Log file Viewer
#mongodb logs for Wednesday the 8th of January, 2014
[02:55:06] <mrpoundsign> oh idk, haven't used one in a while. :) there's more here: http://docs.mongodb.org/ecosystem/tools/administration-interfaces/
[08:11:34] <Varun_Krishna> I would like to move the data of the database in myserver to mongodb. Is it possible to do this in mongodb?
[09:33:03] <master_op> hello, i'm fowllwing this tuto, http://dba.stackexchange.com/questions/22813/mongodb-returning-the-points-nearest-to-a-location-with-distance
[09:38:27] <tiller> I've some mongos/c/d servers running on 4 machines, and I try to set proper iptables
[09:38:40] <tiller> But when I set, on 10.0.0.250 for example, those ip tables: http://pastebin.com/7rAZD9xR
[12:43:20] <djlee> Does a compound unique index apply in both directions. So if i have a unique index on "FieldA" and "FieldB", and i enter 1 into FieldA and 2 into FieldB, can i make it so the unique constraint will also stop 2 being inserted to FieldA and 1 being inserted in FieldB
[13:25:56] <djlee> Nodex: right ok, i presume it was a one way uniqueness so "1,2" != "2,1". Clearly not then, which saves me writing a load of management code :D. Thanks for your reply
[13:57:21] <KamZou> Derick, i'm not using balancer (cause i've got some contraints) and i'd like to know if there's a way to configure mongodb to write not only on one shard but on many ? (at least 2)
[14:05:30] <Derick> mongos will make sure half your data gets written to each shard (with a properly chosen shardkey)
[14:06:20] <KamZou> Derick, some of my shards are currently full of space but i want to keep this retention
[14:06:21] <Derick> KamZou: then you need to do the sharding by hand, and your app needs to know to which of the shards to write
[14:08:09] <KamZou> Derick, my apps doesnt "have to know" which shard to write... The backend is the mongoS who dispatch to the correct shard (i set manually)
[14:08:41] <KamZou> Derick, but i'd like to know if i could have 2 shards at the same time in the way i do with one today
[14:11:04] <KamZou> no, i'd like a document A in the shard1, and a document B on shard 2 (if it's possible)
[14:12:58] <Derick> KamZou: but you said you are already doing that, by having configured mongos yourself
[14:15:15] <KamZou> i disabled balancer on my mongoS. Right know i've many shards. Each time a shard is "full" i manually edit the shard where "all" data goes. So data go to one shard, i'd like them to go to 2 or more
[14:23:11] <KamZou> I'm just discovering this parameter : http://docs.mongodb.org/manual/tutorial/configure-sharded-cluster-balancer/#change-the-maximum-storage-size-for-a-given-shard
[14:24:23] <KamZou> i've got to know if the dedicated-server's volumetry is the only reason why we didn't use the balancer right now
[14:24:54] <KamZou> If it's i'll activate it with this kind of configuration (max storage size / shard)
[15:32:44] <tiller> Question: which instance [mongod|mongoc|mongos] communicate with which instance?
[15:33:23] <Joeskyyy> Depends a bit, if you have a mongos I'd probably start there with firewall rules
[15:33:55] <Derick> tiller: that's the only exception. But all mongos's talk to all mongoc's and mongod's
[15:35:31] <Derick> actually, mongod's will only talk to mongod's in the same shard (and to mongos and the 3 mongoc's)
[16:30:55] <nollidj> i'm interested in mapping homogeneous documents in a collection to a stream of homogeneous documents, which will be inserted into a different collection. is mongodb's map-reduce function the right mechanism for doing this?
[16:45:00] <TehShrike|Work1> What happens if ensureIndex is called for a key that already has an index, but with different options than were used on the first index?
[16:52:05] <TehShrike|Work1> The docs say about ensureIndex "Creates an index on the specified field if the index does not already exist."
[16:53:10] <TehShrike|Work> Derick: not really, I don't know how to tell the difference between an index being rebuilt versus being silently ignored
[17:02:08] <TehShrike|Work> Looks like it silently ignores it, even if the options are different. Good to know
[18:16:02] <Veejay> Hi, if I have a MongoDB replica set running locally, how can I manually kill the primary or make it so that it becomes unavailable (to see if the re-election procedure happens as expected)?
[18:17:34] <Veejay> I was looking for a more mongo-centric procedure but yeah I guess that'd work fine
[18:18:47] <Veejay> Is db.printSlaveReplicationInfo() the way to get the info about which one is the primary?
[18:19:03] <Veejay> And will I have the info about the process PID that way (to know which one to kill)
[19:02:40] <nollidj> is there a good way when running the mapReduce function to have the resulting records merged with an existing collection rather than overwriting it?
[19:08:27] <nineteen1ightd> I have a document with a field that contains an array of subdocuments. Those subdocuments contain a field that itself has an array of subdocuments, e.g. {foo: [{bar: [{baz: True}]}]}. I can find documents in the collection easily enough based on a value in the inner most subdocument, however, I really want to apply updates to a particular set of those inner most subdocuments...
[19:10:44] <nineteen1ightd> I tried something like db.coll.update({'foo.bar.baz': True}, {$set, {'foo.bar.$.qux': true}})
[19:11:17] <loopj> does mongo have any way to sort a collection by the length of an array in each document yet?
[19:11:42] <loopj> we dont have a seperate count field, and I dont want to make one for this task (building an admin page)
[20:04:51] <wc-> hi all, im having some issues trying to do a disctinct query with 2 fields using the aggregation framework
[21:27:41] <tripflex> hey guys so wondering what the "best" and most efficient way to do this would be
[21:28:32] <tripflex> I have a collection that has a value "status", status is either equal to complete or incomplete, in building my web app to output into tables i'm currently using two find calls but my assumption is this isn't the most efficient way to do it
[21:30:23] <jmagder> Good afternoon (EST). Been playing with mongodb for a few weeks now and it is amazing. Been learning about improving search queries with hints, but I can't seem to get a hint to apply
[21:30:32] <tripflex> app.db.models.TPV.find({ OwnedBy: theowner, TpvStatus: 'incomplete' }, function (error, incompletedata){ app.db.models.TPV.find({ OwnedBy: theowner, TpvStatus: 'complete'}, function (error, completedata){ callback (incompletedata, completedata)
[21:31:15] <tripflex> so I can pass the returned data in two objects, completedata and incomplete data, and then use swig template to output the tables
[22:55:44] <djlee> If i have two mongo _id's. Is there any such thing as a "lower" id? Like with a integer id in mysql, 10 is smaller than 15. Is there any such thing in mongo?
[23:03:47] <Joeskyyy> djilee: Not really… the objectID is compromised of a few factors, such as unix epoch, procID, and a random value counter
[23:08:12] <djlee> thanks Joeskyyy, suppose i'm looking for something that mimics mysql's autoincrement field, and theres a good reason mongo doesn't have such a thing. I'll have to start thinking about solving my problem in a different manner
[23:12:00] <nineteen1ightd> Is there an update I can do to $set a new field on all embedded documents in an array?
[23:12:33] <djlee> i don't think there is, but is there a way of implemnting a bidirectional unique constraint in mongo? So if FieldA is 1 and FieldB is 2, i cannot then enter FieldA:2, FieldB:1. So basically the uniqueness would be on both 1,2 and 2,1 ?
[23:14:03] <nineteen1ightd> I'm not terribly familar with Mongo, however, I'm guessing a unique index on (FieldA, FieldB) is unique on that ordering of their values?
[23:15:27] <djlee> nineteen1ightd: yep, uniqueness is only on the order you specify from what i can gather from my tests. Which makes sense since first and foremost it's an index
[23:16:12] <djlee> just going through all the possible good solutions i can think to my problem, before i resort to coding something a little less nice and performant to handle things.
[23:17:10] <djlee> nineteen1ightd: not sure if this helps you but: http://stackoverflow.com/questions/6758339/updating-embedded-document-property-in-mongodb
[23:23:18] <nollidj> so if i have _id fields set to values like { foo : bar, baz : { quux : blech } }, i should see that the index is being used when i make a query like { _id.foo : bar } }
[23:23:36] <nineteen1ightd> Yeah, unfortunately the $ operator in the update part only operates on the first matche element in the array...
[23:25:08] <nollidj> my query is more complicated than that (uses $or more than once), so i'm checking to see if a simpler query does
[23:26:41] <nollidj> when "nscannedObjects" is the cardinality of your collection, that's bad, right?
[23:27:29] <nollidj> i can't confirm that an index is being used --- indexOnly is false, but that makes sense since the query i'm running targets parts of the _id field
[23:29:53] <nollidj> bizarre. given documents like { _id : { key : value, key : { nested : stuff }, ... }, value : thingy }, queries like { "_id.key" : value } don't appear to be using an index
[23:32:39] <Joeskyyy> nineteen1ightd: Did you get yours solved? Because it sounds like you are looking for the positional solution
[23:33:02] <Joeskyyy> Forgive me if I'm not understanding you exactly though :P it is about time to go home so my brain is drained
[23:37:07] <Joeskyyy> Especially since you'd have to create an index for all possible fields to accomplish that, which is not pretty at all
[23:52:33] <Joeskyyy> Ahhhh, I found it nollidj: http://docs.mongodb.org/manual/core/index-single/#indexes-on-subdocuments
[23:52:56] <Joeskyyy> Querying the subdocuments must match the ordering, which makes sense when you think about how the indexing is working on something like that
[23:53:09] <Joeskyyy> db.test.find({"_id":{"foo":"bar", "baz" : { "quux" : "blech" }}}).explain().cursor
[23:53:21] <Joeskyyy> So you may want to create some indexes yourself if you don't have a bunch already.