07:24:50 GMT <Cain_> hey guys
07:25:53 GMT <Cain_> I was wondering about the use case of using Redis beside traditional Relational Databases
08:14:03 GMT <sssilver> Hey guys, I can only store primitive types in Redis?
08:14:19 GMT <sssilver> can I store documents like I can in Mongo?
08:14:30 GMT <sssilver> e.g. key => { document with a bunch of fields }
08:31:54 GMT <Inge-> sssilver: you can store whatever you want in redis, but it would be a binary field - so it wouldn't be able to e.g. index on specific json fields or anything
08:32:19 GMT <Inge-> but storing json strings in redis is viable to do, and a way to chache them "in memory" for fast access.
08:32:23 GMT <Inge-> cache*
09:03:21 GMT <Habbie> Inge-, sssilver, HSET might be of limited interest here
09:04:10 GMT <sssilver> Habbie Inge- thanks! I was using Rust with the Redis lib, and it was too much work so I just switched to Mongo :(
09:04:18 GMT <sssilver> although Redis would be better for this particular project
09:04:21 GMT <sssilver> the Rust lib is too raw
10:06:02 GMT <tinyhippo> is anyone using haproxy in front of redis, at all? I seem to be getting a high number of connection issues Error writing/reading bytes from the server
10:12:25 GMT <KekSi> as a tls terminator?
10:13:24 GMT <KekSi> or are you loadbalancing reads on replicas?
10:14:16 GMT <KekSi> i haven't used redis behind haproxy but found no problem with rabbitmq (messagequeues - they use long-running tcp connections)
10:32:47 GMT <soveran> tinyhippo: I've tried haproxy with Redis in the past, didn't get good results
10:36:57 GMT <tinyhippo> soveran: yeah think I'm going to ditch it
10:39:08 GMT <minus> not really having problems with HAProxy in front of a sentinel setup
13:34:03 GMT <plitter> is there a function to check how big a key is?
13:34:24 GMT <plitter> as in size in B?
13:35:33 GMT <minus> plitter: there's DUMP, but it compresses the data (like in rdb files)
13:36:45 GMT <plitter> I was hoping to get the size it uses inside redis
13:37:15 GMT <plitter> I can see that info shows some info, but not for keys as far as I can see
14:13:18 GMT <badboy_> there is no such command to see the actual in-memory size
14:29:26 GMT <plitter> ok, I'll guess info is my friend then :)
15:26:17 GMT <gobbler> what is a more typical setup; connect frontend clients directly to redis or shield redis by an application server?
15:27:13 GMT <minus> what do you mean by frontend clients?
15:27:32 GMT <minus> if something has access to redis it has access to pretty much everything
15:28:04 GMT <gobbler> frontend clients = web clients or fat UI clients
15:29:03 GMT <minus> oh yes, definitely shield it then
15:30:26 GMT <gobbler> minus: thanks, is it common to connect clients to an app server via websockets and connect the app server to redis via pub/sub? thats my design proposal
15:33:31 GMT <minus> that doesn't sound very good, but it mostly depends on what you want to achieve and on what scale
20:40:59 GMT <badboy_> c
20:41:01 GMT <badboy_> oops
22:53:11 GMT <MasterGberry> How does amazon's vCPU's work with redis?
22:53:16 GMT <MasterGberry> I thought redis was single threaded
22:53:29 GMT <MasterGberry> Yet they seem to only run a single instance and give each one more power?
22:54:26 GMT <MasterGberry> Or does AWS shard or elasticache?
23:24:38 GMT <ddoom_> Is there a way to serve data from disk rather than keep in memory? Data is refreshed daily and we'd like to keep the old data until the new data is finished importing, but don't have enough space in memory to keep both sets
23:51:31 GMT <Habbie> ddoom_, i'm not that great with redis but my guess would be 'no'
23:51:37 GMT <Habbie> ddoom_, is memory that expensive for you?
23:52:44 GMT <ddoom_> we'd need double the amount, 1/2 of which is unused 99% of the time
23:53:38 GMT <ddoom_> it's not that expensive, but just looking at alternatives