03:42:18 GMT <rolo> Hey all, I have a sorted set performance question
03:42:55 GMT <rolo> I'm trying to get the ZRANK of several members, I can only do this with MULTI, right?
03:43:32 GMT <rolo> Also, is ZSCORE faster than ZRANK?
08:38:45 GMT <aiRness> Hello. I have the following issue. I didn't had the overcommit set to 1 so the redis couldn't fork, now everything looks good although it seems that I'm getting out of memory in the box. Should I add more RAM or just set a maxmemory setting at the configuration?
08:38:51 GMT <aiRness> how can I measure this?
09:33:00 GMT <aiRness> also, I've put the new values with config set but still it doesn't seem that redis is using them
09:39:12 GMT <minus> if you don't set maxmemory it'll just exhaust available memory
09:39:27 GMT <minus> check if the values are really set using `config get`
09:40:38 GMT <minus> if you're starting redis-server yourself be sure to pass the config file as argument, it doesn't load one by default
11:03:57 GMT <bitonic> Is there a command that, given a list of keys, tells me which of the keys exists?
11:04:03 GMT <bitonic> `EXISTS` just tells me how many of them exist
11:07:18 GMT <minus> just run exists multiple times?
11:07:38 GMT <minus> in a multi/exec block if you need atomicity
11:18:52 GMT <aiRness> minus: it is running via docker
11:19:21 GMT <aiRness> config get gives me the vlaues that I've set but the memory is not going down
11:23:49 GMT <minus> aiRness: you may be to set the eviction policy as well
11:39:49 GMT <aiRness> minus: I've changed it to volatile-lru, my question is, if I throw a configuration file /usr/local/etc/redis on the docker container it would get read if I restart the container?
12:27:13 GMT <bitonic> minus: so the answer is "no"
12:30:53 GMT <minus> bitonic: probably
12:31:29 GMT <minus> aiRness: i don't know how config is handled with the docker image, the readme of it should tell you how it works
12:35:26 GMT <bitonic> minus: OK, thanks
14:00:37 GMT <derfel> Hi, if I want to search if there's a value inside a range I can use a sorted set with ZRANGEBYLEX and that's ok. But what if I need to save a range into a list/set and check if an integer value is inside a previously saved range? What's the best data structure to model this case?
14:08:25 GMT <_Wise__> Hi *, I have an infra question: I have a Redis 3.0.5 server that ran out on connection (10000). I want to identify the clients that have an idling connection with the server (timeout is 3600 seconds)
14:08:29 GMT <_Wise__> *but*
14:08:50 GMT <_Wise__> since I reached max clients, I can't run a 'client list' command in redis-cli
14:09:22 GMT <_Wise__> how can I get the client list in this condition ?
15:16:54 GMT <adac> When redis is shut down, is there all be  written out the dump file first?
15:23:12 GMT <aiRness> minus: btw, I have 5 GBs of ram left and the dump file is 1.5 GBs, how much should I put at maxmemory? Is there enough memory to handle it because I would need more. How can I calculate this?
15:29:24 GMT <minus> the dump is compressed, how much memory you have to leave free depends on how much memory is required for compression; i don't know how much that is, probably not much
15:40:47 GMT <aiRness> minus: ok
19:06:50 GMT <scriptor> does the MONITOR command maintain the order of writes?
19:07:26 GMT <scriptor> meaning, if client 1 writes "FOO" before client 2 writes "BAR", is it guaranteed that MONITOR will show the client 1 write before the client 2 writ?
19:07:30 GMT <minus> yes
19:07:39 GMT <scriptor> nice, thanks
19:08:56 GMT <zergov> Hi ! I am working on a REST api. Would it be fine if I hit redis on almost every request ? The query would be a simple HGET. I am mostly scared about the round trip time of my web api to the redis instance.
19:09:30 GMT <zergov> By fine i mean, is this okay and a good practice / usage of redis.
19:09:41 GMT <minus> zergov: faster than hitting anything else if it's running on the same machine
19:10:17 GMT <minus> nodejs?
19:11:30 GMT <zergov> minus: Alright then, my initial use case is that I need to read before write on my cassandra DB... but reading before writing is considered a bad practice on cassandra
19:12:34 GMT <minus> because performance or because consistency?
19:12:56 GMT <zergov> read before write is considered bad practice on cassandra because of the performance
19:13:15 GMT <zergov> reads are more expensive then writes.
19:13:36 GMT <zergov> I just need to do a simple business check before writting. And I expect a lot of writes
19:13:36 GMT <redis553> My application is storing a datetime as a long. this makes it difficult to read the value in redis' cli. I wanted to make a lua script to read the key and convert its value into a readable datetime using lua's os.date. Apparently the redis lua interpretor does not load os.date. Is there another way I can do what I want?
19:13:38 GMT <scriptor> nothing wrong with that
19:13:55 GMT <scriptor> no different than doing a SQL query for every request
19:14:47 GMT <zergov> scriptor: are you talking about my first question ?
19:14:52 GMT <scriptor> zergov: yes
19:15:08 GMT <zergov> noice
19:15:30 GMT <minus> yeah, checking things in redis is mighty fine
19:16:05 GMT <zergov> This is making my day !
19:16:20 GMT <zergov> Now I got a shitton of reading to do on redis ! Project looks great haha
19:16:34 GMT <minus> in fact we replicate everything in our DB to redis to get cheaper reads
19:17:05 GMT <zergov> So when you write, you also write on redis, and the reads are made on redis only ?
19:17:16 GMT <minus> yes, kinda
19:17:23 GMT <zergov> And for example, if a key is not found on redis, you fallback on the db
19:17:26 GMT <minus> DB to redis happens asynchronously
19:17:34 GMT <scriptor> minus: how do you do the replication?
19:17:43 GMT <scriptor> I'd assume writes happen to the DB only and reads come from redis
19:17:45 GMT <minus> a custom daemon
19:18:04 GMT <minus> reads are only done on redis, no fallback
19:18:17 GMT <scriptor> hmm
19:18:29 GMT <minus> we're navigating back towards DB-only though, since the replication daemon is rather failure prone
19:18:40 GMT <scriptor> ideally in our setup we'd be using elasticache (aws redis) as a cache for dynamo
19:19:30 GMT <zergov> Thanks a lot for your help ! Have a great day !
20:51:03 GMT <JustinHitla> you can't catch me
21:53:17 GMT <OverCoder> ehm
21:53:25 GMT <OverCoder> redis vs. sql?
21:53:43 GMT <OverCoder> I use MySQL, but then, I wanted to have a huge table with auto-expiring rows, and then, people suggested redis for efficency
21:53:44 GMT <OverCoder> what is this
22:04:40 GMT <_Cory_> Is it possible to set the ttl on a specific field rather than an entire key?
22:12:47 GMT <Inge-> OverCoder: redis is an in-memory key-value store on steroids.
22:16:43 GMT <_Cory_> BEEFCAKE xD
22:24:31 GMT <OverCoder> Inge-, aw, hmmm
22:24:33 GMT <OverCoder> got it
22:24:42 GMT <OverCoder> So it's a tinier but faster storage :P
22:24:44 GMT <OverCoder> that's what I want, yay thx
22:44:30 GMT <Yukiku> https://i.sli.mg/5YM3Im.png am I missing something?