Log file Viewer
#mongodb logs for Friday the 2nd of May, 2014
[05:35:37] <joannac> and if you had asked your question in the first place, you might have an answer by now :)
[05:38:51] <sahanh> thought the same, wasn't sure if it's the best approach, since I have to define several indexes just for the count part
[05:42:35] <joannac> well, no, you don't have to. But indexes will get you better performance than collection scans
[08:06:59] <BaNzounet> Hey guys, if I've a doc multiple keys and I run an update with only one key in the update section will it override other field or will it just update the matching field?
[08:53:48] <ahawkins_> hey everyone. I have an array of strings, I need to query documents where an entry matches a regex. I used $in with $regex before but newer mongo versions are complaining about that. Any other way to construct this query?
[08:57:45] <ahawkins_> rspijker: given a document with an property "things" [foo, bar, baz], find all documents where a member of "things" matches /b/
[09:00:01] <ollivera> Hi, is there any way to check the last time my collection has been updated? I tried db.getLastErrorObj()
[09:00:19] <rspijker> say "things" is a regular string, either "foor", "bar", or "baz". Then you can use $in:["foo", "bar", "baz"] to find any document that has things either one of those values
[09:00:34] <kali> ollivera: not trivially (you can look at the oplog if you have ha replica set, but it's a bit hackish)
[09:02:04] <rspijker> if this is something you want to do routinely, you need to keep track of it yourself ollivera
[10:55:08] <BaNzounet> Is there anythings wrong with this ` db.bars.update({installedOnce: true}, {$set: {installedOne: false}}, false, false);`? Cause it doesn't update my doc
[11:34:04] <oskie> hello, it seems the minimum size of an empty collection is 8MB, is there anything you can do about that?
[11:38:17] <rspijker> all I can recall is that it limits the max size to 512, don't know if it also makes inital files smaller
[11:39:55] <oskie> rspijker: I don't think the files are the issue - I have like 100 databases, each with 16 or so collections with "size" : 0 and "storageSize" : 8192...
[11:42:56] <rspijker> I can't find any documentation that explains what those settings do *exactly*, only somewhat vaguely
[11:43:26] <rspijker> I remember there being a video seminar that went into mongo storage mecahnisms into a little more depth, but that was a while ago
[12:06:12] <kali> smallfiles reduce the size of the preallocation chunks from 2GB to 128M iirc. it's a must for integration tests
[12:06:56] <kali> noprealloc is good when running on filesystem that do not support preallocation (can be usefull for tests, too)
[12:07:12] <kali> but i would not consider these two tweaks to be suitable for production environment
[13:14:09] <Industrial> Hi. I have several collections that represent time based data. Right now they all have one property 't' which is a Date object (so down to the millisecond).
[13:14:35] <Industrial> Is it possible to merge several collections (with the same format) based on one property, also sorted by that property?
[14:18:03] <mac_cosmo> hey! i’m new to mongodb and can’t get my head around basics things … let’s assume, that i have an old projekt with a psql/mysql db, like this: http://pastie.org/9133802 what would be a good design when i have to redo this in a mongo db?
[14:30:01] <mac_cosmo> it’s something like messages in facebook. an user can create a “message group” and invite other users.
[14:32:31] <mac_cosmo> a user should be able to query the parties he joined, and a party should be able to query all of it’s users
[14:35:42] <Nodex> parties : -> ... { _id : ...., oid : ...., moderators: [], users : [{id:1,name:"bar"....}] }
[14:38:05] <Nodex> and drop an index on users.id to be able to query all parties a certain id belongs to
[14:38:50] <Nodex> BUT... there is a 16mb document limit so you will need to rethink this when your app gorws
[14:41:15] <artimuscf> hi, im trying to write a .js script to clone a database. it works fine, but I would like to add some logic so I know if it was successful or not. I currently use var out = db.cloneDatabase('somehost'), but out is an object that im not sure how to parse
[14:42:24] <artimuscf> so basically i just do db.dropDatabase(), then db.cloneDatabase(). i just want to add some error handling. can anyone point me in the right direction
[14:47:43] <emr> Hello how i can find object which have array list and contains item? its something like { "_id" : "37a6259cc0c1dae299a7866489dff0bd", "fruits" : [ "apple", "banana" ] }
[14:48:17] <rspijker> you can just act as if fruits is a simple field containing a string you want to match against emr
[15:12:57] <asturel_> if i made accidently a typo in my field name can i someshow change them to something else?
[15:31:40] <artimuscf> please, can anyone help me determine if db.cloneDatabase() was successful in javascript
[15:44:14] <ekristen> I’m afraid this isn’t the first time this has happened — trying to figure out why
[15:54:05] <jbrechtel> I've searched a bit and haven't found anything....has anyone seen anything done around rollback strategies for failed document schema migrations when using a replicaset?
[15:55:46] <jbrechtel> I'm looking for ideas around what to do (either automate or plan to do) if we do a document schema migration and decide it has failed and we want to rollback. We've considered taking a member out of the replicaset and subsequently building a new replicaset around that in the event of a failure. This requires two application configuration updates though. Looking for possibly simpler approaches.
[16:25:10] <ekristen> I’m having what appears to be issues with one of my replica’s secondaries, certain queries seem to get stuck and will run for hours
[16:25:50] <ekristen> is there anyway to troubleshoot why this is happening? it doesn’t seem to be happening on my primary and/or my other secondary
[16:32:19] <magglass2> ekristen: have you checked to make sure any indexes required by the query are in place? compare the indexes to those on your master
[16:43:49] <dfinch> I'm using a third party component which creates tons of collections implicitly, I want to ensure these are configured correctly and indexes are added when they are created. my plan is to edit the library's calls to update/etc. to insert some step to do this. this will presumably have to be done on every write operation, so what's the best way to do this so I don't seriously affect performance of the queries
[16:46:58] <magglass2> ekristen: can you run the query with a .explain() and compare it to the master?
[17:21:36] <unholycrab> I want to monitor multiple clusters that cannot reach eachother, and so cannot be monitored from a single mongodb-mms-monitoring-agent
[17:39:18] <daidoji> hello, how do I know if I'm running into memory limits or whatever with MapReduce in Mongo?
[17:39:37] <daidoji> alternatively, here's some output explaining an issue I'm having that I can't seem to figure out https://gist.github.com/daidoji/1ef55617e69553afc4f8
[17:46:07] <daidoji> also note, that second bit of output the count is correct, there are three documents with that type not two as in the first set of output
[18:04:48] <pure> So, considering a sharding use case whereby the users of the service can host either shards themselves or redundancy shards. Would there be a nice way to make sure the sharded databases are somehow encrypted from the host but still able to be used by the rest of the DB network?
[18:15:26] <magglass2> pure: not sure what you're asking; typically each shard would be a three-node replica set
[18:16:17] <pure> I mean, each shard in the database should have its own redundancy databases thingy?
[18:29:42] <bemu> I would expect to see some kind of output, perhaps the remnants of cfg, but no dice. Am I doing something wrong?
[18:36:05] <daidoji> bemu: where does cfg come from? I've never done sharding, but empty [] is acceptable output for splice when there's nothing at that index
[18:37:23] <daidoji> also, in case anyone could help with my issue above, its now also a Google Groups post for mongodb-users https://groups.google.com/forum/#!topic/mongodb-user/C68gT2kn7O8
[18:39:54] <daidoji> bemu: isn't splice a javascript function anyways? I don't think its a Mongo specific thing
[18:40:47] <mbeacom> {'id': 1, 'another_key': 2, 'messages': [{'text':'hey', 'timestamp': 1397671947403}, {'text':'hey', 'timestamp': 1397683296710}]}
[18:41:16] <mbeacom> I want to filter my collection where the child messages timestamp is less than a timestamp i provide
[18:42:25] <mbeacom> I have db.messages.find_one({'id': id, 'messages.timestamp': {'$lt': 1397683296710}})
[18:44:06] <xissburg> how about db.messages.find_one({'id': id, 'timestamp': {'$lt': 1397683296710}})
[18:48:01] <daidoji> https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/splice here's the docs on that function though
[18:51:09] <daidoji> Mongo docs assume you're familiar with Javascript though in case you run into other things that don't seem to have documentation in Mongodocs proper
[19:44:43] <q851> mbeacom: I only know of http://docs.mongodb.org/manual/reference/operator/query/elemMatch/ to limit subdocuments.
[19:53:01] <mbeacom> q851: To limit the subdocuments it returns, would you just let it return what it is going to return and then programmatically add htem to an array
[19:53:45] <mbeacom> q851: so I can make a call to the server, grab 15 before a timestamp. then if the user scrolls up more, grab another 15 before a timestamp
[22:40:50] <frodo_baggins> Which brings me to my question... is there a way to have timed tasks occur within mongo, without using crontab?
[23:20:42] <frodo_baggins> Should I create a collection called application, and then a document for settings? Or should I add a collection under system called application_settings?
[23:21:37] <frodo_baggins> Okay, so, documentation is telling me not to create collections with the system.* prefix.
[23:31:08] <frodo_baggins> I want to use findAndModify to query for the presence of an element within an Array on one document.
[23:49:02] <frodo_baggins> Is there a way to tell if an update operation ends up inserting a new document?
[23:50:04] <frodo_baggins> I would like to be able to call a method, or at least specify a conditional $addToSet on the event a new document is inserted.
[23:52:27] <frodo_baggins> Is there a way to tell from the result whether an insert or an update was performed?
[23:56:03] <frodo_baggins> And by "returns" I'm hoping by passing arguments to the callback it receives.
[23:57:18] <frodo_baggins> I'm using the node-mongodb-native driver, and I don't see any mention of the WriteResult object.