Log file Viewer
#mongodb logs for Friday the 25th of October, 2013
[00:52:18] <eph3meral> jnewt, remove those "..." parts and surround the relevant portions with db.somecollection.insert(); for example
[00:52:50] <eph3meral> jnewt, include a use statement at the top, make it *super* easy, otherwise I personally at least (and probably many others) don't have or care to take the time to do it ourselves for your benefit :)
[06:50:34] <omnidan> I have a quick question. I trying to get a list of objects from a collection by searching for the ID of the reference. the reference is stored in "assigned_to" and I tried {"assigned_to.id": "ID"}, {"assigned_to": {"id": "ID"}} and ObjectId("ID") instead of "ID", none of these queries worked
[08:27:15] <jihyun> I added existing mongod as a shard, enable sharding for exising collection (~1GB)
[08:27:52] <jihyun> but collection is not distributed into multiple shard, and all queries are directed into single mongod instance.
[08:34:24] <jihyun> There are more collections, but all of them are in similar status (1 ~ 2 chunks)
[08:40:06] <kali> can you check your balancer ? http://docs.mongodb.org/manual/tutorial/manage-sharded-cluster-balancer/#check-the-balancer-lock
[08:47:39] <jihyun> from config database -> { "_id" : "gm-sdb1-2013-10-25T07:53:22-526a2372cf425fdbcd5c1cdc", "server" : "gm-sdb1", "clientAddr" : "10.17.14.52:49558", "time" : ISODate("2013-10-25T07:53:22.393Z"), "what" : "split", "ns" : "development.gamedata", "details" : { "before" : { "min" : { "_id" : { "$minKey" : 1 } }, "max" : { "_id" : { "$maxKey" : 1 } }, "lastmod" :
[08:47:39] <jihyun> Timestamp(1, 0), "lastmodEpoch" : ObjectId("000000000000000000000000") }, "left" : { "min" : { "_id" : { "$minKey" : 1 } }, "max" : { "_id" : NumberLong(0) }, "lastmod" : Timestamp(1, 1), "lastmodEpoch" : ObjectId("526a23713e04684b30d465d3") }, "right" : { "min" : { "_id" : NumberLong(0) }, "max" : { "_id" : { "$maxKey" : 1 } }, "lastmod" : Timestamp(1, 2),
[08:49:14] <jihyun> Will balancer splits existing chunk? or does it only moves existing chunk to other shard?
[08:50:38] <kali> iirc, when you actually run the initial shard collection command, it blocks while the initial split happens
[09:19:43] <schovi> Hello around. I am fighting with Mongo on Debian. I installed it via apt-get install mongo-10gen
[09:46:12] <_Heisenberg_> Will updates that have been journaled get rolled back in case a primary steps down without replicating that update before?
[09:48:51] <eph3meral> is there any kind of layer similar to elasticsearch where I can provide "boosts" etc to various things so that I get a semi mixed bag of results
[09:49:29] <eph3meral> e.g. I should preferentially see higher paid listings first, but only to a point, after about say 25 miles or so I don't care about paid listings and I would get more use out of listings closer to me
[11:14:19] <omnidan> I have a quick question. I trying to get a list of objects from a collection by searching for the ID of the reference. the reference is stored in "assigned_to" and I tried {"assigned_to.id": "ID"}, {"assigned_to": {"id": "ID"}} and ObjectId("ID") instead of "ID", none of these queries worked
[11:16:29] <joannac> erm, show us the "assigned_to" field and the structure of your query? pastebin it pls
[11:32:42] <joannac> If this isn't a one to many relationship, have you considered using manual references instead?
[11:36:58] <omnidan> joannac: interesting, I'm trying to get this via doctrine though and used .id in another ref and it worked
[11:38:00] <omnidan> $data->findBy(array("assigned_to.id" => $id)); doesn't work but it works in another collection
[14:19:04] <briefcaseofstuff> how can i add a query to this aggregate? (ie where itemtype:"something")... http://paste.laravel.com/11Ls
[14:36:45] <briefcaseofstuff> nvm i got it, i think its supposed to be an obj in the array passed to aggregate, not part of the first obj
[15:14:17] <Nodex> anyone ever sed Mac in a cloud type services? - i/e renting a mac desktop and connecting to it remotely?
[15:42:32] <Xzyx987X> okay, this is bizzarre... why am I getting the error "Can't take a write lock while out of disk space" when my database is located on a disk with over a terrabyte of free space? and also, why is it that when I restart the program that's writing to the database the error goes away, at least for a while, until it starts happening again a few hours later?
[16:35:41] <dharmaturtle> Hi, I'm trying to follow the directions here http://stackoverflow.com/a/3605054/625919
[16:37:38] <kali> dharmaturtle: http://docs.mongodb.org/manual/reference/command/mapReduce/#dbcmd.mapReduce
[16:37:38] <dharmaturtle> oh. how should I modify it then? I'm unfamiliar with finalize... very new to all this.
[16:38:16] <kali> dharmaturtle: http://docs.mongodb.org/manual/reference/command/mapReduce/#mapreduce-out-cmd or that
[16:47:20] <dharmaturtle> Is it possible to write to a collection from finalize or reduce? When I try it, I get TypeError: Cannot read property 'myCollection' of undefined near { db.myCollection.insert(value)' ",
[16:47:55] <dharmaturtle> and when I try to create to instanciate Mongo/db, I get: Mongo is not defined near 'conn = new Mongo();
[16:56:06] <kali> this is wrong. you must consider map, reduce and finalize as pure function, withou side effect
[17:24:15] <t0th_-> i have a "$err" : "unauthorized db:test ns:test.system.namespaces lock type:1 client:127.0.0.1",
[18:37:25] <orngchkn> I'm currently crushing the crap out of our mongodb by using it with delayed::job (which uses findAndModify to atomically lock jobs in the queue)
[19:45:42] <dweeb_> If I run mongod using the nojournal=true option. Does mongo still store the data in the filesystem or are all data lost if I quit the process?
[19:56:33] <bjori> dweeb_: but if mongod or the server crashes before it flushes the new write to the filesystem you may loose some commits
[19:56:34] <dweeb_> I accidentally removed data from my db. So I restored the server from a snapshot. When the server restarted mongodb couldnt start so I removed the mongodb.lock file and now all my databases are empty. I ran repair but all dbs are still empty.
[19:58:31] <bjori> dweeb_: shut down mongod, restore the data, check the filesize, startup mongod again
[20:00:33] <dweeb_> bjori: So I restore my server from snapshot. Shut down mongod process. Remove the lock file. Run mongod with --repair option. And then start mongod again with right dbpath option and file permissions. is that all?
[20:23:12] <orngchkn> Given this query (findAndModify) https://gist.github.com/contentfree/cdfa6b8c4e0c49e63b49 could anyone help me adjust the indexes so they, ya know, get used? This query is generated by a ruby gem called delayed_job_mongoid and I'd love to contribute a fix back to their project, if possible.
[21:02:59] <ygnk> Would anyone know how to write pig schema to load array of ints frmo mongo for mongo-hadoop?
[22:50:48] <eldub> in a 3 node replica set... does 1 node NEED to be arbiter? Or will failover still be handled in a 3 node replica set
[23:02:24] <jyee> eldub: no, in a 3 node replset, they can all be regular nodes (1 primary, 2 seccondary).
[23:04:00] <jyee> assuming you had 2 die, then you'd at least be able to read (just not write, since you wouldn't be able to get a voting majority).
[23:07:33] <eldub> wait a sec... jyee if only 1 goes down... say the master... then there won't be a majority to vote and elect a new primary....?
[23:13:05] <jyee> in a 3 node set, if primary goes down, you have 2 secondaries to vote and one will be elected primary
[23:13:57] <jyee> because the overall set is still 3, so 2/3rds majority if they agree… which they should, unless something odd happens
[23:22:41] <a|3x> when i use db.copyDatabase() command on localhost (3 gb database), its seems to execute very slowly, it looks like it is copying the data quickly but reindexing