[13:34:53] <ksa_> Gentlemen (and women!) - I'm having a problem figuring out how to write a specific map/reduce and was wondering if anyone in here could be convinced to impart their assistance / knowledge unto in return for some cash
[13:48:34] <fiws> does gridfs has its own channel?
[13:55:13] <fiws> ok my question: can a chunk be assigned to multiple files?
[13:59:11] <ksa_> Can anyone help me figure out how to get data in this format: http://pastie.org/private/scmo3ccehstcr2csfsrtsg to be reduced to show the amount of "useful" votes (field: votes.useful) for a single day (field: date)?
[17:13:27] <Dark_Sun> I'd like to ask this question: what happens if a master fails in a master-slave deploy?
[17:14:24] <kali> just nothing. you'll find yourself with a dead master and an idle slave
[17:14:40] <kali> this is why you probably want to use replica set instead
[17:16:44] <Dark_Sun> yes, replicaset as well,but here is my concrete problem: i need to process a large amount of data that takes hours and hours for few updates. i was thinking a master-slaves configuration where i use the master to produce these few updates to propagate to all the slaves that are in replicaset on the other collections
[17:17:43] <Dark_Sun> it came in my mind: what would happen if the master disappear forever ? i mean, i would dump the data from one slave and create another master from that dump, but is this possible?
[17:18:46] <kali> in what you describe previously, i have no idea. in a standard master slave, you stop the slave and restart it as a master and make sure your application talks to it
[17:19:10] <Dark_Sun> or well, i could try replica set using a single node to the massive read and few writes
[17:21:51] <Dark_Sun> but the problem remains. Let's change the question. Is it possible in a (A,B,C) replica set to use only (A,B) as read nodes ?
[17:30:55] <Bubo> Would it be possible to design a schema for an RSS reader? Users have subscriptions, subscription has feeds
[17:31:06] <Bubo> or is SQL database more suited for this?
[17:49:05] <jtomasrl> why "$or" isnt working if one of the attribute doesnt exists?
[19:08:47] <Pinkamena_D> Hello. I have a project to replace an extremely old access 2 db app with something modern. The program is used by a small group of people and has been used by launching from a network drive. I an considering either sqlite or mongo db.
[19:09:21] <Pinkamena_D> I like mongo better but I am not sure if their is some way to have the db "embedded" as sqlite is.
[19:10:10] <Pinkamena_D> (the computers which use the program are not always on)
[19:10:42] <Pinkamena_D> Is their a way to kind of have the server launch with the gui app, or would the only way be to use some remote server?
[19:13:35] <missinglink> is it possible to clear the oplog for a mongodb server?
[19:21:41] <awpti> Pinkamena_D, sounds like SQLite is the way to go.
[21:18:08] <Milos_> is there a way to tell mongo not to return some field from the requested document?
[21:59:23] <Beg> How do I invert a boolean value in a update modifer?
[22:04:19] <drag> Hello all. I'm looking for a NoSQL database to store lots of data of different types including test, json objects, shapefiles, etc... I'm trying to decide between HBase and MongoDB. I'm looking at comparisions online and there seems to be a few comments that say something along the lines of "I wouldn't trust big data to mongodb". Can anyone enlighten me on what exactly may those comments be referring to?
[23:20:30] <astro73|roam> are there any guidelines or suggestions on what should go in a collection together and what should be separated?