Log file Viewer
#mongodb logs for Monday the 9th of December, 2013
[05:58:48] <mark___> joannac: means if i have thousands of collections so that i can do shard to maintain the balance
[08:36:55] <kephu> I got a bit of a noobish question: I want to write a backend-to-a-game sort of a system, with badges and achievements and such. Being new to mongodb, I wonder, if it's better to store badges and achievements as sub-documents to the user document, or their own collection?
[08:37:42] <kephu> if it helps, I'm expecting potentially a lot of users, and, at least initially, just a couple of achievements
[08:39:22] <kephu> kali: like I said, noobish question ;) How would I approach that from the query standpoint?
[08:39:24] <kali> kephu: try to figure out what the most frequent queries will be, and what layout will be optimum
[08:40:49] <kephu> kali: well when you put it like this, the "embedded doc" seems like the way to go. But then again, there's gonna be a dashboard component to it
[08:42:27] <kephu> well, like an administrative bit that lets you see how many users have achievement x, which ones are consistently topping the leaderboards, stuff like that
[08:44:42] <kali> well, if it is on the administrative size, the scalability is less critical. i assume most stuff will be achievable with the aggregation pipeline or map/reduce, but that may be too slow for your taste
[08:44:48] <kephu> another noobish question, then: does mongo allow me to group by username, and then by achievement type? Provided the layout's like {username: 'some_user', achievements: [{name: some_achievement, ...}, ...]} ?
[10:27:50] <Quest> how relations are made in mongo db? like i have 2 tables, person and person_address . how do I get the addres of the person?
[10:28:34] <Derick> or, redesign your schema so that the address is embedded in the person documents (prefered)
[10:30:42] <Derick> because in MongoDB you store your data from an applications point of view, and it makes no sense to do two queries if you don't need to.
[11:49:08] <richthegeek> I've got some badly-saved data in my db (numbers stored as strings) and I'm wanting to do an aggregate $sum on these fields - is there any way to make it work (ie, make the driver assume that $sum fields should be at least considered as numbers) or is there just no way?
[11:51:13] <Nodex> richthegeek : I don't think there is a way, on the plus side you can just loop it and fix the data
[11:59:53] <thelucid> Hey, getting "failed with error 16550: "not authorized for query on xxx"", anyone got any ideas, desperate!
[12:00:33] <joannac> thelucid: authenticate with a user with the right permissions for what you want to do?
[12:01:31] <thelucid> @joannac: I have... using Mongoid. I can auth manually and select the relevant entries.
[12:31:16] <bla> Previously on few fields I did .distinct() and then using a loop just reinserted into new DB
[12:32:00] <bla> But now distinct fails on 16mb limit and I've got no idea how to do it - except for writing external software.
[12:33:08] <ranman> bla: internally distinct creates a document and if there are more than 16mb worth of distinct values it's going to hit the document size limit
[12:34:08] <bla> Can I create a collection with unique index on value and use map reduce to fill it?
[13:25:19] <Derick> bin: IRC is asyncronous - ask your question, and somebody will answer when/if they can
[13:25:35] <sflint> aasd123....mongodb is single threaded....and mongo is fit for production and used at many companies in production.
[13:28:57] <bin> so can i provide rs conf file automatically without connecting to mongo and invoking rs.initiate()
[13:31:52] <kali> also, run a top showing thread on a a system where mongodb runs... that will not leave much place for doubt
[13:35:36] <Derick> bin: it has to be in config files to make adding new nodes easier... and not have stale config files all over the place
[13:36:51] <Derick> bin: it has to be in the replicaset config on the servers to make adding new nodes easier... and not have stale config files all over the place
[15:18:41] <Joeskyyy> Well, you don't necessarily have to, you can run your arbiter on the same machine as another
[16:17:29] <movedx_> Can you place a constraint on a document's fields? For example, can I place a constraint on a field that prevents an integer less than 0 or greater than 100 from being applied, or do I have to do this in software?
[16:36:30] <prateekp> hello, i am building a rails appliction . I am using mongodb .... While i am able to access the database from console but the things are not running on the application
[16:39:39] <prateekp> kali : you can see the error at https://gist.github.com/papriwalprateek/7875432
[16:44:46] <prateekp> but actually i am able to get access the database through terminal but results are not coming on the web app
[18:42:12] <ericsaboia> hey guys, I need to sort users based on a list of ObjectIds, I'm trying to use aggregation with project to accomplish that
[18:57:18] <link1511> and i don't see anything that looks like an issue in the numbers on the mongo side of htings
[19:06:48] <lakmuz> have collection {_id: ingridient, productIds: idsofproducts}, need to update this row by add new product id to productIds if exists or create new
[19:27:35] <sharondio> link1511: Slaves? You mean replica sets? Read-only secondary replica sets is the default.
[19:31:47] <link1511> and it appears that the writing in blocks / second are equally distributed among my hosts irregardless of master / slave
[19:32:07] <link1511> i have a bunch of statistics on theses hosts but absolutely no information on what might be "bad"
[19:34:52] <sharondio> Well, you can login and take a look at the rs.status() to confirm that things are setup the way you want them.
[19:35:33] <maplebed> Hi, I'm looking for the changelog for the mongo gem. Docs point to https://github.com/mongodb/mongo-ruby-driver/wiki/History but that page doesn't exist. Any suggestions?
[19:41:16] <link1511> like what can i tell from an error message there saying it was trying to replicate
[19:43:24] <link1511> joannac: totally up to help get them to you, what would you need? I'm setting up a pastie of the rs.status() on primary
[19:43:57] <maplebed> joannac: that gives me the revision history for the wiki page, not the gem, doesn't it?
[19:44:15] <maplebed> I'm trying to find out what changed between v1.8.2 and v1.9.2 of the mong ogem.
[19:46:03] <link1511> joannac: I can't find any errors in the logs, maybe i'm looking in the wrong place?
[19:47:50] <link1511> hmm, but i'm seeing probles on the frontend in the form of autoreconnect errors
[19:50:07] <maplebed> joannac: Perfect! Thanks. Any chance you know who could update http://api.mongodb.org/ruby/current/ to point there?
[19:51:18] <joannac> maplebed: i think maybe open a jira ticket under the docs project? or maybe the ruby project
[19:52:41] <joannac> link1511: check the mongod logs to see if there's any "failed connection" or "A thinks B is DOWN" or something like that
[20:16:37] <ericsaboia> hey guys, I need to sort users based on a list of ObjectIds, I'm trying to use aggregation with project to accomplish that
[20:22:27] <dnsdds_> I have an array of documents I want to insert to a collection with a unique index. Some documents are identical to existing documents, some aren't. What's the path of least resistance to insert only the ones that aren't already there?
[20:27:34] <luimeme> i try to do an upsert it work well except it remove a field that is on the document but not in the value i upsert how can i update only the field of the upsert and leave the other intact please
[20:29:00] <Joeskyyy> http://docs.mongodb.org/manual/reference/method/db.collection.update/#upsert-parameter
[20:31:25] <dnsdds_> Essentially, I have incoming documents with a single field called "tags" that have CSV data in them ("foo, bar, derp"). I split it with ", " to make an array of tags, and I want to add each tag as a separate document in my tags collection
[20:32:01] <dnsdds_> I can do a bulk insert, but what if "foo" already exists? Doesn't that mean that "bar" and "derp" won't get inserted because the unique flag failed for the first one?
[20:38:36] <Joeskyyy> dnsdds_: Can you expound a bit more? Are you just adding "foo" "bar" and "derp" to a general tags collection, and checking to see whether any of them exist?
[20:39:28] <Joeskyyy> i.e. in your tags you currently have "foo" "bar" "cats" , so if you sent in "foo" "bar" and "derp", it checks to see which of those exists, if it doesn't exist, insert it? (in this case "derp")
[20:40:45] <dnsdds_> Joeskyyy: But I'm guessing there should be a smarter way to do that than to just loop over the array and separately .find() each tag and .insert() if it doesn't exist
[20:41:38] <dnsdds_> Joeskyyy: in the actual items collection, it's just a CSV-style String for presentational purposes. In "tags", it's separate document for each tag
[20:44:11] <dnsdds_> I just find the bulk insert so handy, there must be a way to ignore any unique errors and just press on
[20:52:10] <dnsdds_> "Drivers have a ContinueOnError option in their insert operation, so that the bulk operation will continue to insert remaining documents in a batch even if an insert fails."
[20:56:21] <Joeskyyy> kind of a janky workaround but 2.4 has a setOnInsert when running an update, I'm not sure how fast this would get you but it does work
[20:56:22] <Joeskyyy> http://docs.mongodb.org/manual/reference/operator/update/setOnInsert/#op._S_setOnInsert
[20:56:48] <Joeskyyy> > db.example.update({tag:"derp"}, {$setOnInsert: {tag:"derp"}},{upsert: true})
[21:08:27] <dnsdds_> Joeskyyy: [MongoError: E11000 duplicate key error index: plusminus.tags.$name_1 dup key: { : null }]
[21:44:33] <qswz> If you directly map user/clients json requests (like that: [action, token, collection, json1, json2]) into mongo queries, what are the potential danger? given that 1- collections names are prefixed with "test.", 2- only 3 mongo methods accepted: find, remove, update (upsert mode), and 3- there is an access, update and remove-rights for documents
[22:39:20] <ericsaboia> hey guys, I need to sort users based on a list of ObjectIds, I'm trying to use aggregation with project to accomplish that