06:12:50 GMT <cdembue> Redis Hackathon n SF??? http://bit.ly/2eMrlSg. Anybody Going?
06:21:15 GMT <BHC_> um... any body here?
06:22:35 GMT <BHC_> am i alone in this channel?
08:59:53 GMT <xmj> moin
09:00:05 GMT <xmj> what's a good way to install redis on amazon linux?
09:00:52 GMT <xmj> in other words - is there an RPM repo for 3.2.4 somewhere?
09:26:34 GMT <darren``> hello...any one?
09:26:52 GMT <xmj> (nevermind, found it)
09:34:57 GMT <darren``> except the slave node which provide readonly feature, is there other method to make connection readonly?
09:43:14 GMT <minus> you can disable all writing commands by renaming them, darren``
09:47:01 GMT <darren``> oh i see...thanks
09:47:43 GMT <darren``> btw, read only really cannot used on non-slave node? <minus>
09:50:27 GMT <minus> not afaik
09:59:16 GMT <shaharmor> Hey! Is there anything planned for PubSub in the upcoming cluster changes?
15:19:09 GMT <Faizal> Hi
15:20:12 GMT <Faizal> I am using redis on my ecomerce site
15:20:41 GMT <Faizal> Getting issues such as page not found or not able to login etc.
15:21:01 GMT <Faizal> Gets max concurrency in log
15:22:12 GMT <Faizal> Max
15:22:36 GMT <minus> are you not closing connections/not using connection pool?
15:23:36 GMT <Faizal> Hi
17:56:57 GMT <dgaff> Hey all - over the weekend I used HSET to concurrently run many processes that were storing IDs into memory. I'm looking now and there's a ton missing... any initial thoughts from anyone?
17:59:13 GMT <svm_invictvs> dgaff Do you ever remove them?
18:00:55 GMT <dgaff> svm_invictvs: nope
18:01:21 GMT <dgaff> every op is atomic, yeah? If you have two procs racing to update the same hash it should keep both right?
18:01:27 GMT <svm_invictvs> dgaff Are you setting something based on dependent information?
18:01:47 GMT <svm_invictvs> dgaff Yes, I believe how it works it that only one thread may write at a time.
18:02:12 GMT <svm_invictvs> The HSET itself is atomic
18:02:29 GMT <svm_invictvs> but a read and subsequent write is not.
18:02:34 GMT <dgaff> svm_invictvs: I have several hundred procs, each with a partition of IDs, each hset'ing
18:02:51 GMT <dgaff> I do this until all have been hset, then I do my read stage, walking through and checking for existence
18:02:57 GMT <svm_invictvs> Yeah
18:03:02 GMT <svm_invictvs> hrm....
18:03:05 GMT <dgaff> that sounds safe, right?
18:03:16 GMT <dgaff> I'm hashing the key into something, so I'm checking that hashing process now
18:05:01 GMT <svm_invictvs> yeah
18:05:08 GMT <svm_invictvs> I have a hard time visualizing it
18:07:24 GMT <minus> if your hashing produces collisions you're shit out of luck
18:07:55 GMT <minus> redis doesn't lose keys, that'd be a major bug that would've been found by someone pretty surely
18:08:21 GMT <dgaff> minus: yep, figured that. I think I just found the issue, hold
18:08:31 GMT <minus> :)
18:09:14 GMT <kablamo> hi! i'm still trying to use diskless replication.  i can now get the master to not write to disk but my slaves keep writing to disk at /var/lib/redis/dump.rdb.  any suggestions on how to make them not do that? i've set repl-diskless-sync to yes and i've omitted all "save" statements in the config file
18:14:43 GMT <svm_invictvs> dgaff What hashing algo are you using?
18:19:45 GMT <dgaff> svm_invictvs: Base 36 to base 10 then the one on the memory optimization page on redis' site (that ruby code)
18:19:52 GMT <dgaff> can't recall the URL at the moment
18:20:07 GMT <dgaff> I think I have tracked down the error, and to everyone's shock and surprise... user error
18:20:19 GMT <dgaff> I think I'm fine, at least on the redis front. time to go refill the DB once more
18:20:27 GMT <svm_invictvs> oh?
18:20:36 GMT <dgaff> Yeah, i'm just crossing the wires on my lookups
18:20:58 GMT <dgaff> asking for the wrong thing, in the wrong way, in the wrong DB
18:21:25 GMT <dgaff> Past me is doing future me a real disservice here - anyways, thanks y'all!
18:22:42 GMT <minus> i can see that; you're using ruby
18:22:58 GMT <dgaff> haha
18:58:37 GMT <ankit> Hello, I need some help.  I am trying to setup a redis HA using 3 servers with sentinel/server pair each. The servers are within NAT and also have a public IP.  The firewall however for each servers allows access from only whitelisted IPs (which are the clients). Now if I setup sentinel to point to master using the NAT IP, it works. Howerver on running sentinel get-master-addr-by-name I get the NAT ip at the client, which doesn't work.
18:59:10 GMT <ankit> sentinels cannot reach the master as the servers own IP isn't whitelisted. What should be done in such a situation? Should I whitelist the servers own IP, in its firewall? Is there any other way to tackle this?
20:25:36 GMT <ankit> hi can anyone answer a question related to get-master-addr-by-name when the ip listed in sentinel monitor is a private ip, and we need the public ip of the master?
20:31:12 GMT <ankit> @all
21:04:06 GMT <kerrick> Is there a list of all Redis read-only commands somewhere?
23:18:53 GMT <sagax> hi all!
23:21:42 GMT <sagax> how to calculate how many need memory for redis cluster?