Log file Viewer

[00:00:25] <rafaelhbarros> the solution to that problem was to make python read frm the nearest in the replicaset

[00:00:31] <rafaelhbarros> not just the primary

[00:00:39] <rafaelhbarros> and god that was scary

[07:38:46] <mark____> db.system.users.update({"readWrite":true}); is it ggoing wrong ???

[07:42:34] <mark____> #node.js

[07:44:53] <mark____> http://pastebin.com/gVmuBJaE

[08:15:31] <mark____> http://pastebin.com/swEt57vw when i start node app.js then go to browser localhost and then the content which is to be come dynamicaly come fromm mongodb doesnt comme.then i check the console of mongod there i got it ends the conenection suddenly

[09:36:09] <dsirijus> i'm a project manager on a project, and have no real experience with dbs, and the dev that's supposed to do that has only experience with mysql...

[09:36:32] <dsirijus> and i have to pick a database so was wondering would a mongodb be a good bet on the following use case...

[09:36:57] <dsirijus> pvp turn-based multiplayer game, hopefully hitting *many* ccu...

[09:37:46] <dsirijus> as far as writes go, each game match results need to be stored, as well a 10-20 small vars that help analytics

[09:38:39] <dsirijus> and as far as reads go, player matching algorithm needs to access a few vars but across the whole body of players, and at the end of the gaming session few types of global leaderboards need to be shown

[09:38:54] <dsirijus> that's kind of it, as far as i know

[09:39:08] <dsirijus> data is not sensitive, so if it drops to store a score or two, it's ok

[09:42:03] <kali> well... if you have only one developper and he has experience with mysql, why not using mysql ? that's not a rethoric question: have you identified a problem with mysql in your use case ?

[09:42:37] <dsirijus> kali: i agree. but we're on a bit of a deadline so we'd like to have a good pick right off the bat. and he's not even that good with mysql (phpmyadmin dude :))

[09:43:41] <dsirijus> kali: in essence, i don't need someone to tell me "use this or that database engine" but "hey, your use case it bottlenecked at there and there"

[09:44:34] <dsirijus> i suppose doing a query for leaderboards or for player matching algorithm is that here

[09:44:42] <dsirijus> since it has to go through all the players, each time

[09:46:44] <kali> it sounds like you have a very striclty structured dataset. i don't think document models will bring much to the case, so i would stick with relational

[09:48:24] <kali> and honestly, if your guy is "not that good with mysql", i'm not sure it's a good idea to uproot whatever he knows and put him in a situation where he has to re-learn everything

[09:50:44] <dsirijus> oh, we're both reasonable people and have no issues with "putting someone in situation". we just want to make this work the best we can is all

[09:51:44] <dsirijus> kali: thanks for mentioning phrase "document model", looks like that's really the underlying principle of nosql i was ignoring

[09:51:55] <kali> no.

[09:52:02] <kali> it's one variation of nosql

[09:52:11] <kali> there are other families

[09:52:27] <dsirijus> but i assume most nosql families are document-oriented databases?

[09:52:28] <kali> like column oriented, or graph oriented which both could be more relevant to your use case

[09:52:32] <kali> not really

[09:52:56] <kali> dsirijus: http://en.wikipedia.org/wiki/NoSQL

[09:55:33] <dsirijus> so, my primary decision factor should be picking the database model?

[09:56:24] <dsirijus> since we're kind of "winging it", relational might be a problem, since we might change stuff all the time

[11:31:57] <mark____> any help???

[11:35:55] <mark____> anybody here for help

[11:51:26] <dsirijus> mark____: https://www.youtube.com/watch?v=yWP6Qki8mWc

[11:53:30] <mark____> dsirijus: he he he

[13:39:05] <Todilo> hello people. I feel like an idiot but I just cant find a way to query my documents to get the "newst" of each type. pseudo code: find 1 post per post.section where lastEdited is newest. I do need an aggregate for that right?

[13:40:39] <Todilo> example document post1 = {section="top", lastEdited="2013-01-01"}, post2 = {section="top" lastEdited="2013-03-04"} post3={section="bottom", lastEdited="2013-05-05} and I want post2 and post3 but not post1(as it is older and same section as post2

[15:15:23] <Zeeraw> Now I've even tried to configure my server just as the docs tell me. It still takes forever ti flush the mmaps.

[15:15:30] <Zeeraw> I'm completely out of ideas

[15:45:23] <mark____> what is moongoose?why we use that with mongodb? can u please explain me littlebit

[15:45:33] <mark____> #node.js

[18:28:52] <rafaelhbarros> for pymongo users out there

[18:29:11] <rafaelhbarros> keep track of the max_pool_size

[18:29:13] <rafaelhbarros> its default to 10

[22:52:37] <mdeboard> I'm creating a MongoReplicaSetClient instance per request (via pymongo). Each request will create a new connection until `max_pool_size` connections is reached, right?

[23:23:55] <jcardinal> I have a mongodb instance that seems to be having issues under heavy load. I noticed the filesystem the db is on is not mounted with noatime/nodiratime. Everything I've read says to mount with those options, but can anyone help me understand how significant the impact of that will be on performance?

[23:27:40] <jcardinal> Related…can I safely shut down mongos using "service mongodb stop", remount the volume with the correct options, then "service mongodb start"? I assume the stop command should shut it down cleanly without risk of data corruption, right?

[23:28:12] <jcardinal> shut down mongod, that is

[23:42:55] <cheeser> jcardinal: yes

[23:44:00] <jcardinal> Thanks cheeser. I figured as much, but I'm just getting started with mongodb and don't want to make any stupid mistakes and damage the db.