00:51:10 GMT <lunaa> hi :)
00:51:11 GMT <lunaa> with rejson
00:51:23 GMT <lunaa> am i able to pass whole objects to JSON.SET ?
00:51:30 GMT <lunaa> multilevel ones
08:04:36 GMT <minus> lunaa: looks like yes, just check out the docs, specifically http://rejson.io/path/
13:15:42 GMT <k1ng440> hi
13:15:48 GMT <k1ng440> anyone has experience with stunnel and redis?
13:17:25 GMT <k1ng440> http://paste.ubuntu.com/25146281/
13:46:52 GMT <k1ng440> can anyone help me please?
14:25:00 GMT <Habbie> k1ng440, what's the problem?
14:33:14 GMT <k1ng440> i fixed it. i was trying out sentinel but couldnt make it work
14:33:16 GMT <k1ng440> so i gave up
14:33:34 GMT <k1ng440> habbie, will the slaves work if the master went down?
14:33:41 GMT <k1ng440> the slaves are writable
14:33:56 GMT <Habbie> uhm, i suppose so then
14:34:38 GMT <Habbie> they should, yes
14:35:05 GMT <k1ng440> and the master will be updated when its back up right?
14:35:34 GMT <Habbie> from the slave? that sounds weird
14:35:52 GMT <k1ng440> writiable slave
14:35:58 GMT <Habbie> oh, you mean you promote a slave to master when the master goes down?
14:36:09 GMT <k1ng440> no
14:36:19 GMT <Habbie> then no, there is no replication back from the slave to the master
14:36:21 GMT <k1ng440> i cant get the sentinel to work since i am using stunnel
14:36:53 GMT <k1ng440> so the master will never be updated even the master is up?
14:37:08 GMT <Habbie> as far as i can tell, that's correct
14:38:01 GMT <k1ng440> can i setup multi master?
14:38:09 GMT <Habbie> no, redis does not have this
14:38:13 GMT <Habbie> sentinel is the only thing that comes close
14:38:36 GMT <k1ng440> even i use sentinel, the master will not be updated from slave
14:38:39 GMT <k1ng440> goddamn it
14:38:45 GMT <Habbie> if you use sentinel, the slave will be promoted to master
14:38:53 GMT <Habbie> and the master that went down will be a slave when it comes back up
14:38:56 GMT <Habbie> as far as i understand
14:39:06 GMT <k1ng440> i understand that too.
14:39:20 GMT <k1ng440> i just need a cluster that will sync both way.
14:39:27 GMT <Habbie> why?
14:39:37 GMT <k1ng440> since the redis are in multiple location over the earth
14:40:08 GMT <Habbie> ah
14:40:18 GMT <Habbie> i don't think redis can do what you want
14:40:20 GMT <Habbie> (but i could be wrong!)
14:40:49 GMT <k1ng440> then why slave is writiable?
14:40:57 GMT <Habbie> because you configured it to
14:41:20 GMT <Habbie> by default slaves are not writable
14:42:55 GMT <k1ng440> so basically i have to write or find a script that will sync multiple redis
14:46:59 GMT <Habbie> that would be one way, yes
14:49:23 GMT <k1ng440> @habbie, i am confused about sentinel
14:49:33 GMT <k1ng440> if i setup sentinel on all servers
14:50:05 GMT <Habbie> yes?
14:50:07 GMT <Habbie> brb
14:50:11 GMT <k1ng440> sentinel monitor redis-cluster 127.0.0.1 6378 2
14:50:22 GMT <k1ng440> it will be same on all servers?
14:50:44 GMT <k1ng440> or it has to be pointed to master node?
14:51:54 GMT <Habbie> https://redis.io/topics/sentinel
14:51:57 GMT <Habbie> please read this full page
14:52:09 GMT <Habbie> specifically the text under the configuration example in the Configuring Sentinel section
14:53:37 GMT <k1ng440> well its hard for me to understand. i did try to read but complex to understand.
15:02:50 GMT <Habbie> you point sentinal at the master
15:02:53 GMT <Habbie> it will discover the slaves
15:07:06 GMT <k1ng440> i understand now. thank you
22:22:32 GMT <kixelated> hey guys, I'm running redis cluster 3.2.8
22:22:40 GMT <kixelated> on bare-metal
22:23:24 GMT <kixelated> we've disabled rdb persistence but some instances persist anyway for some reason
22:24:32 GMT <minus> they do for replication, kixelated
22:25:09 GMT <kixelated> oh huh so a slow slave would cause it?
22:25:42 GMT <minus> any slave afaik
22:26:00 GMT <minus> there's the diskless-replication option that might prevent that
22:27:24 GMT <kixelated> we've got 16 masters and 16 slaves, two of the masters started to persist every few seconds today
22:27:40 GMT <kixelated> while other instances haven't persisted for weeks
22:27:49 GMT <minus> i don't actually have any experience with those things though, so i can't really help
22:27:58 GMT <kixelated> no it's been good insight
22:28:02 GMT <kixelated> thanks!
22:31:01 GMT <kixelated> actually it makes sense, the hosts get super overloaded at peak
22:31:15 GMT <kixelated> slaves blip and reconnect to the master
22:31:28 GMT <kixelated> which triggers a disk replication
22:31:54 GMT <kixelated> and seems to create a chain reaction