Log file Viewer
#mongodb logs for Saturday the 5th of December, 2015
[00:16:17] <Boomtime> @Doyle: https://docs.mongodb.org/manual/release-notes/3.0-upgrade/#change-replica-set-storage-engine-to-wiredtiger
[00:16:31] <Boomtime> specifically the note: "Before enabling the new WiredTiger storage engine, ensure that all replica set/sharded cluster members are running at least MongoDB version 2.6.8, and preferably version 3.0.0 or newer."
[01:25:14] <shortdudey123> anyone used mplotqueries? trying to group by the index used for the query, and my regex isn't working :/
[01:34:56] <leptone> how can i db.collection.find({a: 3, b: 4}, function(err, doc){blah}) such that doc is item in collection with a = 3 and b = 4?
[01:57:14] <leptone> why isn't product.description being populated with data from the .findOne() like the rest of the product object? https://gist.github.com/leptone/3751359934e104c01737#file-productroutes-js-L18
[10:37:23] <lost_and_unfound> Greetings, I am have some difficulty with Indexing a single key with using a range selection.
[11:15:51] <lost_and_unfound> I have attempted to add indexes based on various articles and help pages. So, I added the new index, and the queries are still extremely slow. Can anyone shed some light whether this is related to indexes / query or if this might be hardware related that the server has insuffient resources? http://pastie.org/private/kwcg1diozuq4vdkuhbrq
[11:52:17] <lost_and_unfound> doing some more reading, https://docs.mongodb.org/manual/tutorial/ensure-indexes-fit-ram/ - Is this free RAM required part of the default mongod running, or is that an additional X-GB ram it requires once loading the index to do the query?
[13:10:56] <lost_and_unfound> eh... ok, so we just changed the hardware. From 4GB to 20GB RAM, from a dual core CPU to quadcore CPU, running same query, downfrom 171542ms to 108406ms.
[13:12:31] <lost_and_unfound> This is still way to slow, its a DEV server, so It does not even have proper data.
[13:14:29] <lost_and_unfound> I have the query before and after indexes where created with minimal change to time
[13:16:04] <lost_and_unfound> StephenLynx: that is what I did in the pastes provided, the first query the explain showed the index was not used, then I created the index, ran query with explain() again and it shoes the index was used.
[13:18:09] <StephenLynx> I don`t think that handling nearly 3 millions documents is trivial. and you didn't ran explain, it might not be using the index.
[13:18:49] <lost_and_unfound> My bad on the first paste, the <pre> had the eplain way at the end, I reformated the text now
[13:26:55] <lost_and_unfound> in the link, http://pastie.org/private/khxbmun8461mjguq12shiw you will see the "before" heading where I ran a find().explain() and it showed "BtreeCursor template_id_1". I then created the new index and rean the find().explain() again and to shows I am using the index "BtreeCursor template_id_1_log_date_1"
[13:27:46] <lost_and_unfound> So the index was created and used. I have also added more hardware to the server and restarted it.
[13:40:51] <kexmex> mongodump is horribly slow latest--- is that a known problem? is there a way to diagnose it?
[13:42:57] <lost_and_unfound> I have done a vanilla install on FreeBSD and have not configed and settings
[13:54:54] <daedeloth> I'm fairly new to document based databases, and I was wondering if I could run my use case here
[13:56:43] <daedeloth> Alright, so. I have following situation. I have a list of games. Each game has a list of of groups and a list of users. All users are assigned to one or more groups. Games itself are atomic, so a user from game A can never be in a group of game B.
[13:57:19] <daedeloth> BUT. Groups have a property called "unique token", which should be unique across the whole thing
[14:00:56] <StephenLynx> a game would be more or less like this {gameName:'x',groups:[{groupId:1},{groupId:2},]}
[14:02:22] <daedeloth> well it's used to login, you login into a group and that group is assigned to a game
[14:03:56] <daedeloth> it's not supposed to be very secure :) I just need to generate an access token that is used to login as a certain group
[14:07:32] <daedeloth> yea it's a bit confusing, I'm building a quiz game where users can connect their cellphones to have local multiplayer
[14:10:07] <daedeloth> well problem is that one of the ids represent the "game master", and we'd like to avoid users taking over the quiz master controller
[14:11:15] <daedeloth> ah but user doesn't represent an authenticated user; everyone who connects is automatically a user. Other than that token, there is no authentication done whatsoever
[14:13:52] <daedeloth> now, any thoughts on how to keep these objects in sync across multiple nodes? :p
[14:14:17] <StephenLynx> if you really wish to keep your game free of an account system, 3 collections is the best compromise.
[14:23:03] <StephenLynx> 1: to have servers to not have to keep track of that in the first place, which would increase readings on the database
[14:26:59] <daedeloth> I'd prefer to go for the second option, sounds more bullet proof. Also, I'm talking to a client who wants to use this super big scale (10k players), so I'd like to be able to spread the traffic across multiple servers
[14:27:41] <daedeloth> socket.io connections are scaled with redis; I think I could inject some commands in there to tell all servers to "update the documents with provided ids"
[15:47:42] <kexmex> 2015-12-05T15:42:05.574+0000 I QUERY [conn14] getmore api2.requests cursorid:27698932230 ntoreturn:0 keyUpdates:0 writeConflicts:0 numYields:114 nreturned:959 reslen:4206760 locks:{ Global: { acquireCount: { r: 230 } }, Database: { acquireCount: { r: 115 } }, Collection: { acquireCount: { r: 115 } } } 16045ms