09:10:00 GMT <shmoon_> anyone on ? need some help
12:03:33 GMT <virhilo> hi guys
12:03:57 GMT <virhilo> questions about particular language clients implementations are ok here too? :)
12:59:08 GMT <unj> hey
12:59:39 GMT <unj> is there a redis-idiomatic way of dealing with 2 agents writing to the same key in redis?
12:59:52 GMT <unj> like transaction type of stuff, if I had lets say array of ints cached
13:00:15 GMT <unj> and 2 agents would like to add and int to that array (each agent a different one obviously)
15:01:38 GMT <ren0v0> Hi can anyone help with this?
15:02:20 GMT <ren0v0> 18583:X 07 Feb 15:00:42.100 * +sentinel sentinel 4358d005d5556f9a0644e6e831de851ce54894d3 ip 26380 @ mymaster ip 6380
15:02:21 GMT <ren0v0> 8583:X 07 Feb 15:01:12.152 # +sdown sentinel 4358d005d5556f9a0644e6e831de851ce54894d3 ip 26380 @ mymaster ip 6380
15:02:36 GMT <ren0v0> it recognises the sentinel, but then its flagged as down ?
15:03:41 GMT <ren0v0> pulling my hair out!
15:11:11 GMT <badboy_> ren0v0: it recognizes the monitored instance as subjectively down
16:08:41 GMT <bsamorim> Hello fellas, I've got a pretty simple question: does reddis require all of my DB to be on RAM? If not, what are the parameters and their values that would server as threshold to Redis applicability to a given DB?
16:08:52 GMT <bsamorim> that would werve*
16:08:55 GMT <bsamorim> serve*
16:10:58 GMT <minus> bsamorim: yes, it does
16:11:23 GMT <bsamorim> minus: :( :( :(......ok thanks very much!
16:12:28 GMT <bsamorim> is there some sort of rule of thumb that would determine how much space would my dataset take with redis?
16:13:00 GMT <bsamorim> more broadly, can I expect significant differences between the space taken in SQL DB's and Redis ones?
16:14:42 GMT <ren0v0> badboy_, isn't it saying the sentinel is down ?
16:15:40 GMT <badboy_> no
16:15:49 GMT <badboy_> sdown is an event about monitored masters
16:16:06 GMT <ren0v0> badboy_, i've been on this for 2 days, and can't get it working. I have 3 servers, 3 instances of redis-server and 3 sentinels on each. and this is the log  >>  http://pastebin.com/Ra1NVgQK
16:16:32 GMT <ren0v0> but master is running  >>    redis    18300  0.1  0.2  38748 11296 ?        Ssl  14:51   0:08 /usr/bin/redis-server 0.0.0.0:6380
16:17:33 GMT <badboy_> is ip1 a valid ip the sentinels can reach?
16:17:40 GMT <ren0v0> And, if i run  "sentinel sentinels mymaster" i see this   >     10) "s_down,sentinel,disconnected"
16:18:26 GMT <ren0v0> badboy_, yes.
16:19:27 GMT <ren0v0> check this >   http://pastebin.com/3BXqgg64
16:20:19 GMT <ren0v0> and then sentinels   >>   http://pastebin.com/dhhGY4C6  (only pasted 1)
16:20:49 GMT <badboy_> oh wait, let me refresh my sentinel knowledge quickly
16:20:58 GMT <ren0v0> the docs absolutely suck
16:20:59 GMT <badboy_> are ip2 and ip3 also valid?
16:21:09 GMT <badboy_> "absolutely" is quite harsh but I can agree :D
16:21:12 GMT <ren0v0> yes, i have mongo replication working across the same 3 servers just fine
16:21:43 GMT <badboy_> can you telnet to the sentinel from each other server running a sentinel?
16:22:00 GMT <ren0v0> i can ping, i don't use telnet but can try something specific if you need !
16:22:07 GMT <ren0v0> ports are open etc
16:22:22 GMT <badboy_> are you sure? did you try?
16:22:23 GMT <badboy_>  ;)
16:22:37 GMT <ren0v0> to ping? yes, nmap ports? yes
16:23:03 GMT <ren0v0> i've spent 6 hours today debugging firewall, to find out that flushing all rules makes no difference, so it must be redis/configs!
16:23:17 GMT <badboy_> as i don't know your config
16:23:28 GMT <badboy_> easiest would be to start from scratch once more
16:23:34 GMT <ren0v0> i've already pasted it all into ##networking, let me get them
16:23:40 GMT <badboy_> bring down instances, remove sentinel config, restart
16:24:29 GMT <ren0v0> remove sentinel config? (yes its annoying that it keeps rewriting mine)
16:25:20 GMT <badboy_> it persists the state in the config file, yes
16:25:44 GMT <ren0v0> here is my config before and after >     http://pastebin.com/CYwc4gtf
16:26:49 GMT <badboy_> that looks just fine, yep#
16:27:32 GMT <badboy_> meh, my dev setup for sentinel is broken
16:28:52 GMT <badboy_> ooooh wait
16:28:57 GMT <badboy_> I made a mistake
16:28:59 GMT <badboy_> you are right
16:29:07 GMT <badboy_> +sdown sentinel reports the sentinel is down
16:29:21 GMT <*> badboy_  checks his customer's instances
16:30:03 GMT <badboy_> ren0v0: what about the other 2 sentinels, do they report that the other sentinel is down as well?
16:32:29 GMT <ren0v0> let me check
16:32:53 GMT <ren0v0> yes
16:33:53 GMT <badboy_> then for some reason that I don't know packets between those instances are lost
16:34:01 GMT <badboy_> which ip do the others report about the down instance?
16:34:51 GMT <ren0v0> http://pastebin.com/Ra1NVgQK
16:34:54 GMT <ren0v0> this is all i see in the logs
16:35:27 GMT <ren0v0> and on each respective server its the same, but with the other 2 servers ips
16:35:33 GMT <ren0v0> i added verbose logs, and i see this   http://pastebin.com/hAXjYt77
16:35:37 GMT <ren0v0> additionally
16:38:06 GMT <badboy_> something with your network is flaky
16:39:26 GMT <Mahesh_> A newbie question..  Trying to access Redisover a load balanced VIP where in there is only one active server at a time. Is there any special configuration needed either in the VIP (F5) configuration or in the Redis configuration ?
16:42:28 GMT <ren0v0> badboy_, like what? i don't see anything wrong
16:42:35 GMT <badboy_> Redis does^^
16:42:37 GMT <ren0v0> the only thing that isn't working here is redis
16:42:51 GMT <ren0v0> badboy_, it's not telling me anything is wrong, what am i expected to do ?
16:43:06 GMT <badboy_> then drop down to packet dumps and see if you can spot anything suspicious
16:43:12 GMT <badboy_> at least that's what I would do
16:43:15 GMT <badboy_> also I have to leave now
16:43:17 GMT <ren0v0> nothing
16:43:20 GMT <ren0v0> it sucks
16:44:10 GMT <badboy_> it sure as hell does
16:48:36 GMT <ren0v0> Adding a new Sentinel to your deployment is a simple process because of the auto-discover mechanism implemented by Sentinel.     <  YEA RIGHT
16:50:10 GMT <badboy_>   "(given that your network holds)"
16:50:25 GMT <badboy_> also: they auto-discovered each other :P
16:51:13 GMT <ren0v0> where can i find a list of ALL sentinel config options?
16:51:17 GMT <ren0v0> i mean man, the docs are soooo bad
16:52:14 GMT <badboy_> http://download.redis.io/redis-stable/sentinel.conf
16:52:59 GMT <ren0v0> well, nothing there i don't have, and it doesn't include some stuff that sentinel itself writes to my config
16:53:14 GMT <ren0v0> i'm deleting the configs and starting again, bringing them up 1minute apart
16:53:40 GMT <badboy_> because what sentinels writes is state, not config
16:54:05 GMT <ren0v0> then why put it in the f**** config ?
16:54:12 GMT <ren0v0> its just annoying
16:54:29 GMT <ren0v0> call it sentinel.state
16:54:38 GMT <badboy_> it needs to be put somewhere. that it is in the config file is IMO one of the bad decisions indeed
16:54:45 GMT <ren0v0> everytime i run it, i have to change my config
16:54:51 GMT <badboy_> but given that I'm not responsible for it...
16:55:25 GMT <ren0v0> its removing lines from mine....
16:56:11 GMT <badboy_> it might remove default values afaik
16:56:36 GMT <badboy_> it will not remove changed values and if it does that's a bug
16:57:50 GMT <ren0v0> i'll test that. The .conf you linked doesn't include  "loglevel", yet this works and i needed it, so why aren't all configs there ?
16:59:31 GMT <badboy_> some are valid across all modes
16:59:45 GMT <badboy_> but yeah, they should still be added there
17:00:02 GMT <ren0v0> well yea because "some" doesn't tell me which :D
17:00:25 GMT <ren0v0> you are right about rewrite removing "default" values, but thats stupid as well, because i have to dig around again and find the configs so i can edit them...
17:00:27 GMT <badboy_> likely all that don't touch data storage things^^
17:00:42 GMT <ren0v0> i've just found protected-mode and bind directives in this config
17:01:07 GMT <ren0v0> i didn't set this, sentinels still discovered each other according to logs, so not sure why they would be needed?
17:01:38 GMT <badboy_> protected mode is not for sentinels
17:01:49 GMT <badboy_> and bind defaults to 0.0.0.0
17:02:03 GMT <ren0v0> thats not what the config says
17:02:19 GMT <badboy_> not?
17:02:25 GMT <ren0v0> and by default bind defaults to open? thats also not what it says
17:02:29 GMT <ren0v0> top part  http://download.redis.io/redis-stable/sentinel.conf
17:03:02 GMT <badboy_> oh i see
17:03:09 GMT <badboy_> i was wrong again (and tired)
17:03:12 GMT <badboy_> sorry
17:03:23 GMT <badboy_> which version of redis/sentinel are you running?
17:03:56 GMT <badboy_> ok pretty new
17:04:03 GMT <ren0v0> yea 3.2.6 i think
17:04:05 GMT <badboy_> anyway, gotta go. happy debugging!
17:04:11 GMT <ren0v0> ok, turned on this protected-mode,    no sdown yet
17:04:28 GMT <ren0v0> docs don't even tell me what it does :D :D
17:04:29 GMT <ren0v0> thanks! cya
17:17:31 GMT <ren0v0> badboy_, for your interest, it seems some "configs" or "states" however you want to call them, (to me sentinel master <name> <ip> is a config because you set it prior to running), are overwritten, especially the one i just mentioned because its used to store master's host
17:18:11 GMT <ren0v0> i thought the master i set would always be re-elected once it came back online
17:18:30 GMT <ren0v0> assumption based on how mongodb works that was all, its annoying its not the case because my configs will never match the right master :D
20:40:01 GMT <badboy_> ren0v0: mongo promotes back a re-joining old master? o_O
20:40:06 GMT <badboy_> that sounds ... horrible
20:40:16 GMT <badboy_> (afaik it did not do that in my setup)