03:57:51 GMT <sector_0> hey
04:00:08 GMT <sector_0> I need a small, fast in memory storage of key/value pairs, which is exactly what redis is for, but how hard is it to get a production-ready setup of redis going?
04:00:18 GMT <sector_0> also, what's the footprint of redis?
04:02:57 GMT <sector_0> the aim is to make my setup as scalable as possible
04:05:15 GMT <sector_0> how heavy is a redis process?
04:43:36 GMT <minus> sector_0: ~5mb per instance. There's not really anything special to do to get it production ready. Use cluster to scale up (though that is more difficult). One redis instance is a single process/thread (plus one additional while saving to disk)
04:45:11 GMT <sector_0> minus, oh ok thanks
04:45:39 GMT <sector_0> well I don't need persistence so that last part isn't relevant in my case
12:34:19 GMT <gojinjedi> what would you guys recommend as the best tutorial or course to learn redis?
13:09:14 GMT <ebalsumgo> minus: re: sentinel and cluster… well thats interesting documentation doesnt make that very clear. can you elaborate?
14:36:29 GMT <badboy_> ebalsumgo: https://fnordig.de/2015/06/01/redis-sentinel-and-redis-cluster/index.html
14:37:02 GMT <ebalsumgo> badboy_: yeah thanks, i actually have been reading and making a bit more sense of things all morning
14:37:56 GMT <ebalsumgo> i guess one question i have is that there arent that many production examples. so if i want to run a 3 master cluster is it ok if i run the slaves on the same box (as in is that the accepted practice) or should i be looking at 6 box total
15:13:00 GMT <badboy_> that depends on how fault-tolerant you want to be
15:13:15 GMT <badboy_> how likely is it that just a single process dies vs. the whole box?
15:25:34 GMT <ebalsumgo> badboy_: well the whole box will def go down at least once a month for maintanance. so i guess that answers my own question as the docs seem to indicate that redis-cluster doesnt like it when that happens or at least thats how i am reading it here https://redis.io/topics/cluster-tutorial
17:44:05 GMT <metsuke> Using Sentinel, data should only go to the master right?
17:46:08 GMT <badboy_> metsuke: that's true for all of Redis deployment modes
17:47:59 GMT <metsuke> @badboy_ thanks, so then if the master fails and a slave takes over, then there'd need to be some mechanism to point to the new host.  Is there a recommended proxy or load balancer for that?
17:48:14 GMT <badboy_> Ask the sentinel
17:48:22 GMT <badboy_> Sentinel knows
17:48:29 GMT <badboy_> Client libs do have support for it
17:48:34 GMT <badboy_> there might be some proxies handling that as well
17:49:07 GMT <metsuke> so the sending application needs to query one of the sentinels?
18:20:04 GMT <badboy_> yes