04:46:48 GMT <gracefu> is it possible to migrate a redis instance from one server to another with absolutely no downtime or data loss? if not, is it possible to set things up such that you can?
16:00:32 GMT <codepython777> anyoneknows how to fix this: https://paste.ubuntu.com/24587596/
16:03:30 GMT <codepython777> oops - my bad - got it running
17:17:57 GMT <rrva_> investigating if I can use redis pub/sub to replicate a dataset from one redis client process to another redis client. For incremental changes this works fine, but what about bootstrapping (getting all changes for a channel, or re-fetching them separately). Is there anything in pubsub which helps with this?
17:38:49 GMT <xintron> Is there an easy way to do a setnx but if the value exist and matches a specific value it will instead increase the TTL in an atomic way. This relates to the distributed lock described here: https://redis.io/topics/distlock but the locks would be held by one "master" application doing all the work until that node dies or does an unlock (on app exit)
19:14:51 GMT <xintron> Why is there a differentiation done on KEYS and ARGS for Lua scripts? Aren't they really the same?
19:18:30 GMT <xintron> Ah, found it. Used for sharding in a cluster setup. Makes sense as you would need to know the keys specifically in such a case.
20:21:50 GMT <stephen87> hello.  i have a redis cluster using sentinel. I want to add a new redis slave to my sentinel config, but i cant figure out how to do it
20:22:31 GMT <stephen87> i tried editing the /etc/redis/redis-sentinel.conf filer but that didnt seem to do the trick
20:22:39 GMT <stephen87> is there some redis-cli command i need to run?
20:31:48 GMT <minus> stephen87: afaik sentinel picks that up automatically when you add it as slave
20:32:07 GMT <minus> also, don't mix up cluster and sentinel, those are two disjoint things
20:32:13 GMT <stephen87> ok
20:32:24 GMT <stephen87> so i see 3 sentinels when i look at info
20:32:30 GMT <stephen87> but i only see 1 master and 1 slave
20:32:38 GMT <stephen87> when i should see 3 sentinels, 1 master, and 2 slaves
20:32:48 GMT <stephen87> how can i make sentinel aware for the 3rd slave?
20:52:48 GMT <stephen87> ah i connected to redis on the new server and did a slave of master.ip port
21:00:02 GMT <codepython777> how do i fix these warnings: https://paste.ubuntu.com/24588991/ ?
21:04:14 GMT <minus> codepython777: 1) increase ulimit -n or lower the backlog setting to 128 2) enable overcommit kernel option 3) disable THP kernel option
21:05:18 GMT <codepython777> minus: Is there a way to tell redis to limit itself to say 32MB and over that swap from disk?
21:05:57 GMT <minus> there's maxmemory; no you can't tell it to swap
21:08:28 GMT <codepython777> Thanks
21:08:46 GMT <codepython777> what is ulimit -n for?
21:08:52 GMT <codepython777> its currently 1024 on my machine
21:08:59 GMT <minus> maximum number of open file descriptors
21:09:09 GMT <minus> wait, that's the wrong one
21:09:46 GMT <codepython777> backlogis set to 511 in the default conf
21:09:47 GMT <minus> might a kernel setting also
21:10:12 GMT <minus> >WARNING: The TCP backlog setting of 511 cannot be enforced because /proc/sys/net/core/somaxconn is set to the lower value of 128.
21:10:21 GMT <minus> it tells you what to do anyway
21:26:46 GMT <codepython777> minus: thanks
21:45:18 GMT <mrgreenfur> I’m ashamed to ask this but here goes: After intermittent connectino problems, i looked in logs and see that about every 2 minutes, SIGTERM is causing redis to shutdown and restart
21:45:41 GMT <mrgreenfur> There is no memory or disk pressure.  On each restart it syncs with a slave
21:45:51 GMT <mrgreenfur> anyone know where to look for the culprit of these SIGTERM ?