01:23:33 GMT could I disable the following commands and still have a functional redis service in production? 'bgrewriteaof', 'bgsave', 'config', 'debug', 'del', 'flushall', 'flushdb', 'keys', 'pexpire', 'rename', 'save', 'shutdown', 'spop', 'srem' 02:00:02 GMT Hey guys, I have an issue where the redis in my docker container somehow keeps getting set to save RDB snapshots. What would cause this to get enabled after a bit of usage? Eventually my sets stop working and I have to restart redis 09:19:53 GMT crayon: if you don't need those commands, sure 12:57:25 GMT hi, i have a setup with 3 redis-servers and 3 redis-sentinels. It seems that sentinel often gets confused and as a result my redis-cluster becomes broken. Anyone have an idea what this might cause? 13:00:36 GMT Guest22192: connecting them through any proxies? 13:00:42 GMT HAProxy, stunnel, etc 13:02:54 GMT minus: correct, I connect to the current master using haproxy, forgot to mention 13:03:35 GMT sentinels and redis-servers connect through the correct port, they don't use haproxy. Only my clients do that 13:18:04 GMT should be fine, but you should never ever attach a slave through haproxy 13:18:35 GMT if you do it once, sentinel will remember and try to use it (because haproxy is still running and seemingly working) 13:18:42 GMT I figured as much when I built the architecture, thanks for the heads-up 13:19:35 GMT I allready cleared sentinel's memory, to ensure historic errors were the cause of this issue 13:19:39 GMT you need to remove it from the sentinel configs (stop, remove, start) if it works 13:20:23 GMT just make sure there's no artifacts of it left on your config 13:21:06 GMT took us quite a while to identify the problem, and it's hard/impossible to prevent 13:24:30 GMT What do I need to remove minus ? 13:27:30 GMT in the sentinel config there probably is an entry that points to the haproxy ip/port, that has to be removed 13:42:27 GMT easiest is probably to delete sentinel configs and recreate them (especially if you can do that via ansible/puppet/chef/salt) 14:48:28 GMT minus: ill look into it, thank you for your input! 17:25:39 GMT Hi everyone. I am using Redis as a Broker(buffer) for storing data received from 100's clients and consuming it simultaneously. I am storing Logs data in different Lists in Redis and using LRANGE to consume in batches of 1000's. I have started facing slowness in consuming from redis since a week. Any Idea? 17:26:58 GMT I used to had ~50 connections to Redis sending data. Since a week, I've added more servers to send data to Redis. Actually, I am using Redis as Broker in ELK setup. I've tried tuning the Kernel parameters, redis settings, and tried almost everything. But could not find the issue yet... :( 18:14:17 GMT AnkurThakur, maybe your CPU is simply maxed out? Redis is single-threaded after all. 18:51:06 GMT colo-work: That is where the problem comes. I had been keeping my eyes over CPU, Mem and other things using Zabbix on Redis as well as Logstash and ES. It looks like it isn't problem of Redis as I am able to pull 1000's of Access Logs, but cannot pull 100's of Large Logs. 18:51:50 GMT colo-work: Everything is under control. Redis isn't using much CPU and Memory. Also, I've tried using taskset to make it use only 1 CPU. But nothing helped. Looks like its my own issue of logstash. 18:52:13 GMT can other redis clients consume the data quickly? 18:52:28 GMT you will need to pinpoint the bottleneck before you can hope to widen it 18:53:43 GMT gtg, ttyl 18:54:55 GMT colo-work: Thanks for the help though. 22:22:11 GMT Hello - I'm storing a number of session variable in redis, such that: 'A' = [x], 'W' = [x,y], 'R' = [x,y,z], 'V' = [x,y,z,w]. Is there some other data structure that can hold nesting sets like that? 22:22:52 GMT So, 'V' is always a superset of 'R', which is always a superset of 'W', etc.