Log file Viewer
#mongodb logs for Tuesday the 26th of August, 2014
[00:07:41] <hahuang65> if I get an error message saying my config servers are not synced, how do I fix the bad one? I know which one is out of sync, since the other two look good.
[00:51:21] <joannac> (so probably shut a good config server down, then copy the files to the bad config server, then start both)
[08:22:50] <dejhcb0c> hi, i have a strange problem with mongodb capped collection: i defined a capped collection with size of 170GB but the database seems to expand itself because the size on disk is now around 224GB? somebody have an idea?
[08:33:13] <dejhcb0c> i don't exactly know what the numbers meen but i have 5 indexes (including _id_) and all have a size (after db.collection.stats) around 8.000.000.000
[08:34:28] <dejhcb0c> ah, i understand. so the capped collection size does not include indexes, right?
[09:46:08] <jordana> efazati: https://blog.compose.io/how-we-scale-mongodb/ has some info about "Core sharding"
[10:50:46] <nextfun> I have a .populate going on inside of a .findOne, but I can't seem to figure out how to exclude _id and __v from being sent with the user object inside .populate - any ideas of how I would do that? https://gist.github.com/anonymous/1ecf47fdf1342fb67aa7
[12:53:17] <whaley> DarkSasori: DBRefs have a fetch function on them (at least on the built-in javascript driver)
[12:54:12] <whaley> DarkSasori: safe to assume other drivers have that function/method also (the java driver does at least)
[12:59:16] <kuku> I'm using geonear, and it gives me weird results, showing that an address in a different state is a few miles away. Do I have to reindex for every search ?
[13:44:49] <whaley> DarkSasori: apparently it expects you to explicitly find it yourself using its collection/database/id properties
[13:46:52] <hugh> I'm using Tingodb (http://www.tingodb.com/) and am having troubles with files for the collections not being automatically created. Anyone know if theres some flags/options needed to specify auto-creation of them?
[13:47:33] <hugh> What's worked so far is just 'touch'ing each of the files of the collections I'm indenting to use, but this seems like an unneccesary step
[13:50:19] <W0rmDr1nk> we need something like http://benchmarksgame.alioth.debian.org/ for databases
[14:22:28] <richthegeek> is there an (easy) way to make a replica set only replicate a single database?
[14:24:00] <richthegeek> short story, because I want to try out some things on a single database inside of a system which contains lots of databases and i'd rather not do it on an important secondary and also don't want to copy every db over the wire if possible
[14:25:23] <richthegeek> the long version is my boss wants to look at the data in tableau so i need to use MoSQL to get it into PostgreSQL (rather than a FDW) and, similar to above, dont want to do every database in the world for this one thing
[14:25:56] <richthegeek> although looking at it now, i think i can just make MoSQL look at the primary and it'll sort it for me
[14:26:07] <kuku> I'm using geonear and I get an error: no geo index even though I have a 2dsphere index
[15:01:42] <JonathanN> I'm wondering if anyone else has had this issue...I upgraded my mongo replica set from 2.4.5 to 2.6.4 following the instructions (upgrade the secondaries, primary step down, upgrade final node). However, all of my clients (using Ruby driver) got a ConnectionFailure the next time they used their connection.
[15:21:24] <ome> So because of how indexing works (http://askasya.com/post/largeembeddedarrays ) it is not advised (http://docs.mongodb.org/manual/tutorial/model-referenced-one-to-many-relationships-between-documents/) to use arrays that frequently update.
[15:21:48] <ome> So what is the best way to store analytical data in mongo? or should I consider something like Redis alongside mongodb.
[15:24:26] <halfamind> Can anyone point me to an example of performing a javascript query from Mongoid/Moped?
[15:24:36] <ome> jordana: Page hits, User Activity (clicks), navigation history. Something close to Google Analytics, but a very specialised case.
[15:30:41] <jordana> ome: hmm, depends how much throughput you expect, but provisioned IOPS might not be needed
[15:31:21] <jordana> If you have a collection /collections that take in each of your clicks views etc
[15:36:33] <jordana> I'm pulling out thousands of documents from these raw collections in very quick tim
[15:38:09] <jordana> If you're worried about speed on the front end use redis to mitigate anything you think will be too slow
[15:40:23] <ome> Guess I will just use redis for the analytics and sync to disk at intervals while mongo for actually data and user.
[15:43:36] <jordana> In our case we do use that most recent data but not HUGE amounts until the following day after which it's been consolidated
[15:44:12] <ome> jordana: It is very similar to a blog platform in terms of data flow and access patterns. Lots of Documents that don't change on itself but there is going to be a lot of rapidly changing data around these Documents.
[15:44:58] <ome> think of comments (rapidly changing) around Documents (articles) that don't change that often.
[15:45:29] <jordana> In which case, seperate out the bits that don't change from the more volatile data
[15:46:32] <ome> But the rapidly changing data are very small to be considered a standalone document, mostly just a timestamp with an ip and a document id.
[15:48:30] <ome> No. not incremental data, incremental number of data. visit = [ {ip, timestamp} , {ip, timestamp} , {ip,timestamp}......]
[15:53:43] <akshay_> I could not find anything about excerpt/snippet extraction .. does mondoDB support it
[15:57:53] <Derick> akshay_: mongodb only stores data, it doesn't extract anything (like you probably mean from elastic search)
[15:58:08] <jordana> Because you can have realtime updates at a precision of an hour? Like 100 views in the last hour? or..
[15:58:52] <jordana> ome: http://blog.mongodb.org/post/65517193370/schema-design-for-time-series-data-in-mongodb might help you
[16:00:17] <akshay_> and from what I see .. probably FTS does not have fuzzy matching in current version
[16:00:53] <mjmdavis> Is it possible to sort arrays created with push in an aggregation query? Can I just sort on the parameter I want the arrays sorted in before the group stage with the push?
[16:01:25] <mjmdavis> Also, if I already have an index on that property, does that guarantee that the arrays created with the push will already be sorted?
[16:36:12] <tjb1982> I noticed that pymongo supports wire version 2, but I'm not able to find any documentation on what the differences are between wire version 1 and wire version 2. Also, I noticed that my mongod is serving v1 to pymongo. Is there any way to ask mongod to use a specific protocol version?
[16:51:20] <feathersanddown> how to remove subdocuments ?? only need to exists to remove that subdocument
[18:08:28] <davidski> Hey all, I have a tricky aggregation that's causing me to pull my hair out. My documents have a field which is an array of subdocuments in the structure of {collection: [{"type": "key1", "value": "value1"}, {"type": "key2", "value", value2"}]}, I need to reach into the subdocument, pull the value out of all type == key1, project that into the main document,
[18:08:51] <davidski> I can't quite seem to wrap my head around how that would work though. Any pointers/tips?
[18:34:15] <Talles> hi guys , please can some one help me, http://stackoverflow.com/questions/25512370/best-way-to-make-a-medical-history-db-in-mongodb
[18:41:39] <justind> In a collection with documents structured as {arr: [ObjectId]}, how do I find the top 10 occurring ObjectId's?
[18:55:56] <Talles> hi guys , please can some one help me, http://stackoverflow.com/questions/25512370/best-way-to-make-a-medical-history-db-in-mongodb cheeser
[18:58:37] <Talles> im making a medical history sistem, my doubt is in the structure , do i put all patient evolutions, records, medical prescriptions , exams , and surgeries in a subdocument , or do i create a new document for each one of these categories ?
[19:00:48] <Talles> ex: patient{name:example,medications{},prontuary{},surgeries{} .... } or prontuary{pacient:_id} medications{patient:_id} prontuary{patient:_id}
[19:04:39] <Talles> but the language is not the problem , i need to know the best structure do use in this mongodb project
[19:05:20] <Talles> becouse it will be a big project with a lot of records, i've chosen mongodb , becouse im not confident in the performance of mysql joins
[19:05:57] <awasthinitin> i am also new to mongodb. I am using it to store chat messages in a chat server build on nodejs
[19:08:05] <awasthinitin> the only reason i choose mongo db was that nodejs is asynchronous and relational databases would cause problem with scale
[19:24:46] <sflint> Talles: can you describe the data you are storing and the query pattern you will need to use
[19:27:32] <Talles> sflint: i will store all the patient medial history , like , medication that he will take , procedures , exams, and prontuaries , prescriptions and etc
[20:12:44] <sarath> Can anyone help me> i am unable to launch mongod on ec2 it cant bind to the external elastic ip
[21:56:56] <feathersanddown> I'm trying to save a class into database, is a inheritance from other class, and what I see inside db is only fields from child class, not the base attributes