00:28:38 GMT <polypus74> threading hiredis question: if i have a redisContext which i can guarantee will only ever be used from one single thread at a time, but i cannot say which thread will be doing so. is that ok to do?
07:21:30 GMT <kaushal> Hi
07:21:50 GMT <kaushal> can i add a new slave to the current redis master?
09:31:28 GMT <kaushal> Anyone around here?
09:31:46 GMT <kaushal> is there a way to verify if redis slave is in sync with master?
09:41:36 GMT <scottc1> redis-cli info replication should give you master_repl_offset which you can compare against your slaves
09:56:00 GMT <kaushal> scottc1: master_repl_offset:0
10:01:19 GMT <kaushal> scottc1: shall i pastebin the output?
10:03:17 GMT <kaushal> scottc1: https://paste.fedoraproject.org/430927/74279390/
10:05:13 GMT <scottc1> kaushal: Could you run the same command on the master please?
10:06:08 GMT <kaushal> sure
10:08:36 GMT <kaushal> scottc1: https://paste.fedoraproject.org/430929/27971014/
10:08:55 GMT <kaushal> scottc1: i have pasted both the master and slave redis-cli info replication
10:21:08 GMT <kaushal> scottc1: you around?
10:27:14 GMT <scottc1> kaushal: on and off I'm afraid :) It appears the offsets don't match so the slave isn't completely in sync with the master, are you seeing issues or were you just wanting to know
10:29:00 GMT <kaushal> just wanting to know
10:29:27 GMT <kaushal> scottc1: is it line 6 and 20?
10:29:56 GMT <kaushal> those numbers should match?
10:30:16 GMT <kaushal> scottc1: i have just added this new slave
10:30:59 GMT <scottc1> kaushal: You can use 5 & 6 on info replication on the master....you have the slave offset (offset=) & the master_repl_offset. On the master it should report the offset for all of the connected slaves
10:31:24 GMT <kaushal> scottc1: ok
10:32:14 GMT <kaushal> scottc1: also can i add a new slave on the fly?
10:32:32 GMT <kaushal> or do i need to have a planned downtime for the redis master to add a new redis slave?
10:34:42 GMT <scottc1> kaushal: In an ideal world they would match but how much of an issue that lag is can depend on your environment / requirements.
10:34:57 GMT <kaushal> sure
10:35:05 GMT <scottc1> kaushal: You should be able to use CONFIG SET & SLAVEOF commands to add a new slave without restarting
10:35:16 GMT <kaushal> ok
10:35:36 GMT <kaushal> scottc1: at the redis-cli> prompt?
10:35:48 GMT <scottc1> Yeah
10:35:54 GMT <kaushal> scottc1: OK
10:36:15 GMT <kaushal> scottc1: sorry caps on
10:38:21 GMT <scottc1> I believe OK being capitalised is the preferred form :)
10:38:50 GMT <kaushal> scottc1: ok
10:39:00 GMT <kaushal> scottc1: also i am following the redis docs
10:39:07 GMT <kaushal> http://redis.io/commands/config-set
10:39:27 GMT <kaushal> scottc1: please share me some examples of CONFIG SET and SLAVEOF?
10:40:35 GMT <scottc1> kaushal: The syntax for SLAVEOF is SLAVEOF host port
10:41:26 GMT <scottc1> You can get the list of accepted parameters for CONFIG SET by using CONFIG GET *
10:41:26 GMT <kaushal> scottc1: ok
12:21:21 GMT <stan_svec> hello, does anybody know whether redis guarantee that an element is returned with zscan command under this condition -> the element is contained in the sorted set from start to end of iteration but, the score of such element has changed (even several times) during iteration? thanks
14:18:14 GMT <thenotorious> join
18:41:24 GMT <tgodar> hi, looking at moving from memcached to redis.  Redis works as a distributed cache too right?  If I have three servers in pool a key/value is stored on one server, not all?
18:44:22 GMT <minus> with redis cluster, yes
18:45:09 GMT <tgodar> minus: is that a totally different implementation?
18:45:18 GMT <minus> no
18:45:29 GMT <minus> it's a redis feature since 3.0 or 3.2, don't remember
18:45:41 GMT <tgodar> ok cool.  I see redis provides disk persistence by default?  What sort of overhead does that entail and if significant can that be regained by disabling persistence?
18:46:39 GMT <tgodar> follow up to that, does the disk persistence make space finite?  as in if full its full rather than purging least used key?
18:47:53 GMT <minus> there's two modes for persisting to disk; full snapshots (periodic, depending on the number of writes, can be configured with the "save" directive) and appendonly where new stuff is always appended to a kind of log
18:48:16 GMT <minus> full snapshots tend to use less space then the inmemory representation because they're compressed
18:49:06 GMT <tgodar> but slower throughput I assume?
18:49:15 GMT <minus> to save a snapshot redis forks and the child does the saving and then exits. simple and effective, but may be problematic if you have a huge DB
18:49:31 GMT <minus> so it does not affect the performance of redis
18:49:41 GMT <minus> at least not directly
18:49:59 GMT <tgodar> to start out I'm looking to replace memcache with redis, not really looking to leverage anything else that redis can do beyond that at this time.
18:50:24 GMT <tgodar> Another thing that jumped out was upper memory limit...
18:50:30 GMT <minus> then i suggest you disable persistence and give it a try
18:50:50 GMT <minus> memory limit and eviction policy are configurable
18:50:58 GMT <tgodar> let's phrase it this way, am I right in thinking I can drop this in as a replacement and via configuration get the same behavior I have now?
18:51:00 GMT <minus> by default there's none
18:51:08 GMT <minus> probably
18:51:19 GMT <tgodar> Works for me.  Thanks minus
18:51:39 GMT <minus> not sure if redis supports the memcached protocol, but i vaguely remember having heard something like that
18:52:08 GMT <tgodar> for a php app that I want running on 7.0 and redis seems to have a lot more momentum currently than memcached (in terms of client / extension) (well and everything else really...)
18:52:47 GMT <minus> i'd consider redis a safe technology
18:52:55 GMT <tgodar> minus, its pretty well abstracted... using it as a key/value store I can't imagine needing to refactor much
18:53:02 GMT <minus> i.e. not gonna die any time soon
18:53:16 GMT <minus> yeah, that's cool
18:53:42 GMT <minus> once you get it running you might even be able to optimize your data structures leveraging redis features
18:54:11 GMT <tgodar> i.e. lists and datatypes other than strings?
18:54:57 GMT <tgodar> yeah, I think its a move in the right direction.  Would expect I will find ways to leverage more of the features redis does offer on down the road
18:55:15 GMT <minus> list, set, sorted set, hash
18:56:10 GMT <minus> and if you wanna go all fancy you can implement logic in lua scripts inside redis, and in future also custom native plugins for redis
18:56:23 GMT <tgodar> sorted sets and time series stuff has me thinking...
18:56:38 GMT <tgodar> yup - feeling good about this move
22:03:06 GMT <tgodar> trying to understand how to use redis as I do memecahed across multiple servers... confused
22:04:04 GMT <tgodar> redis calls that sharding?
22:09:21 GMT <tgodar> not clear on why these nodes need to talk to each other...
22:27:29 GMT <Seventoes> ok, let's say i want to track "active pages"
22:27:46 GMT <Seventoes> aka, which pages have more than X views in the last Y minutes. is there a smooth way to do that with redis that doesn't involve 150 lookups? :D
22:34:23 GMT <Seventoes> or, essentially
22:34:45 GMT <Seventoes> is there a way to use a set/list/hll which lets me set expiration on individual elements?
22:34:53 GMT <Seventoes> kinda the effect i'm going for i guess