Log file Viewer
#mongodb logs for Monday the 7th of January, 2013
[00:58:37] <timah> am i correct in saying that regardless of content, the default and max document size is 16MB?
[01:06:02] <timah> but it seems like my document size is consistently 16MB regardless of how little, or how much i stuff into it.
[01:18:27] <multi_io> i.e. such that you can perform multiple inserts/updates, all other reads and writes are blocked and don't proceed, but if the lock isn't held, read operations run in parallel?
[01:19:18] <multi_io> (I'm not talking about mongodb's own internal per-database r/w lock -- that one isn't customizable afaics, i.e. you can't make it span multiple write operations)
[01:41:47] <_aegis_> http://docs.mongodb.org/manual/reference/operator/atomic/ and http://docs.mongodb.org/manual/tutorial/isolate-sequence-of-operations/ ?
[07:17:35] <bosyak> hi all. I can't query find() method with Spring Data. Please drop me an example?! mongo.executeCommand("{'find':'firststeps', 'article':'art222'}");
[07:30:10] <Ro__> Hello, I need GridFs with version-able files. After seeing some examples I found that python client has get_version method. Although php driver does not have this method (or I haven)
[07:30:21] <Ro__> Hello, I need GridFs with version-able files. After seeing some examples I found that python client has get_version method. Although php driver does not have this method (or I haven't found any).
[12:06:40] <BadCodSmell> Occasionally it runs out of memory, I have a fix for this, stop the server, delete the data files, reimport the data, restart the server, restart everything else
[12:08:22] <BadCodSmell> The difficulty is in detecting when mongo enters that condition programmatically
[12:09:48] <algernon> though, hm, since you reimport, capped collections won't really do what you want
[12:10:27] <BadCodSmell> It's just is there a good way to trap the error you get outside of the application and mongo when you hit that limit?
[12:54:23] <osallou> Hi, I am trying to call the find method on a connection using python reflexion "db_conn.objklass.find(filter)" but I have errors: TypeError: 'Collection' object is not callable. This works fine if I use db_conn.User.find(filter). Anyway to use a class reference directly?
[13:57:40] <Virunga> I configured mongodb to listen on port 27017 and accept connection from a comma separated list of ips, and starting the service i get mongodb start/running, process 1804. Is it enough to let the service work?
[14:18:19] <timah> jeremy: if i were asking that very same question i would first start by challenging myself in finding a much healthier approach in accessing the data.
[14:22:32] <jeremy-> I doubt there is a healther way, i understand its an expensive query but its designed as opposed to doing an odd few dozen queries instead
[14:23:05] <jeremy-> Ie, against 3 different fields, ($or), search against a user defined array for each ($in)
[14:23:53] <jeremy-> The alternative is to only allow the user to do one of the 3 fields at a time, then it can be indexed, not sure if i want to restrict the user like that
[14:25:36] <jeremy-> I thought the above worked, (i could have sworn i got something similar working a few days ago).
[14:28:44] <jeremy-> Cause ive got a test case and its just one line i guess i can paste in here: { $or : [ { "firstfield" : $in ["val1", "val2"] }, {"secondfield" : $in ["val3", "val3"] } ] }
[14:34:37] <timah> jeremy: thank you, but rest assured that next time i'll have finished at least 1 cup of coffee.
[14:54:35] <ichilton> I have a Groups model and I have a need to store 'affiliated groups' - I was thinking of having an array of group id's, but I need to store other details with them, such as the date they were affiliated and an accepted yes/no. What's the best way of storing that in MongoDB/Mongoid?
[14:54:40] <ichilton> In SQL, i'd obviously use a separate link table, in effect doing a many-to-many with itself.... but what's the right way to do that in MongoDB?
[14:59:12] <wereHamster> group is { name: 'The Group', affiliatedGroups: [ { _id: X, accepted: true, affiliatedAt: Date }, ... ] }
[15:17:45] <jeremy-> Hello, on the node we were discussing before, I found out where my issue was after you guys gave me the advice, i created a pastie to try and explain my problem
[15:18:23] <jeremy-> Basically, I want to programmatically design some conditions for mongo that may or may not exist using php
[15:18:56] <jeremy-> If i write the statement and create a php associative array for each mongo statement, i can give each statement a curly bracket, but if i try and stack all my conditions in one array, they all share a curly bracket so it doesnt work
[15:21:30] <sanderj_> Can anyone recommend where I can get my php/mongodb application hosted as a platform?
[15:21:41] <jeremy-> I forgot to add to the pastie how i am getting my test case that doesnt work, obviously: array( '$or' => array( $or_array ));
[15:22:26] <NodeX> jeremy.... the best way I try to attack nesting problems is make it work in Json then reverse it in php
[15:25:36] <jeremy-> Just the bottom one, where i build the statement programatically, array( '$or' => array( $or_array )); the json_encode output turns out like the last line
[15:29:26] <jeremy-> Succinctly, I want print_r(json_encode($associative_array)) to put a curly bracket around each key/value, instead of only one curly bracket around the entire thing
[15:30:09] <hersha> All of the sudden one of the databases on my shard cluster started throwing a "setShardVersion failed" on one of my nodes. This database has been running fine for 4 months with no problems. Any ideas?
[15:32:35] <jeremy-> NodeX, if you look at my original paste, see how the one you created does the same thing
[15:33:26] <hersha> "$err" : "setShardVersion failed host: 10.11.12.86:27017 { oldVersion: Timestamp 0|0, oldVersionEpoch: ObjectId('000000000000000000000000'), ns: \"tekno_records.tekno_records\", version: Timestamp 1060000|21, versionEpoch: ObjectId('5064a757e45bc0a1544406b9'), globalVersion: Timestamp 1061000|1, globalVersionEpoch: ObjectId('5064a757e45bc0a1544406b9'), reloadConfig: true, errmsg: \"shard global version for collection is higher than trying to set to
[15:33:59] <jeremy-> so the way that i know thats each to make each statement in its own curly bracket is to have each as an array
[15:34:25] <jeremy-> so you can write it like my original code, but then I dont know how to make it dynamic (conditionally not have one of the elements)
[15:34:39] <kchodorow_> hersha: try running db.adminCommand({flushRouterConfig:1}) on the mongos giving you errors (or restart it)
[15:36:42] <NodeX> it's fine jeremy- ... $or's can be confusing, give me a few mins, it's crazy busy here
[15:37:18] <jeremy-> now that i have actually written down the problem i think i may eventually get there myself too
[15:37:56] <jeremy-> Oh, you know how you could do it, create a for loop, and write every value of that array into a new array
[15:38:11] <BadCodSmell> I just did this to cram my db: http://pastebin.com/EX6ftESr I understand how you can get slack space when things aren't packed and mongo doesn't unallocate after removal, but this is pretty huge (6x), is there any documentation coverring specifically this kind of issue and minimising bloat?
[15:39:46] <timah> i've asked my question a few times. however, seemingly at just the wrong time each time.
[15:40:04] <BadCodSmell> The reason I find this bloat weird is because all I rarely update and mostly mongo restore every now and again to it (poor man's replication), it's read only otherwise.
[15:41:57] <timah> here's my document: { _id: { m: <int>, o: <objectid> }, a: <int>, b: <int>, c: <int>, d: <int>, e: <int> }
[15:42:37] <timah> aside from the size of the various ids, the size of the document does not fluctuate, at all.
[15:43:20] <timah> however… it seems that my minimum document size is always 16MB… no matter how small i make it.
[15:45:41] <NodeX> {"$or":[{"abc":{"$in":[1,2,3]}},{"def":{"$in":[4,5,6]}},{"ghi":{"$in":[7,8,9]}}]}
[15:45:45] <timah> all of my seemingly endless research has only led me to posts/docs concerning increasing the minimum document size from 16MB.
[15:49:03] <timah> i don't know what my /actual/ document sizes are… but i can bet they are much smaller than 16MB.
[15:51:02] <hersha> timah: pass your document to Object.bsonsize() to see the actual size of the document.
[16:05:30] <hersha> Any other ideas on my setShardVersion failed issue? About ready to rebuild the collection.
[16:06:28] <sanderj_> NodeX, when it says replication.. Does it mean sharding/spreading of data around? That's atleast what I want.
[16:08:11] <sanderj_> hersha, so basicly most mongodb hosting is useless if you want to scale writing with over 1 server?
[16:08:46] <NodeX> sander__ : I dont have a clue, I dont use them and never would because they're far to expensive
[17:20:10] <w0rmdr1nk> cos I dont like the limitation of 3 config servers, especially if there is one production and one disaster recovery data center
[17:21:37] <w0rmdr1nk> and in general I would prefer a config server on each node - especially with large clusters, otherwise there is more sensitivity on downtime for specific nodes.
[17:34:54] <drunkgoat> what would you say a better approach for a chat server with mongo: capped collection with tailable cursor or an indexed capped collection using a regular cursor?
[18:10:00] <timah> is it more performant to ensure a consistent document width within a collection vs. keys with null values?
[20:03:24] <Virunga> With the node.js mongodb driver, if one wants to perform an insert/update must retrieve the collection first and then make the operation. Does really retrieve the whole collection? Is there a more efficient way to do this?