18:12:45 GMT <netcrash> Hello, I'm running redis inside a docker container , and passing the config file has argument for the CMD , anyway, the file has "daemonize no" but redis executes has a daemon, when I remove the file it runs normal(not a daemon). What am I missing?
20:10:17 GMT <ledil> hello ive got a list of keys with the prefix "state_" if I use scan to retrieve the keys chunked (1000) are they returned in a special order ?
20:13:16 GMT <minus> ledil: no special order, and setting the limit to 1000 will get you at most 1000 iirc. scan has the performance penalty of having to scan the entire key space fyi
20:16:37 GMT <ledil> what is the best way to fetch them all, actually ive got over 100k items that I need to fetch ... I supposed it would be better to use scan with a limit/counter
20:22:47 GMT <ledil> minus: any ideas ?
20:22:48 GMT <minus> it's best if you know what to access. the next best thing is scan. you can fetch multiple values at once using mget
20:23:07 GMT <ledil> ah ok ... good idea
20:23:11 GMT <ledil> so scan + mget
20:23:17 GMT <minus> if it's a one-off operation scan is fine
20:23:44 GMT <minus> if you do it frequently you might run into problems; but you'll find that out anyway
20:24:47 GMT <ledil> why might I run into problems ?
20:25:23 GMT <ledil> the keys are cached values that I need to fetch frequently ...
20:25:50 GMT <minus> scan isn't cheap
20:26:13 GMT <minus> but then again, processing 100k items isn't cheap either
20:26:40 GMT <minus> will probably take a couple of seconds to scan and get those
20:27:56 GMT <ledil> the time doesnt matter as I will fetch only 1000 items at times ... my call is using a limit/offset that the user must use
20:28:10 GMT <ledil> so I will never fetch more than 1000 items at a time
20:28:32 GMT <minus> aye
20:30:40 GMT <minus> i'm not sure how stable scan iterators are (if you intend to pass those on to the user-facing side), and also you will get varying number of results (see https://redis.io/commands/scan#the-count-option)
20:31:33 GMT <minus> you can create an index for your data in a sorted set (ZSET), which will give you control over ordering and precise number of results
20:31:40 GMT <minus> but you'll have to manage that index