15:12:46 GMT <edge226> I'm trying to set up 2 docker containers with replication...
15:13:07 GMT <edge226> All I can manage to get is connection refused messages.
15:13:27 GMT <edge226> the master and slave are on the same host but using different ports.
15:15:18 GMT <edge226> the description about slave-announce-port is unclear whether this needs to be set on the master or slave.
15:18:59 GMT <minus> docker does NAT; you can disabled that by passing --net=host to docker run
15:48:28 GMT <edge226> minus: so would I do that on the master/slave or both?
15:55:14 GMT <minus> both; they do run on separate ports after all, edge226
15:56:01 GMT <minus> on a side note: docker networking is a fucking pain and doesn't work well together with any kind of firewall that changes iptables rules
16:01:14 GMT <edge226> minus: I'm not really doing any of that.
16:01:37 GMT <edge226> minus: I also still get the same error when I do --net=host
16:03:01 GMT <edge226> minus: to confirm I do have access to the redis database at the port referenced to, it just does not do the replication.
16:03:25 GMT <minus> that error means nothing is running on that port/you're using the wrong ip/port
16:03:40 GMT <edge226> minus: but I'm not.
16:04:02 GMT <minus> check what port redis is running on and verify with the slaveof setting
16:04:16 GMT <edge226> I'm pretty sure one setup on 127.0.0.1 and the other setup with 127.0.0.1 works... The ports are the same.
16:04:23 GMT <minus> ss -tlpn | grep redis
16:06:06 GMT <edge226> do that where?
16:06:28 GMT <minus> on the host, that should show both redises
16:09:18 GMT <edge226> I launched the master to see and it does not.
16:10:24 GMT <aah> morning all.  question re: psubscribe and when redis stops sending messages due to internal buffer overflow / connection latency.  basically, it looks like my connection stays open forever, even though the redis instance will no longer send me messages.  is that .. unexpected?  is there anything that would let me know that I'm no longer going to get data?
16:11:37 GMT <minus> edge226: oh, you might have to run ss as root to see process names
16:12:22 GMT <minus> aah: i'd expect it to kill the connection, though i don't really know
16:12:30 GMT <edge226> minus: so doing this I will have to change the port expose in the images right?
16:13:06 GMT <aah> minus: yeah, that's what I expected too. :P  I'll poke around.
16:13:56 GMT <minus> edge226: possibly, depends on how the image works
16:14:58 GMT <edge226> minus: well the default redis image exposes 6379 by default.
16:16:37 GMT <minus> unless you can somehow get it to start redis-server with a different --port it won't work
16:19:29 GMT <edge226> minus: ok it just worked lol.
16:20:58 GMT <edge226> minus: thanks a lot I've been trying to "solve" that for literally hours.
18:02:24 GMT <rubas> Hi, have anyone had the experience where a terminal says: " accept: Too many open files "
18:16:20 GMT <minus> rubas: you might have too many open connections
18:16:41 GMT <rubas> minus: to redis?
18:17:00 GMT <rubas> Okay - I'll try to limit it, will quickly test, thank yoy
18:17:02 GMT <rubas> you *
18:17:33 GMT <minus> yes, might be redis, might be something else. i don't know what you're doing so i can't tell
18:17:45 GMT <rubas> minus: makes very good sense
18:18:07 GMT <minus> uname -n tells you the limit of open files (though it can vary depending on where you started your processes from)
18:18:50 GMT <rubas> minus: you're right.
18:19:05 GMT <rubas> minus: it fixed the problem, I started closing connection, and then reopening it
18:19:15 GMT <rubas> instead of just opening
18:19:21 GMT <rubas> minus: thank you very much!
18:19:56 GMT <minus> depending on how your application looks you might be able to keep one (or a pool of) connection open
20:01:45 GMT <tmberg> "uname -n tells you the limit of open files" =)
20:05:55 GMT <minus> yes?