03:47:14 GMT <jpastore> Hi, I'm very new redis. Just installed 4.0. I'm using bash to load a bunch of data from csv files. I loop through the origin data to do an hgetall on a hash and it fails because ip comms goes out and comes back.
03:48:48 GMT <jpastore> I saw in the log file I needed to adjust some values in sysctl.conf and redis.conf, but the same problem is occurring.
03:50:06 GMT <jpastore> and pro tips would be appreciated. I haven't tried this in a clustered environment yet. Is there a limit to how much I can throw at a single node? bench marking shows 1m queries in 7.2 secs so I figured looping through 5k lines at a time to see if they all made wouldn't be a problem
03:50:17 GMT <jpastore> s/^and/any/
08:30:09 GMT <thekrynn> hello, was wondering if anyone here has had problems with the redis implementation of hyperloglog
10:37:12 GMT <shayla> Hi guys, it seems i've got some limitation using the key of hash type. I can't use something like HSET 595e051e2620e15b708b4571
10:37:43 GMT <shayla> In documentation i can't find anything about that
10:37:46 GMT <shayla> Si that right?
10:38:04 GMT <shayla> *is
10:41:34 GMT <_ikke_> shayla: What is the limitation you are seeing?
10:42:19 GMT <shayla> WRONGTYPE Operation against a key holding the wrong kind of value
10:42:22 GMT <_ikke_> shayla: for me, redis just accepts that key
10:42:33 GMT <_ikke_> shayla: that means the key already exists as a different type
10:42:42 GMT <shayla> Uh
10:42:54 GMT <shayla> Well, tahnks _ikke_ i will try to find out
10:43:26 GMT <_ikke_> shayla: "TYPE 595e051e2620e15b708b4571"
10:43:42 GMT <shayla> Thank you _ikke_ !
16:35:48 GMT <zskynet> hello
16:36:06 GMT <zskynet> I have an ask about redis sentilnet and multicast
16:36:28 GMT <zskynet> if someone can help me, thanks a lot
16:36:38 GMT <Exuma> Is there a way I can make Redis evict values in a set rather than keys when memory is full/
17:02:20 GMT <_ikke_> zskynet: not that I can help you, but it usually helps to first ask an actual question
17:04:12 GMT <zskynet> thanks ikke
17:09:42 GMT <blubberdi> Hello, I have a short question I couldn't figure out in the documentation. Is maxmemory checked against used_memory_rss or used_memory? I ask because I want to monitor my redis instance and be notified before the application can't write anymore into the redis instance.
17:50:58 GMT <badboy_> used memory, blubberdi
17:52:30 GMT <blubberdi> badboy_: so that's independed from the fragmentation? I can always rely on: if i want to write something into redis and it's smaller than maxmemory-used_memory it will work?
17:55:24 GMT <badboy_> pretty much
17:56:36 GMT <badboy_> we usually warn at around 70-80% usage (if at all, some apps just have an eviction policy)
17:56:53 GMT <badboy_> (we = my workplace where we run Redis)
18:18:47 GMT <blubberdi> badboy_: ok, thanks
19:37:08 GMT <traverseda> So I am developing a "game" in python. My hope is to leverage python data-science stuff, to make good voxel games. I would like to ship redis with this game. Is there some way redis could run "locally", maybe a reliable static redis build?
20:14:22 GMT <minus> i only see redis linked against m, dl, pthread and c, should be pretty portable
20:14:41 GMT <minus> shouldn't be difficult to link completely statically
20:14:51 GMT <minus> but why would you use redis there?
20:15:00 GMT <minus> can't you do things in-process in python?
20:15:06 GMT <minus> ^ traverseda
20:16:53 GMT <traverseda> minus, I want a game that scales-out, basically. Redis seems like the way to do IPC in that instance.
20:17:57 GMT <traverseda> So we can have one process that does geometry simplification, another that does crude voxel water simulation, and all those other tasks that aren't real-time
20:18:44 GMT <minus> storing things in redis seems like unnecessary overhead and delay there
20:20:21 GMT <traverseda> I'd really prefer not to write things like clustering code myself. And that is what I'm aiming for, a cluster.
20:28:57 GMT <minus> well, if redis works for you, it works for you
20:29:12 GMT <minus> to build redis statically: LDFLAGS=-static make
20:29:18 GMT <minus> worked on first attempt
20:52:01 GMT <JordiGH> Any clue why AWS what could cause this? One Redis instance has its memory and connection usage fluctuate while both instances have absolutely zero queries: http://jordi.platinum.linux.pl/piccies/why-redis-why.png
20:52:33 GMT <JordiGH> It's not a lot of memory, about 5 megs, but still, eh?
20:58:08 GMT <minus> correction: one test fails with the static build (stack trace functionality)
21:32:41 GMT <AsenLenBeyraklar> What's the default behavior when the exact same script gets SCRIPT LOAD-ed twice? Is the same hash returned?
21:33:14 GMT <badboy_> yes
21:33:21 GMT <AsenLenBeyraklar> Thank you
21:33:24 GMT <badboy_> scripts are identified by their sha1 hash
21:34:03 GMT <badboy_> so unless you generate a SHAttered Lua script it is pretty unique
21:34:12 GMT <AsenLenBeyraklar> shattered?
21:34:33 GMT <badboy_> Name for the SHA collisions Google recently produced
21:34:42 GMT <AsenLenBeyraklar> Oh right, thanks
21:34:53 GMT <badboy_> to my understanding you don't have to fear that for Lua scripts
21:35:11 GMT <AsenLenBeyraklar> Meh, I just load like 5 scripts max
21:37:29 GMT <AsenLenBeyraklar> I guess it doesn't really matter what Google released, because any hashing algorithm will have collisions. They just detected one.
21:38:05 GMT <badboy_> and broke some tools along the way
21:38:07 GMT <badboy_> but yes
21:38:59 GMT <AsenLenBeyraklar> I wrote that sentence without context. I was going to ask a dumb question of whether there is any plans of replacing SHA1 in Redis cuz collisions then answered myself :)
21:57:21 GMT <AsenLenBeyraklar> In Rust you can "prepare_invoke" a Redis script, which would return you struct of type ScriptInvocation. Does the ScriptInvocation map to some actual Redis feature or is it purely Rust library's implementation? Also, would it be a good idea to pass this ScriptInvocation around and re-use it, instead of the client?