Log file Viewer
#mongodb logs for Monday the 27th of January, 2014
[00:06:44] <rhqq> guys, i've been looking through documentation but i couldnt find the answer. what user roles are required to view the http interface?
[00:37:34] <george2> I need to run a radius of gyration function (http://paste.chakra-project.org/6314/) over my dataset. Should I be using map/reduce, or something else?
[00:38:25] <akarin> Hey guys, I had a question... I have a unique application and I feel mongodb would be the best choice for a starting point, but I have very specific needs regarding replication. Is it possible to somehow write a plug-in for mongodb where I could implement my own version of replication? If not, I suppose I could always write some sort of proxy server that sits between my application and mongodb and contact each server manually, but this is
[00:39:35] <george2> The builtin functions don't work for your purposes? http://docs.mongodb.org/manual/replication/
[00:42:10] <akarin> george2: I need to implement a special form of replication that basically allows for multiple primaries, and then they'll sieve results together upon rejoin if disconnected
[00:44:42] <akarin> basically you can post updates to multiple servers, and if for some reason they're disconnected, it'll pick the latest update to be the final value for whatever item
[00:45:09] <akarin> most forms of replication only allow for one primary, and basically i'm doing something special
[04:00:32] <daniel-s> A question about database size: Why are databases so big? I mean, I have only been learning to use mongodb and experimenting with it a little. My DBs have only a few documents, but are several hundred MBs in size.
[05:31:58] <george2> I currently have thse functions http://paste.chakra-project.org/6322/ saved to my db.system.js. Can anyone help me figure out how I should run them over my dataset http://pastebin.kde.org/pc163fbcb/08004631/raw ? I want to collect all the posts matching userId "X" and put them through gyration() to get a radius of gyration measurement for each individual user.
[05:33:09] <george2> the "frequency" mentioned in the JS is just the number of data points for the user at location (lat, long)
[05:33:57] <george2> I'm unsure whether I need to do something with map/reduce (which I haven't used before), or if I can just use an aggregation pileline, or if I need to do something entirely different
[05:35:18] <george2> also, if it matters, I will be running this over a few hundred million data points, with probably a hew hundred thousand unique userIds
[07:26:12] <_rgn> is polling mongo for new objects a bad idea? to use it as a queue of sorts for background jobs
[08:31:23] <imslavko> _rgn Meteor is an example of a system that looks for mongo updates. In particular, Meteor tails oplog
[09:43:59] <soko> Trying mongoose.connection.collections['collectionName'].drop( function(err) { }); returns TypeError: Cannot call method 'drop' of undefined. I assume I need to define a collection first? What's the recommended way to do so? (I'm following the code from a book and I have either missed something or it's an error in the book.)
[10:18:19] <kali> first line is a "slow request" warning line (which is truncated because the request is pathologically huge)
[10:21:00] <kali> can't offer specific help here, but you may want to test case the query with another client
[10:32:50] <bzikarsky> How does mongos balance data I insert into a sharded cluster when no sharding is configured for the db and collection? Does it all go to shard1 until shard1 is full?
[10:40:24] <bzikarsky> Hm. We have a scenario with many smaller databases. It's a steadily growing numer, each is hovering around 1-10GB. It would be nice to balance the databases to multiple shards, so writes hit different shards (replsets-primaries)
[10:41:25] <Derick> bzikarsky: you can set a primary shard for each database (I think, only before you create it though)
[10:41:52] <eutheria> hi, i am just starting to look at mongo, i was wondering does it support schema constraints?
[10:43:01] <bzikarsky> Currently we load-balance manually to different clusters. I'm afraid that introducing sharding with manual configuration would only introduce an unneccesary layer of complexity
[10:43:02] <Derick> eutheria: no, it does not. That's something that is now pushed in your appliction
[11:06:31] <soko> Trying mongoose.connection.collections['collectionName'].drop( function(err) { }); returns TypeError: Cannot call method 'drop' of undefined. I assume I need to define a collection first? What's the recommended way to do so? (I'm following the code from a book and I have either missed something or it's an error in the book.)
[11:07:11] <soko> I've been stuck on this for a while and I can't see anything that I haven't followed. Any help is appreciated :-)
[11:19:01] <eutheria> i am trying to figure out if i should use mongodb, i am currently using orientdb, but struggling to integrate that with my java webapp
[12:48:51] <idank> how do I tell a sharded cluster to rebalance after mongorestore on one of the mongod's?
[13:24:13] <kali> if you shake the whole setup (stop everything then restart, for starters) the sharding may pick it
[14:38:08] <pwaleczek> hey guys, anyone can tell me gow to limit mongod do only one instance? It's spawning multiple subprocesses and I'd like to stop that.
[14:58:19] <byalaga> Hi, one of our replicatset member runs the configdb server, is there way to check whether its a mongodb primary or not?
[15:53:59] <byalaga> joannac: I have connected to mongo configserver shell, I want to know if there is way/command to check if the sever is also running mongd.. rs.status() doesn't work as i have already connected to configserver
[16:15:55] <saml> replSet=set0 master=true I have these in config file. but i get "not master" when i insert to the mongod
[16:40:35] <saml> https://gist.github.com/saml/13c65c697195ba514784 i connect to both :27017 and :27018 using mongo shell. and non of them is master
[16:48:39] <goreorto> I'm new to mongo and trying to understand how mapReduce works. I have a collection with about 3K records. When I perform a dump mapReduce, the records do not add up. db.events.mapReduce( function(){ emit('x', 1); }, function(k, vs){ return { x: vs.length }; }, { out: { inline: 1 } }). x ends up with a value of 6. Any hints are truly appreciated.
[17:54:32] <shadfc> hey, i've got a test suite for a python app that drops and recreates a db from fixtures for each test. That appears to be really slow. Any pointers on how to make this faster? I'm currently doing it all from pymongo via drop_database and then bulk insert calls
[19:22:24] <shadfc> so this is pretty cool. When running my test suite against a local mongod, I have sporadic errors when querying using secondary preferred read preference. After removing the read pref, all tests pass. I am not using a replica set locally -- { "ok" : 0, "errmsg" : "not running with --replSet" }
[19:31:21] <ranman> shadfc: you should use smallfiles and no prealloc as command line params or config options for mongod to minimize the amount of time creating the new database creates
[19:32:59] <ranman> shadfc: in pymongo you should refrain from using a MongoClient to connect a replicaset, use a MongoReplicaSetClient for that. If you're not running with replication the secondaryPreffered readpref won't really work for... well I mean you don't have any secondaries so :P
[19:33:26] <steve1> I just had a brief site outage that stemmed from the following events: 1) 5-node replicaSet, 2) took 2 nodes down (one was master). A new master was elected from remaining members, but site stayed offline 3) A new node was added to replicaSet, 4) The replicaSet continued to be un-query-able until the down'd nodes were taken out of replicaSet. Can anyone help explain what happened to cause the outtage?
[19:35:08] <ranman> steve1: what were the responses from the mongod that lead you to believe it was unqueryable?
[19:35:22] <ranman> steve1: also please feel free to open a community support ticket in JIRA https://jira.mongodb.org/browse/SUPPORT
[19:36:43] <ranman> steve1: I need more info that that sorry, specifically what was it hanging on? what URI were you passing in? were you passing in a good seedlist?
[19:39:10] <steve1> ranman: connection string was for all good IPs: mongodb://172.16.0.20,172.16.0.21,172.16.0.22/?connectTimeoutMS=10000
[19:39:54] <ranman> and you don't have any more detailed info than the connection from the driver was hanging on that URI?
[19:42:07] <steve1> ranman: right. I was hoping it was just our process failure that you or someone could say "yeah, duh … look at this page". I'm creating a MongoDB/JIRA account now as you suggested, and will go through the logs to see if perhaps an error message was overlooked.
[19:45:17] <joannac> I would be curious if it's a problem with the PHP driver, or whether we didn't actually get a new primary elected (i.e. a server bug)
[19:48:26] <starfly> steve1: if the future, you might consider manually stepping down the primary when you know you need to take it offline: http://docs.mongodb.org/manual/tutorial/force-member-to-be-primary/
[19:51:08] <steve1> starfly: I suggested that, but I think the sysadmin prefers to be a chaos-monkey and test the fault-tolerance.
[19:51:59] <starfly> steve1: sure, I'm all for testing fault tolerance, just not on a system or replica set where uptime matters
[19:52:44] <cheeser> so you're always walking by my desk ... and not working. that's what I'm hearing.
[19:53:42] <steve1> starfly: preaching to the choir. :) but by the same token, if fault-tolerance is not tolerant in production, then it's not really tolerant.
[19:56:53] <steve1> ranman may be right; it could be a mis-configuration on the app end. we've traditionally had a lot of pain with the fault-tolerance of MongoDB replicaSets. At the very least, the downtime may be enough motivation for the business to allow us spending time on finding those weak links.
[20:01:37] <starfly> steve1: yeah, you don't know what you've got until it's gone… Most innovation stems from catastrophes
[21:06:18] <NyB> Is it just me or is the specification at bsonspec.org somewhat incomplete at points?
[21:10:25] <NyB> cheeser: for example I cannot find the specification for both UUID versions, ObjectId, MD5, DBPointer, Timestamp
[21:11:52] <NyB> I have been reading the MongoDB BSON decoder code, but I sure would like an actual document with those details
[21:12:49] <NyB> For example, what is the difference between the old and the new UUID binary subtypes?
[21:14:02] <NyB> ...the version of the Java driver that I am using (latest 2.x.x) does not seem to even support the new UUID serialization...
[21:23:08] <NyB> cheeser: interesting... I am writing a new BSONDecoder implementation, so this may come handy
[21:25:19] <cheeser> good luck. i was just saying i'm not sure I could write one with what's up there. it *does* seem a bit incompletely.
[21:27:29] <NyB> cheeser: well, to be honest for my application getting some cases wrong might not even be noticed. Why would anyone want to store regexes in a BSON document ? :-p
[22:19:09] <saml> cheeser, did you send to solr each mongodb document? or did you collect updated/inserted/deleted documents and sent to solr in batch ?
[23:15:46] <TheDracle> couldn't add user: Only admin.system.users entries may contain 'otherDBRoles' fields
[23:19:39] <DevRosemberg> http://pastie.org/8673538 in there, Achievements, Permissions and Purchased Items dont work
[23:21:12] <joannac> TheDracle: http://docs.mongodb.org/manual/tutorial/add-user-to-database/, last example "Only the admin database supports the otherDBRoles field."
[23:23:21] <DevRosemberg> like a player joins back and its like if they had never joined, permissions, comets and stuff work
[23:52:39] <paulkon> as I understand it, the $ projection requires a field of the array in question to be present in the query document itself. Is it possible to project an array field if it is present without it being a part of the query itself?
[23:52:41] <paulkon> I want to return the array element if one of its subdocuments match a subdocument field but still return the document itself if it doesn't exist