14:03:52 GMT <pyios> what is the diff of redis cluster and redis sharing?
14:03:56 GMT <pyios> sharding
14:06:04 GMT <minus> redis cluster does sharding
15:56:16 GMT <akp> hi, i need to do some testing and was  wondering can i just cat my redis dump.rdb file to itself to cheat and double my data?
15:56:26 GMT <akp> ie i dont care that its duplicate data, i just need more
16:30:45 GMT <codedmart> I am still not sure how to solve or track down my problem. I am using Azure for my vm running redis and keep getting  `ETIMEDOUT` issues when connecting to it from my local setup. The initial connection is fine, but I end up getting that timeout. I had the exact same config setup on Digital Ocean and never had any issues.
16:31:00 GMT <codedmart> Anyone have some suggestions or ideas?
16:36:45 GMT <akp> codedmart: is it connecting through any type of load balancer or anything?
16:37:18 GMT <codedmart> akp: I am not totally sure how azure works, but I did have to set the port open in their interface.
16:37:46 GMT <codedmart> I am using rethinkdb on the same machine with no conneciton issues though.
16:39:07 GMT <akp> dunno then, i havent used Azure since my demo account expired.
16:39:08 GMT <akp> sorry =/
16:57:53 GMT <codedmart> akp: Do you know with nodejs/redis how often a keepalive runs?
16:57:58 GMT <codedmart> By chance?
16:58:32 GMT <codedmart> Nevermind.
16:58:34 GMT <codedmart> I got it.
17:43:14 GMT <cpama> hi all.  just a redis noob.  but I have two questions.  first question is what is the max number of slaves allowed in a master / slave set up?
17:44:12 GMT <cpama> we are just starting to check out redis and come up with a design for our application.  I've been asked to consider if we can use redis to replicate to up to 150 read only slaves... with one master.
17:44:27 GMT <cpama> as a programmer.. this "feels" wrong.  so I'm also looking at pub/sub options
17:50:34 GMT <cpama> my second question is this:  I just read this article http://redis.io/topics/notifications and am trying the test suggested at the bottom of the article about trying out pub / sub.
17:50:41 GMT <cpama> Here's a pastebin showing what I've tried:  http://pastebin.com/KkUCDRE0
17:51:05 GMT <cpama> Not sure what I'm doing wrong but I don't see anything happening in the first terminal screen.  Trying various types of commands pertaining to my keys
18:36:25 GMT <minus> cpama: i don't know if there's a hard limit, but with too many slaves you might run into a bandwidth bottleneck on the master. you can use a hierarchy of slaves instead as well
18:36:58 GMT <cpama> minus, can you point me to an article about how to set up a hierarchy of slaves?
18:37:32 GMT <minus> i don't know any, sorry, your googling is as good as mine
18:37:48 GMT <cpama> hm. ok
18:37:49 GMT <dhanesh> ping
18:38:02 GMT <cpama> minus, do i have to set up a cluster to do this?
18:38:06 GMT <minus> a slave can simply be a slave of another slave, nothing magic about it
18:38:11 GMT <dhanesh> how to use redis as cache of mysql
18:38:32 GMT <cpama> ok minus i will just play around and see what I can find.
18:38:33 GMT <cpama> thank you
18:38:36 GMT <minus> if you just want read replicas and not sharding, no, you don't need cluster
18:39:12 GMT <cpama> minus, yeah... i just asked because when i googled "redis hierarchy slaves" that's the first thing that came up... so i thought it might be a requirement
18:39:16 GMT <cpama> minus, thanks for the input
18:39:24 GMT <minus> redis cluster or redis sentinel can do HA though (mainly sentinel) in case you want that
18:39:54 GMT <cpama> don't want partitioned data.  need the entire data set on each slave
18:40:15 GMT <minus> cpama: regarding your pubsub/keyspace notification question: use double underscores instead of single ones
18:40:23 GMT <cpama> oh!
18:40:33 GMT <cpama> <retrying...>
18:41:32 GMT <cpama> minus!  bingo!!  that worked!!
18:41:37 GMT <cpama> thanks1
18:42:21 GMT <cpama> minus, any comments on setting up a bunch of slaves vs. using client that sub to a generic topic that would include all the keys/values I need to keep track of?
18:43:39 GMT <cpama> minus  - in other words, when i suggested that we would have to have many many slaves (hundreds)... did you choke? laugh?  or was it... "that's easy..." (keeping in mind of course, your comment about bottlenecks)
18:45:45 GMT <minus> i'm unsure what you're asking
18:46:17 GMT <minus> the biggest redis setup i work with is a HA setup with 2 instances
18:46:35 GMT <minus> one with lots of slaves is definitely a challenge to manage
18:52:36 GMT <cpama> minus... just want to get a pulse from someone with more redis experience to see if this "design" is what redis is supposed to do... vs. trying to make redis do somethign it was purposed for.
18:52:47 GMT <cpama> i come from a postgresql background and from messaging background
18:52:58 GMT <cpama> when I need to send data to many many clients... I use pub / sub
18:53:24 GMT <cpama> but my team lead is interested in replication to all these slaves.  it's just a new paradigm for me
18:53:36 GMT <cpama> wanted to bounce it off more experience redis people
18:58:15 GMT <minus> a pubsub to reach many client makes sense of course, probably works with slaves though i've never used that.
18:58:26 GMT <cpama> ok.
18:58:37 GMT <cpama> will do.  you've been a tremendous help.  thank you very much for taking the time.
18:58:43 GMT <cpama> virtual beers for you!
18:58:44 GMT <cpama> :D
18:58:48 GMT <minus> if your clients must not miss any pubsub events you might not want to use pubsub though
18:59:06 GMT <cpama> +1
18:59:26 GMT <minus> (because if a client loses connection for a bit it will obviously not get the messages meanwhile; redis doesn't store them)
19:00:36 GMT <cpama> hm. i thought read that a) it doesn't GUARANTEE all messages... but... since the script name that is requesting the subscription is cached somewhere, even if the client goes down... it can recover the missed messages.  Maybe I misunderstood.
19:00:38 GMT <minus> oh, and if every connected client needs to access the data that was just updated (when using keyspace notifications) it might make more sense to just use PUBLISH yourself and push the data right along, but you should probably run some experiments with that
19:01:06 GMT <minus> cpama: it buffers the messages as long as you're connected
19:01:12 GMT <cpama> ok.
19:01:27 GMT <cpama> i have to run to a meeting now... but thanks again minus!!
19:02:19 GMT <minus> always glad to be able to help; if you do set up something big with redis let me know how it goes
19:02:35 GMT <cpama> for sure.  i will look you up here and report in.
19:03:23 GMT <minus> cool, looking forward to it