08:35:17 GMT <tgy> what's the best c++ client for redis ?
08:44:51 GMT <minus> i've only ever used hiredis and can't exactly recommend it with C++. checking through the examples/repos of those listed here https://redis.io/clients#c-- seems like the thing to do
08:59:43 GMT <minus> https://github.com/cylix/cpp_redis and https://github.com/hmartiro/redox seem to look best, both require C++11+
08:59:47 GMT <minus> ^ tgy
09:02:09 GMT <minus> first one seems somewhat neater to use, latter probably offers better performance
09:49:03 GMT <badboy_> there's a redis client called redox? oO
09:52:44 GMT <minus> what's so surprising about it, badboy_?
09:54:00 GMT <badboy_> There's a operating system and kernel, written in Rust, called redox :D
09:54:09 GMT <badboy_> that's what threw me off
15:22:25 GMT <__Yiota> what's the best way to healthcheck redis?
16:52:37 GMT <badboy_> ping it
17:14:38 GMT <commonerror> I'm working on a redirect service (based on query params) I'm thinking of storying serialized JSON in a list ["someRegx => http://example.com", ...]. Is there a better way to go about this?
17:41:07 GMT <GNU\colossus> you want to map ONE set of query parameters (=key) to ONE destination URL (=value)?
17:41:14 GMT <GNU\colossus> or how is this supposed to work?
17:43:41 GMT <commonerror> GNU\colossus: I'll be storing multiple regexp keys with destination values. For incoming HTTP requests, I'll be checking if any of the expressions pass and if they do I'll redirect to the destination.
17:44:32 GMT <commonerror> In other words, the rewrite_rules List will be composed of a number of differing "regex => destination".
17:46:24 GMT <GNU\colossus> and your list of redirect rules is so dynamic that you need that backed by redis?
17:46:47 GMT <commonerror> Indeed it is
17:48:23 GMT <commonerror> Heh, I could probably just have a form for admin to fill out that appends their redirect to a file and redeploys the service :D No one would be the wiser.
17:49:09 GMT <GNU\colossus> I ask this because if you're not careful, your RE engine will end up compiling each resulting automaton anew for each incoming request and each stored regex.
17:51:59 GMT <commonerror> I'm not sure what you mean by "compiling each resulting automation anew". Could you expand? It's not so dynamic that the list is having redirect appended every minute, it's more of a periodical update.
18:01:06 GMT <GNU\colossus> commonerror, before a regular expression can match anything, it will be compiled into a DFA. that's why the POSIX RE has distinct compile and match functions. some runtimes (PHP, for instance) combine two steps (and transparently cache compiled REs)
18:01:41 GMT <GNU\colossus> if your runtime isn't celver enough, it will probably parse and compile each of your REs for each match you want them to perform, even though the resulting RE/DFA hasn't changed
18:01:57 GMT <GNU\colossus> in case you are serving loads of redirects, that might be a performance bottleneck of your service
18:02:08 GMT <GNU\colossus> if you expect a few hundreds matches a second at most, just forget what I wrote
18:03:41 GMT <commonerror> I'm looking at millions/second, but each of the REs will be unique. Granted, I don't expect to have more than a few 100 redirect rules in place at a given time.
20:27:25 GMT <GNU\colossus> is it correct to use Redis replication (read-only slaves) to achieve read scalability? or will a massive write storm destroy read performance due to the replication trying to keep up regardless?
20:27:52 GMT <Habbie> (1) yes (2) of course this is possible
20:29:10 GMT <GNU\colossus> is there a general rule of thumb for an upper bound for the number of slaves?
21:18:46 GMT <chris11> what are some common reasons I'd get connection refused for localhost:6379?
21:19:38 GMT <GNU\colossus> redis listens on a different port or interface. redis isn't running. redis has crashed, and isn't running any more. you have misconfigured your packet filter.
21:19:41 GMT <minus> yes, if redis isn't running
21:19:45 GMT <GNU\colossus> that about sums up the problem space
21:21:00 GMT <chris11> ps aux | grep redis gives me this: 0:00 redis-server *:6379
21:21:10 GMT <chris11> I am running redis from a docker container
21:21:14 GMT <chris11> in a virtualbox
21:22:06 GMT <xiaomiao> chris11: docker has its own network weirdness
21:22:17 GMT <xiaomiao> most likely you're not in the same network namespace as the container
21:22:24 GMT <chris11> ugh that's annoying
21:22:38 GMT <xiaomiao> ... and I almost feel tempted to ask "why", but ... then I don't want to hear the answer ;)
21:24:01 GMT <minus> chris11: so you need to connect to the ip of the redis container via the vbox vm
21:24:46 GMT <minus> you can avoid docker giving containers a separate ip by disabling network isolation by passing --net=host to the docker run command
21:25:09 GMT <minus> better just install linux on your host, and redis too
21:33:50 GMT <__Yiota> chris11 try `redis-cli` from container
21:38:48 GMT <chris11> connection refused.
21:39:19 GMT <minus> docker exec my-container redis-cli?
21:39:30 GMT <chris11> docker ps gives me 0.0.0.0:32771->6379. I call localhost:6379 not localhost:32771, right?
21:40:14 GMT <chris11> I'll try that
21:40:48 GMT <chris11> still same error
21:42:09 GMT <minus> you should go figure out how to use docker if you must use it
21:42:28 GMT <chris11> you might have a point.
21:45:47 GMT <chris11> so what's that command supposed to do? I read that its supposed to run a commad on the docker container. But I saw that I ran the command in the wrong spot so I reran it. I did not get any response. No error though.