Log file Viewer
#mongodb logs for Monday the 17th of August, 2015
[00:40:53] <coderman2> im running db.col.aggregate([{"$group" : {_id:"$source_group", count:{$sum:1}}}]) on a collection with 5m records. i have an index on source_group, but the query is doing a colscan. is that normal?
[00:46:11] <coderman2> db.col.aggregate([{ $sort: {_id: 1}}, {"$group" : {_id:"$source_group", count:{$sum:1}}}])?
[00:49:30] <coderman2> .aggregate([{ $sort: {source_group: 1}}, {"$group" : {_id:"$source_group", count:{$sum:1}}}]) been running over a minute
[01:35:38] <nofxx> j. chrst , why mongo doesn't pay a github or use gitlab at least? JIRA is a waste of time, to put in kind words..
[01:41:05] <cheeser> it's definitively *not* a waste of time though I can see how some wouldn't like using it.
[01:41:15] <nofxx> joannac, do you want'em in chronological or alphabetical order? ;) kidding but, really just old ideas, doesn't compare. But some examples: status: 'on code review' ... good, where's the code? what changed? there's no repo<>issues integration, not to mention pull request concept.
[01:47:46] <nofxx> and the cherry on the top: the design, is so bad and ill-done I can't even start. And I don't mean the beauty, I mean the usability. Currently I'm trying to read a issue and there's so many panels and nonsense info the description and comments have exactly 20% of the screen
[01:57:50] <nofxx> Annoyance and despair: http://i.imgur.com/zph7EUD.png Clarity and information: http://i.imgur.com/hC2GQlL.png cheeser joannac
[02:04:24] <cheeser> nofxx: you're barking up the wrong tree here. no one here can change the issue track decision
[02:05:58] <joannac> Like cheeser said, no one here has the authority to change the decision to use JIRA
[02:07:24] <joannac> If you contributed, you might get some traction. If you're just a user, I don't think you're going to have much leverage here
[02:22:16] <nofxx> cheeser, joannac , no, not directly, neither claiming to. Just a 'user' rant, outside of the box view. But was naive of me to think there's devs here or that mongo devs don't have any active voice in this?
[02:23:53] <cheeser> do you have any idea how hard it is to migrate from one tool to another? apart from massaging the data from one form to another, there are processes and links aplenty. it's non-trivial and needs a bit more motivation than "just a 'user' rant"
[02:26:15] <nofxx> cheeser, with access to the jira db directly, a script to export all to gitlab or github is pretty trivial. I would use some ORM and just map to their respective API.
[04:23:18] <devians> heya, can I query a find across dbrefs in 2.6.*? ie db.mycollection.find('dbref.field', 'value'); I've tried and get no results so I'm obv missing something.
[04:27:20] <devians> joannac, essentially this http://stackoverflow.com/questions/6195286/how-to-query-mongodb-with-dbref
[04:28:25] <devians> they're querying the id thats contained in the dbref, i want to query a field thats on the referenced document.
[04:30:20] <devians> ok. then how can i achieve this in a performant manner? can I query foos and bars independantly and then get the product of those queries somehow?
[04:31:19] <devians> ok, so the latter query, is that essentially an 'where dbref.$id in' kind of query I guess?
[04:32:20] <joannac> "To resolve DBRefs, your application must perform additional queries to return the referenced documents. Many drivers have helper methods that form the query for the DBRef automatically. The drivers [1] do not automatically resolve DBRefs into documents."
[04:36:43] <devians> I think this is what I need: http://doctrine-mongodb-odm.readthedocs.org/en/latest/reference/priming-references.html
[04:40:19] <devians> it means I cant use count() anymore, which is a bit annoying but I'm limiting by a date range so the numbers should be low enough that the speed penalty isnt too bad.
[05:12:38] <h1fuelcell> I'm stuck using an implementation of Mongo that disallows the aggregation framework
[05:14:45] <h1fuelcell> I'm having trouble understanding what this means: "initial (object) – initial value of the aggregation counter object."
[05:14:52] <h1fuelcell> http://mongodb.github.io/node-mongodb-native/api-generated/collection.html#group
[05:15:14] <h1fuelcell> command (boolean) – specify if you wish to run using the internal group command or using eval, default is true.
[08:30:32] <Mia> Hi channel -- I have a mongodb document structure as {_id:"...",content:{...},color:"red"} - and I want to get a random document from my collection with a specific color --- such as "a random document that's red" -- how can I query this in mongodb?
[08:44:57] <Derick> Mia: I don't think there is a way to fetch a random document. However, if there are not too many documents in the collection, you can do something like this: 1. count( { color: "red" } ); 2. calculate random number in range 0-($count-1) 2. find( { color: "red" } ).skip($random).limit(1);
[08:45:48] <Mia> Derick, .skip is a no-no, I tried it, I have around 200k documents in te collection now and if it's a big number it takes seconds
[09:37:04] <bartzy> I’m currently using MongoDB 2.4 on Debian, via the 10gen repos. The package name is mongodb-10gen. I want to upgrade to 2.6, where the package name is mongodb-org. These packages conflict. How should I upgrade? (This is a production environment)
[13:15:16] <vagelis> I ask something even if i have already did some time ago because i really cant do it with mongo. Image we have 3 simple documents like this: {num: 1}, {num: 2}, {num: 3}. How can i get the documents that have sum of 3 (so the first 2 documents) ?
[14:06:44] <Pinkamena_D> Hello, I have a special query in mind: first a specific query is fetched and sorted, and then a less specific query is fetched and sorted. The more specific entries should be returned first. Is this possible with aggregation or should I just pull and store and conbine the two queries client side?
[16:54:47] <baid0c> Hello, I`m getting MongoCursorTimeoutException, although the server load is very small between 1-2, the server has 8 cpus
[17:11:28] <baid0c> the collection matters, if the slowquery runns on collection A, can the timeout query run on collection B?
[17:45:00] <stuntmachine> I'm setting up Mongo 3.0.5 from scratch on CentOS 6 and curious how to properly configure an arbiter w/ two nodes as replicas. I see a lot of the documentation talking about running commands against mongo to set up the replica set but can't I codify this into the Mongo configuration file before starting the arbiter and replicas?
[17:49:39] <Derick> stuntmachine: nope, configuration for a set is not written down in config files, as that makes adding a new member while it runs not straightforwards
[17:50:24] <stuntmachine> I just don't want it to start and try replicating data to the arbiter. I think I'm thinking of this too much like a traditional RDBMS
[17:51:22] <Derick> it won't start replicating to an arbiter if you don't add it as a normal host, but an arbiter
[17:51:32] <stuntmachine> I see now why it says to specify journal.enabled to false and smallFiles to true
[17:51:49] <Derick> usually on the "primary": rs.initialize(); rs.add("hostnameb"); rs.addArb("arbnodehostname");
[17:53:41] <stuntmachine> oh, you know what i was getting confused about last time i looked at this? i install mongo from the official repo and it creates a config file that isn't YAML, but then the docs say the new method is YAML.
[17:54:50] <stuntmachine> so then what's the proper way of adding those two values in the old format? journal.enabled and smallFiles
[17:59:42] <stuntmachine> "These settings are specific to arbiters. Do not set journal.enabled to false on a data-bearing node. Similarly, do not set smallFiles unless specifically indicated."
[18:48:58] <jr3> with mongo one can shard data by actual values on said data right? If property is A put into A shard, if B put into B shard
[21:23:51] <Trindaz> followed the tutorial in the docs, getting TypeError: 'Collection' object is not callable. If you meant to call the 'register' method on a 'Database' object it is failing because no such method exists.
[22:47:31] <movedx> jmorphio: No. I find both MongoDB and ElasticSearch to be fine for general purpose applications.
[22:48:30] <movedx> jmorphio: In fact, NoSQL databases are somewhat designed/perfect for such applications, because they're structureless (in that they don't force a structure on you or force you to define one before hand), so they can grow with your application.
[22:49:46] <movedx> jmorphio: Understandable. Personally I prefer ElasticSearch as its interface for accessing data is easier to get to grips with.
[22:50:25] <movedx> Possibly more mature than ES too. I've never looked up the dates/facts on that one.
[22:50:46] <jmorphio> i'm still getting used to the way relationships are defined in mongo, but it doesn't seem that bad
[22:51:13] <jmorphio> there seems to be a stigma around querying embedded documents in a collection - like it's a performance problem
[22:53:03] <movedx> For me it's just the query language. I've also never used MongoDB in any detailed capacity - I'm a Systems guy.