07:50:22 GMT <antirez> dict.[ch] is a terrible piece of code inside Redis because of the exported API... soon or later I'll get rid of it entirely
08:13:00 GMT <colo-work> is there support for transparently compressing values inside redis? (I'm basically looking for something that supports Redis' API for very large plaintext datasets, and don't very much care for performance. I've found ardb which provides Redis compatibility with a disk-based storage engine, but would like vanilla Redis more, for reasons of familiarity.)
08:48:45 GMT <minus> colo-work: custom redis modules would allow you to do that
08:50:03 GMT <colo-work> minus, yeah, I guess so. we're considering just wrapping the redis client to LZO-compress on write, and LZO-uncompress on read
08:50:41 GMT <colo-work> it's the less painful way to do it in this case and for us, but since Redis can already compress values in RDB dumps, this might be an interesting (albeit niche) feature to have that could be easy to implement in Redis proper
08:53:30 GMT <minus> if you do it client side you also don't have to transfer as much over network
08:54:23 GMT <colo-work> yes, but that's not a concern for us here. also, redis won't be bottlenecking on CPU if the (de)compression chews too many cycles.
08:54:32 GMT <colo-work> (but again, not a concern for us in this case)
09:36:46 GMT <dror> Hi there! How is it possible that we reached 120,000 ops with standalone but with cluster we only reach to about 60,000-65,000 ops?
09:37:28 GMT <dror> 8 master nodes here, with slave for each one
11:23:57 GMT <shafox> I have100000 keys in my redis, I want to delete all the keys which has prefix api: * , how would i do it ?
12:43:25 GMT <minus> shafox: writing a lua script would be one option
12:57:14 GMT <shafox> minus: any blog post or links for doc would be great.
13:02:13 GMT <minus> shafox: EVAL 'local n = 0 for i,k in ipairs(redis.call("KEYS", "prefix:*")) do redis.call("DEL", k); n=n+1 end return n' 0
13:02:44 GMT <shafox> minus: thanks :)
14:49:17 GMT <kaushal> Hi
14:49:46 GMT <kaushal> I am setting up sentinel in my test setup while referring to http://redis.io/topics/sentinel
14:50:43 GMT <kaushal> I have a question regarding sentinel configuration. I have redis setup in Master Slave configuration using 3.2.0 version
14:51:04 GMT <kaushal> Do i need to configure sentinel on both Master Slave box?
14:51:21 GMT <kaushal> Master is running on a separate VM and Slave is running on a separate VM
14:51:27 GMT <kaushal> Please advise
16:17:01 GMT <nacsurte> i'm adding about 40 million list items into my redis queue a day, and removing a few million as well, and i've noticed that it bumps the server load up significantly
16:17:28 GMT <nacsurte> it seems like redis-rdb-bgsave is always running, which looks like what is causing the increase in load
16:17:31 GMT <nacsurte> is that a common issue?
17:31:53 GMT <danns23> I have a question regarding redis cluster failover. I am currently testing out a 9 node cluster - 3 masters with 2 replicas each. When I take one of the masters out by stopping the redis service I see that the cluster notes the node is failed but I do not see either of the slaves being promoted to master.
17:32:26 GMT <danns23> redis-cli cluster nodes shows the failed node and the two slaves as still being slaves to that node