09:11:26 GMT <wk-work> is there a list over sentinel commands?
09:11:50 GMT <wk-work> I'm trying to figure out if you can failover to a specific node using the "failover" command, but I can't find any proper documentation for that anywhere
09:16:53 GMT <badboy_> wk-work: https://redis.io/topics/sentinel#sentinel-commands
09:17:11 GMT <badboy_> no one jumped up and wrote proper docu for the command page yet
09:17:18 GMT <wk-work> badboy_: thanks!
09:17:50 GMT <badboy_> but to answer your question: no, you can't say which node should be picked
09:18:23 GMT <wk-work> even if we've weighted our nodes?
09:18:38 GMT <badboy_> the usual things apply, sure
09:19:14 GMT <badboy_> so slave-priority will help
09:19:23 GMT <badboy_> but you have to set that before doing the failover obviously
09:19:37 GMT <wk-work> yeah, it's set in the config, but it's still just picking a random node, strange
09:22:03 GMT <badboy_> is it?
09:22:23 GMT <badboy_> did you check with CONFIG GET slave-priority that the value is set correctly?
09:22:26 GMT <badboy_> (on the slave nodes
09:22:32 GMT <wk-work> yep, 10, 20, 30 etc
09:23:20 GMT <badboy_> weird indeed
09:25:47 GMT <badboy_> so according to the code it should do the right thing
14:16:24 GMT <film4242424242> Is there a good write-up somewhere of why redis-rb does not plan on supporting redis cluster? I've seen bad-boy (trying not to highlight) reply to a few issues saying there were not plans to support. Just wondering if you're waiting for someone from the community to implement it, or if it's more than that. Thanks!
14:36:29 GMT <badboy_> film4242424242: Time.
14:36:52 GMT <badboy_> and I'm sure it can be implemented on top of redis as its own gem
14:37:32 GMT <badboy_> if the community implements it as part of redis-rb and it is good we might still reconsider merging it (and get that person on board for maintaining it)
14:47:15 GMT <film4242424242> badboy_: Totally understand. Thanks for the response. Do you think think moving to redis cluster now is a good idea, or better to rely on partitioning keys with sentinel for each master?
14:48:10 GMT <badboy_> I moved customers to Redis Cluster because it was appropiate for their use case
14:48:14 GMT <badboy_> it highly depends on what you need
14:48:26 GMT <film4242424242> Yeah. Was just about to add that.
14:48:32 GMT <badboy_> In general I suggest to avoid Cluster if your data fits into a single machine (and your not hitting other bottlenecks)
14:49:23 GMT <film4242424242> In our case we have sentinel with 32GB of mem and we're about 70% full and growing.
14:50:25 GMT <badboy_> We took the route of using a proxy in front of Redis Cluster, so there was no need to have cluster support in the clients
14:50:49 GMT <film4242424242> Yeah. That's what we're thinking as well. You guys using corvus, or something else?
14:59:16 GMT <badboy_> yup
18:17:08 GMT <crayon> is there any way to get a persistent written output of MONITOR so we can ingest this into a siem?
20:26:58 GMT <badboy_> crayon: the AOF file? :)
20:50:41 GMT <crayon> badboy_, is that where it is stored?
20:52:19 GMT <crayon> badboy_, is that where it is stored?
21:04:00 GMT <badboy_> crayon: you need to enable it https://redis.io/topics/persistence#append-only-file
21:14:21 GMT <crayon> does this file log/output commands executed on the redis instance though?
21:14:30 GMT <crayon> that's whats important to me