02:41:46 GMT <trufa> hey there! I've got a warning just like the one on this issue: https://github.com/NodeRedis/node_redis/issues/1081 but I'm not quite sure what I should be doing about it, as far as I understand it will work on redis 2.x but should I already be doing it differntly? how would that be?
02:48:06 GMT <trufa> btw, I'm use node_redis, and doing a simple proof of concept hmset()
12:00:46 GMT <john_dee> Hey #
12:01:03 GMT <john_dee> Can someone point me in the direction of the AOF format description?
12:01:24 GMT <john_dee> What does the *digit mean?
12:43:37 GMT <john_dee> Looks like the length of the transaction, right?
12:46:23 GMT <chargen> Hi. I'm setting up three servers with sentinel but it doesn't seem to failover when I the master sleep for a while.
12:47:18 GMT <chargen> this is part of the log: https://gist.github.com/jmolero/6eaef012f4cb27f2330f3004f0370aa5
12:47:34 GMT <chargen> I see a -sdown there but doesn't go any further
12:47:59 GMT <chargen> # -sdown master mymaster 192.168.99.101 6379
13:15:32 GMT <john_dee> chargen: If I were you I’d just google)
13:15:44 GMT <john_dee> This chan is like a nuclear wasteland %)
13:15:58 GMT <chargen> I'm already googling
13:16:32 GMT <john_dee> Haven’t worked with Sentinel, so can’t help. Sorry.
13:16:52 GMT <chargen> Ok. Don't worry. Thanks anyway.
13:17:18 GMT <chargen> Maybe it's because I just set only one sentinel master
13:26:11 GMT <chargen> I saw a problem. I was monitoring the redis server instead of redis sentinel :S
14:05:42 GMT <madsage> morning
14:06:33 GMT <madsage> hey I am having redis master/slave issues. I cant get my 3 node , 1master, 2slaves, 3 sentinals to fail over.
14:07:00 GMT <madsage> each sentinal reports the others down.  +sdown sentinal
14:07:25 GMT <madsage> they can talk to eachother. and when I try to connect to the sentinals. it tells me i'm in protected mode.
14:07:43 GMT <madsage> but i've set the config protected-mode no
14:08:08 GMT <madsage> and from the redis-cli.
14:08:50 GMT <madsage> any ideas? I have a deadline for this in a couple hours. pulling my hair out over here
14:08:57 GMT <madsage> heh
14:09:12 GMT <madsage> redis 3.2.1
14:10:14 GMT <madsage> i notice some mismatch ID's not sure how to resolve those. i've blown the config and dump out a few times and started over
14:11:55 GMT <chargen> madsage: If you have set up a password on the server you need to add that to the sentinel config: "sentinel auth-pass mymaster yourpassword"
14:19:57 GMT <madsage> chargen, yes
14:20:04 GMT <madsage>  redis-cli -h 10.53.42.81 -p 26379 10.53.42.81:26379> info DENIED Redis is running in protected mode because protected mode is enabled
14:20:18 GMT <madsage> connecting to another sentinal
14:20:39 GMT <madsage> sentinel*
14:21:04 GMT <chargen> I think that's normal because you didn't authorize that client
14:21:14 GMT <chargen> I might be wrong
14:21:19 GMT <madsage> yes I have requirepass, masterpass and auth=pass all set the same
14:22:08 GMT <madsage> in redis.conf and sentinel auth-pass on each sentinel
14:22:09 GMT <chargen> But I think that password is used by your sentinel, not when you connect with a client
14:22:21 GMT <madsage> ahh ok
14:22:33 GMT <chargen> see the info on the local sentinel
14:22:53 GMT <chargen> It should show how many sentinels are connected and who is the master
14:23:02 GMT <madsage> it does
14:23:06 GMT <madsage> everything looks good
14:23:43 GMT <chargen> I'm also trying to figure out how to get it working. Seems that this channel isn't very active.
14:24:03 GMT <madsage> master0:name=mem02-sl-sj,status=ok,address=10.53.42.82:6379,slaves=2,sentinels=3
14:24:18 GMT <chargen> yes I have the same
14:24:32 GMT <chargen> My problem is that it doesn't failover
14:24:36 GMT <madsage> all i get in the logs is +sdown sentinal reports
14:24:41 GMT <madsage> yeah same here
14:25:07 GMT <chargen> I think that's ok. +sdown just checks the status
14:25:25 GMT <chargen> wait
14:25:26 GMT <chargen> no
14:25:56 GMT <chargen> do you get lots of +sdown ?
14:26:27 GMT <madsage> doesnt failover.
14:26:50 GMT <madsage> no but after the +slaves I do see +sdown
14:26:51 GMT <chargen> My problem was I had the redis port in the sentinel conf instead of the sentinel port
14:28:26 GMT <chargen> you need to use the sentinel port. Mine is 26379
14:28:45 GMT <madsage> mine look good i think
14:29:03 GMT <chargen> I see that port on your paste 6379
14:29:19 GMT <chargen> that is normally the redis-server port
14:29:59 GMT <madsage> which paste. master0:name=mem02-sl-sj,status=ok,address=10.53.42.82:6379,slaves=2,sentinels=3 ?
14:30:05 GMT <chargen> yes
14:30:26 GMT <madsage> those is the master redis server
14:30:42 GMT <madsage> ip and redis port 6379
14:30:48 GMT <chargen> I think it has to point to the redis sentinel
14:31:43 GMT <madsage> sentinel monitor mem02-sl-sj 10.53.42.82 6379 2
14:32:09 GMT <madsage> master is set proper in sentinel config
14:32:36 GMT <chargen> but 6379 isn't the sentinel port, is it?
14:32:43 GMT <madsage> no
14:33:27 GMT <madsage> port 26379 daemonize yes pidfile "/var/run/redis/redis-sentinel.pid"
14:35:21 GMT <chargen> yeah, sorry. I think it has to be the redis port
14:36:11 GMT <madsage> hmm
14:36:56 GMT <madsage> nothing in the configs say the sentinels are anywhere else other then 26379
14:37:21 GMT <madsage> only the masters and slave redis on 6379
14:38:47 GMT <madsage> i can put the announce ip and port in the sentinel configs. but im not using containers or nat.
14:39:10 GMT <madsage> internal private network
14:39:13 GMT <madsage> obviously
14:39:33 GMT <madsage> and no firewall restrictions between machines
14:40:17 GMT <madsage> something changed. 2.8 would failover. but not back.
14:40:33 GMT <madsage> so i upgrade to 3.2.1 and now cant failover. lol
14:40:55 GMT <madsage> supposedly a bug in 2.8 prevented failback
14:41:17 GMT <madsage> or rollback whatever when the master showed up
14:50:21 GMT <madsage> maybe the dev branch works
14:50:27 GMT <madsage> hah
14:50:31 GMT <chargen> I'm at 3.2.1 and still no luck
14:50:54 GMT <madsage> same. there must be problem. i'm sure my configs are good.
14:51:28 GMT <chargen> I'm using docker though
14:51:46 GMT <madsage> 2.8 branch would failover but not rollback when the real master showed up. now no failover period
14:51:55 GMT <madsage> unless there is some change that is not documented
14:52:02 GMT <madsage> or I havnt found it
14:52:21 GMT <madsage> i need to buy more time. omg this is killing me.
14:52:35 GMT <madsage> deadline was over a week ago
14:53:19 GMT <madsage> i think I will try 3.3 branch
14:54:15 GMT <madsage> hey quick question. if i blow out the dump file on all 3 nodes. it should created all new ID's right? or am I missing some place where these are stored
14:55:14 GMT <madsage> becasue I do see one ID mismatch when I cycle the master. says cant resync partial. found ID blahblah but expected ID blahblahblah
14:55:38 GMT <madsage> maybe the problem?
14:56:34 GMT <madsage> i've blow out dump.rdb on all 3 several times
14:57:00 GMT <madsage> and started over. also removing ID from sentinals configs
14:57:14 GMT <madsage> sentinels. heh cant spell. :]
14:57:27 GMT <madsage> i will try it again
15:04:43 GMT <madsage> yeah nuts! I cannot get -down sentinals no matter what
15:05:07 GMT <madsage> i get -sdown slaves from each sentinel
15:05:26 GMT <madsage> and -sdown master
15:05:53 GMT <madsage> info shows proper number of slaves and sentinels
15:06:31 GMT <madsage> but still..   +sdown sentinals in the logs for every sentinel looking at eachother
15:09:18 GMT <chargen> yep, me too: 459:X 11 Jul 15:08:17.093 # +sdown sentinel 30243a6fc6d3807a33b3df120b4bc1b5d87b1af8 192.168.99.102 26379 @ mymaster 192.168.99.101 6379
15:09:19 GMT <chargen> 459:X 11 Jul 15:08:17.095 * +sentinel-address-switch master mymaster 192.168.99.101 6379 ip node3 port 26379 for b5123f3140ed4e2835ad69e368d7f522ff08793f
15:10:17 GMT <madsage> this is nuts. haha
15:10:55 GMT <madsage> i've floowed so many howto's they all do the same and are simular configurations.
15:11:06 GMT <madsage> followed
15:11:21 GMT <madsage> atleast 2.8 branch did failover for me.
15:11:51 GMT <madsage> seen elections an vote. all worked well.
15:12:26 GMT <madsage> just no vote to return to master when it came back online
15:13:39 GMT <madsage> i think I will try 3.3 branch
15:51:58 GMT <chargen> madsage:  I no longer get the +sdown. This is what show now for every line 477:X 11 Jul 15:51:50.955 * +sentinel-address-switch master mymaster 192.168.99.101 6379 ip node1 port 6379 for b4904f46b8f7065ad2413da0af2c42771bac5887
15:52:21 GMT <chargen> I just logs and logs that address switch
15:54:40 GMT <madsage> hmm
15:57:03 GMT <horatiohb> The doc for PUBLISH (http://redis.io/commands/publish) doesn't specify the type of message. Should I assume it's required to be a string?
16:26:43 GMT <madsage> chargen, you gets your working?
16:26:52 GMT <chargen> nope
16:27:52 GMT <madsage> hah wish one of the devs would enlighten us. is there a known working build? with master/slave fail-over and rollback?
16:52:47 GMT <chargen> madsage: do you use "bind" in your sentinel config?
16:53:12 GMT <chargen> I got it working once in some servers and it's config has it
16:53:39 GMT <chargen> I'm trying to get it working again on docker to avoid all that pain next time :)
16:54:29 GMT <chargen> I've tried binding the address but I get that error: 946:X 11 Jul 16:51:43.654 # Creating Server TCP listening socket 192.168.99.101:26379: bind: Cannot assign requested address
17:22:16 GMT <chargen> madsage: are you still here? :)
17:54:24 GMT <madsage> chargen, yes
17:55:17 GMT <madsage> had to explain why I just missed my 10am testing with validation team now wondering why i'm dumb
17:55:18 GMT <chargen> Idk what to do, nothing seems to work
17:55:19 GMT <madsage> haha
17:55:38 GMT <madsage> i know.
17:55:56 GMT <madsage> so hey.. i am rolling back to 2.8
17:56:18 GMT <madsage> atleast I seen a promotion of slave to master
17:57:02 GMT <madsage> i'm using HAProxy so utilizing tcp-option and some expect I can determine which node is master.
17:58:12 GMT <madsage> obviously we dont want to write to a slave. even with readonly no
17:59:41 GMT <madsage> this will buy me some time for more work in dev environment. maybe this project will catchup on sentinel issues later. ultimatly I want to cluster but our application is built around servicestack connector that doesnt support that.
18:00:30 GMT <chargen> I see. Maybe I'll downgrade too.
18:01:04 GMT <madsage> i think it was redis 2.8.7  i last seen slave to master promotion working through sentinel.
18:01:19 GMT <chargen> Sentinels can talk to each other but for some reason none of them trigger the odown
18:01:42 GMT <madsage> whatever is in ubuntu 14.04 repo. give me a sec. i will let you know. in case you want to try it also
18:02:42 GMT <chargen> I'm checking the server where I had it working, it's at v=3.0.7
18:05:27 GMT <madsage> Redis server v=2.8.4 sha=00000000:0 malloc=jemalloc-3.4.1 bits=64 build=a44a05d76f06a5d9
18:05:38 GMT <madsage> that one worked for me.
18:05:49 GMT <chargen> thanks I'll try it
18:06:26 GMT <madsage> sentinals actually promoted a slave to master. (atleast)
18:52:13 GMT <madsage> chargen, it worked right away.
18:52:19 GMT <madsage> i knew it wasnt my configs. =]
18:53:23 GMT <madsage> soon as I dropped the master. a slave took over.
18:53:39 GMT <madsage> sentinels promoted... rather
18:56:43 GMT <chargen> Good to know. I'll try it now thanks :)
19:09:46 GMT <madsage> so hey, do you know how to return master role back to my master node?
19:10:29 GMT <madsage> i think slaveof no one
19:10:39 GMT <madsage> but that will partition my replication.
19:11:00 GMT <madsage> they is why I was chasing a fix so it would auto return
19:11:48 GMT <madsage> s/they/this
19:16:36 GMT <chargen> madsage: seems that apt version is 3.0.6 for redis-server
19:16:46 GMT <chargen> that's what I used last time
20:53:06 GMT <danilobrl> hello  all
20:53:13 GMT <danilobrl> anyone can help me?
20:54:07 GMT <minus> not if you don't ask something
20:54:13 GMT <danilobrl> -failover-abort-not-elected master mymaster 10.49.1.223 6389
20:56:03 GMT <danilobrl> when the master dies, the slave no assumes
20:56:15 GMT <danilobrl> i have 3 servers with sentinel.
20:56:32 GMT <danilobrl> +try-failover master mymaster 10.49.1.223 6389
20:56:38 GMT <danilobrl> +vote-for-leader 35e094f7d538402449aa49160e4c603163a75279 16
20:56:45 GMT <danilobrl> -failover-abort-not-elected master mymaster 10.49.1.223 6389
20:56:53 GMT <danilobrl> Next failover delay: I will not start a failover before Mon Jul 11 17:49:12 2016
20:56:58 GMT <danilobrl> this is my error message
20:57:03 GMT <minus> dunno
20:57:24 GMT <minus> but maybe someone else does
21:00:38 GMT <danilobrl> Anyone?
21:07:45 GMT <truthadjustr> gentlemen, do can I prevent redis from writing a dump.rdb? I don't use *.conf, but instead custom-built 3.2.0. I already disabled *save lines* in server.c and my current maxmemory is '0'. But still, i am seeing dump.rdb.. My redis is still consuming 25MB
21:10:52 GMT <minus> are you sure you're not loading a (wrong) config file?
21:12:26 GMT <truthadjustr> i am sure.. i have custom-built three versions for x86, powerpc and s390x architectures
21:14:46 GMT <truthadjustr> my cache usage is only a max of 50MB, but highly secured data.. so i'm custom bulding
21:18:31 GMT <truthadjustr> 'info memory' --> http://sprunge.us/UXCW
21:19:54 GMT <minus> are you sure you're running the right executable?
21:20:10 GMT <minus> connecting to the right machine?
21:20:22 GMT <minus> eh, scratch latter question
21:20:24 GMT <truthadjustr> yes, 'info persistence' -> http://sprunge.us/QRZJ
21:22:56 GMT <truthadjustr> and my setup is: 1 master, three slaves. Could the 'slaveof' caused this writes?
21:23:26 GMT <truthadjustr> im deleting the dump.rdb , and i'm observing if this .rdb file re-appears...
21:23:33 GMT <minus> yeah, could be
21:23:58 GMT <minus> there's a in-memory-only snapshot feature for slaves, maybe give that a try
21:23:59 GMT <truthadjustr> but the .rdb appeard in master, and each of the slaves..