the 3rd kind of of data - but my client library (Jedis) is very resistant to that kind of modification.
18:25:59 GMT <bironran> (and of course use explicit key hash slot assigning, e.g. {logical_slot}_%name%_%nodeid%)
18:45:47 GMT <nicodemus_> hello
18:46:31 GMT <nicodemus_> is there any way in a two-node master-slave setup to promote the slave to master if the master redis fails?
18:48:03 GMT <nicodemus_> (since clustering recommends 6 instances, I'm looking for all the possibilities)
18:54:57 GMT <minus> nicodemus_: sentinel
18:56:32 GMT <nicodemus_> minus, got it. Thanks!
20:09:16 GMT <Dirkos>  20:06:59.680 [[main]<redis] WARN  logstash.inputs.redis - Redis connection problem {:exception=>#<Redis::CommandError: ERR Error running script (call to f_3236c446d3b876265fe40ac665cb6dc17e6242b0): @user_script:2: WRONGTYPE Operation against a key holding the wrong kind of value>}
20:09:27 GMT <Dirkos> Somebody who can explain this to me what is happening?