11:47:46 GMT <adac> guys, is there a proxy one can use for a redis service?
11:52:05 GMT <minus> like twmeproxy? HAProxy?
11:52:25 GMT <zetab> haproxy in tcp mode?
11:52:31 GMT <zetab> i have not tested but could work
11:54:19 GMT <minus> it does, we're using that
12:19:24 GMT <adac> minus, zetab thanks guys!
12:20:42 GMT <adac> minus, may I ask on how your setup looks like with haproxy? You have two or more redis instances  and in front there is haproxy right?
12:20:49 GMT <adac> not master slave or such?
12:21:15 GMT <minus> it's a 2-redis 3-sentinel failover setup with haproxy n front
12:21:21 GMT <minus> *in
12:22:15 GMT <adac> minimalism, ok I see so it is a cluster
12:22:32 GMT <adac> uh wrong sorry, meant minus
12:22:35 GMT <adac> thanks!
12:22:45 GMT <minus> no, not a cluster, at least not in the redis sense
12:23:58 GMT <adac> minus, hmm did you follow an online tutorial on that? Could you send me a link maybe so I can read about this?
12:27:20 GMT <minus> there's a couple, you should be able to find them easily
12:27:36 GMT <minus> basically you get haproxy to do a PING to redis and call it a day
12:28:32 GMT <minus> for failover + having haproxy connect to the master you make haproxy send INFO REPLICATION to both redises and choose the one that replies with role:master as active backend
12:28:58 GMT <minus> i don't really run that myself; the service is managed by our hoster
12:29:19 GMT <adac> minus, ok thanks for the infos!
16:44:00 GMT <ddd_> seq 3 | redis-cli -x lpush test; redis-cli lrange test 0 -1  => "1\n2\n3\n"
16:44:43 GMT <ddd_> not sure why this makes one value instead of three... also not sure why the newlines are showing up
17:17:55 GMT <minus> gotta use xargs
18:57:17 GMT <aah> 'lo all.  question.  I've got a redis list that represents the last N events; events are added via LPUSH, and the list is trimmed to ensure the size is limited to N.
18:57:33 GMT <aah> events are buffered and written every few seconds.
18:58:23 GMT <aah> I'm wondering if I should special case the path where I know I'm writing N events.  i.e., instead of LPUSH ..., LTRIM ... instead I'd do DEL ..., LPUSH ...
18:59:08 GMT <aah> without looking at the redis code at all, I could imagine that the push would be more efficient when the key doesn't exist?
19:00:32 GMT <minus> doubt it
19:01:08 GMT <minus> and unless you're pushing redis to its limits i wouldn't worry about that
21:13:36 GMT <damnski> how do I run redis from startup as a singleton?  no sentinel, no slaves.  I just want to start it as 'slaveof no one.'  I have that set in redis.conf but it becomes a slave at startup.  then I have to set it back with redis-cli
21:18:32 GMT <minus> damnski: are you possibly loading the wrong config or connecting to the wrong port?
21:23:45 GMT <damnski> minus, it is always possible, but no.  we typically run with three separate hosts, with sentinel.  my case now is for a test env and I don't want replication.  I was the redis node to come up and become master, disregarding the need to replicate.
21:24:24 GMT <damnski> I am using a startup script to point to the config file.
21:25:21 GMT <damnski> I can run 'redis-cli slaveof no one' and the node/instance becomes master.  I have the same line in the config file but it still starts up as a slave.
21:44:10 GMT <badboy_> Redis doesn't set itself as a slave. so either it is in your config or it gets triggered externally, e.g. from a Sentinel or similar