PMXBOT Log file Viewer

Help | Karma | Search:

#mongodb logs for Friday the 26th of April, 2013

(Back to #mongodb overview) (Back to channel listing) (Animate logs)
[01:26:36] <kexmex> lol
[01:26:37] <kexmex> i upgraded to latest mongo
[01:26:40] <kexmex> now it doesnt start up
[01:26:43] <kexmex> complaining about dupe users
[01:44:26] <niftylettuce> anyone else having issue: `[Error: kerberos.node: undefined symbol: init]`
[08:45:48] <Bilge> Is there any way to achieve control over how documents with missing or null keys are sorted?
[08:46:02] <Bilge> By default missing key documents are included at the head of the sort order
[08:46:23] <Bilge> How can I sort a key in ascending order but have documents with missing keys included at the end of the series?
[08:47:39] <Nodex> not sure you can tbh
[09:09:03] <Bilge> So you're saying mongodb sucks? Why would you say that Nodex
[09:09:16] <Nodex> when did I say that Bilge ?
[09:09:46] <Nodex> I have been using MOngoDB for ~3 years and not once have I said that
[09:12:21] <pcmaniac> hey, does anyone know why the primary shard gets most (95%) of the queries?
[09:12:30] <pcmaniac> shard key is { _id: 'hashed' } and queries do not use the shard key, but data should be uniformly balanced on both shards
[09:12:52] <pcmaniac> current chunk ratio between these 2 shards s 703 : 704
[09:12:57] <pcmaniac> *is
[09:15:59] <pcmaniac> if i "movePrimary" to another shard, then that's the one that will get most of the queries..
[10:19:14] <halfie|bbl> hi, I get "ERROR: Insufficient free space for journal files" when I try to use a RAMDISK for storing data files.
[10:19:51] <halfie|bbl> seems like a bug in mongodb as I have 64GB of space in the RAMDISK and I can create very huge files (using dd) in that RAMDISK
[10:23:59] <Bilge> You just said you can't sort null/missing keys which is the same thing as saying that it sucks
[10:24:05] <Bilge> Who would even design a database search like that
[10:26:36] <Nodex> then go use something else - simple :)
[11:00:02] <rawl79> Hi there. How to I enable server-side Javascript? I'm getting the following message when I try to run sh.status();
[11:00:03] <rawl79> server-side JavaScript execution is disable
[11:19:29] <Nodex> anyone from 10gen confirm somehting about "mongoexport" ?
[11:20:31] <Nodex> - it seems to spit out json without slashes added - I have written something to output json to mirror this (but it adds slashes as per the JSON spec) and I would like to know if mongoimport will choke on this
[11:22:19] <Nodex> I have managed to write a sh** hot incrememntal backup tool for mongo \o/
[11:33:47] <rawl79> Looking for someone who can tell me what I'm missing or how do I enable server-side JS execution. Here's I'm getting:
[11:33:48] <rawl79> mongos> sh.status();
[11:33:50] <rawl79> Fri Apr 26 07:29:40 uncaught exception: group command failed: { "errmsg" : "server-side JavaScript execution is disabled", "ok" : 0 }
[11:36:38] <kali> rawl79: you probably run with --noscripting
[11:36:55] <kali> rawl79: or its configuration file equivalent
[11:46:03] <avinoamr> Hey fellas. Quick question. I run a find() query on a collection that uses a compound index on 2 identically-configured shards. On one shard it's done blazingly fast, and explain() shows that it uses the index. The same query on the second shard, doesn't use the index and takes a long time to complete. When I use hint() for the query, it uses the index and completes as fast as the other shard. Can you please give me any pointers i
[11:46:05] <avinoamr> use the index?
[11:49:09] <DinMamma> Hello everybody. Ive added a new shard to my cluster, and waiting for the chunks to redistribute which takes for ages. Since 5oclock last night I have migrated 8million docs, when the cluster is evenly distributed there should be 35.5mi docs on each shard.
[11:49:21] <DinMamma> Anywho, I dont want to wait another two days for it to catch up
[11:49:55] <DinMamma> Is it safe to use the cluster as normal while this is going on?
[13:07:49] <pcmaniac> @DinMamma: as far as i know, you should be able to use it as normal
[13:08:16] <DinMamma> pcmaniac: coolio. Thanks
[13:08:31] <pcmaniac> the only problem i saw while quering during the balancing was that the count() reported inconsistent numbers
[13:21:05] <pcmaniac> can anybody who already used sharding help me with this? https://jira.mongodb.org/browse/SERVER-9480
[13:31:06] <[S^K]> How can I sort a nested list?
[13:35:16] <Nodex> dot notation
[13:35:30] <Nodex> db.foo.find({}).sort({"nested.list":1});
[13:35:40] <Nodex> anyone used thise guys before ? -> http://greenqloud.com/
[14:16:48] <honi> Hi, is there a channel for MongoDB JS driver?
[14:17:49] <Nodex> some people in here help/support it
[14:21:35] <honi> Well, I'll go ahead and ask then :)
[14:21:38] <honi> So I'm running an express app, and I want to wrap my db connection in a module which will be required in many other modules. I'm using MongoClient to connect to the db. But since the connection is async, how do I export the db object in my db module?
[14:35:26] <saml> honi, whynot connect(url, callbcak)?
[14:35:46] <honi> saml: I'm reading this: http://stackoverflow.com/questions/11991544/nodejs-mongodb-use-an-opening-connection
[14:36:02] <honi> basically what you are saying? Or I'm wrong?
[14:36:21] <Petrogad> honi: any reason you're not using mongoose?
[14:36:23] <saml> i mean i wouldn't want importing a module to mean open db connection
[14:36:27] <saml> that's wierd
[14:37:03] <honi> I'm just lazy and don't want to repeat 10 lines of code everywhere I need a db connection
[14:37:26] <honi> but I think I could wrap that in a function and export that function
[14:37:32] <saml> abstract it out ?
[14:38:00] <saml> i don't know node.js.. you can't do blocking call?
[14:38:08] <Petrogad> you don't want to do any blocking calls in node
[14:38:17] <Petrogad> it's one instance
[14:38:21] <honi> you can't block when requiring
[14:38:36] <saml> http://mongoosejs.com/ looks like blocking
[14:38:42] <saml> mongoose.connect('mongodb://localhost/test');
[14:39:07] <honi> I don't know how they do it, should check out the source
[14:39:14] <Petrogad> i'm fairly certain it isn't
[14:39:17] <saml> actully no
[14:39:18] <Petrogad> i use mongoose with a large user base
[14:39:22] <Petrogad> and it's outstanding
[14:39:44] <saml> var db = connect(url); db.on('success', ...)
[14:40:23] <saml> connect(url, callback)
[14:40:32] <saml> javascript bro
[14:41:01] <saml> connect() can implement connection pool and reuse
[14:41:11] <saml> that's how openresty does
[14:41:50] <honi> ok, thanks
[14:50:34] <honi> for reference: https://groups.google.com/forum/#!msg/node-mongodb-native/mSGnnuG8C1o/Hiaqvdu1bWoJ
[14:55:40] <saml> patch taht up
[14:56:04] <saml> i'm sick of being integration plubmer i want to code real stuff
[14:56:19] <saml> everyday i read docs for this library and that library and glue crap together
[14:56:33] <saml> oops wrong chan
[15:00:50] <dhodgin> hello all, i have a question about replica sets. i've set one up in a dev environment and everyday after about 24 hours of it running i start getting Assertion: 10295:getFile(): bad file number value errors
[15:01:05] <dhodgin> i have to shut down the primary, wipe its db, bring it back up and let it sync off a secondary
[15:02:10] <dhodgin> i have a small database initialization script in our application that wipes the db by dropping all the collections, and then inserts about 200 sample records. when i run this after the replica set has been up for a day or so i start getting these assertions
[15:02:43] <dhodgin> when i initialize the replica set fresh i can run the db script a dozen times in a row with no issues, a day later when i come back in to work i get the assertions
[15:03:34] <saml> dhodgin, how does your script connect to replica set?
[15:03:51] <saml> if you're using replica set, connection needs to be different
[15:04:03] <saml> rs0/mongo01,mong02,mongo03,....
[15:04:12] <saml> you connect to something like that
[15:04:26] <saml> replica set is worst thing ever
[15:05:16] <dhodgin> we have a node.js app that uses mongoose. i use mongoose.connect(connectionString, options, callback{})
[15:05:45] <dhodgin> my connection string is of the form: mongodb://192.168.0.10:27017/dbname,mongodb://192.168.0.10:27018
[15:06:05] <dhodgin> with authentication options passed in seperately
[15:06:13] <Petrogad> hmm for some reason I thought mongoose figured out which was the primary for writes automatically
[15:06:28] <Petrogad> (not really on topic though sorry :P)
[15:06:55] <Petrogad> dhodgin: are you doing a ton of document inserts?
[15:06:58] <dhodgin> the latest version of mongoose fixed an isue with the authentication options and i have tested that if i take down the primary my app automatically starts reading from secondary and the secondary does get elected primary after about 3-5 seconds
[15:07:00] <Petrogad> like just growing a particular document?
[15:07:59] <Petrogad> what version of mongo are you on?
[15:08:03] <dhodgin> not a lot of document inserts. the script drops all collections and then begins reinitialzing them, about 20 collections, about 200 documents in total accross db to initiale it. usually takes about 2.5 seconds to run on a single local instance of mongo, i run this replica set on a second server on internal lan
[15:08:06] <dhodgin> im on mongo 2.4.1
[15:08:45] <Petrogad> not sure here then :/
[15:09:18] <dhodgin> in the log files just before i start getting the bad file assertion i get another error: warning assertion failure n == 1 src\mongo\db\index.cpp 221
[15:23:24] <jiffe98> so something's messed up here, none of my mongodb replicas show my database exists and trying to access it or create a new one I get 'can't find a shard to put new db on'
[15:34:34] <[S^K]> is it possible to install mongos without mongod?
[15:42:45] <redsand> [S^K]: yes
[15:48:34] <[S^K]> redsand: how?
[15:59:52] <redsand> [S^K]: you just run it on another box
[15:59:54] <redsand> what u mean how?
[16:00:00] <redsand> they're all independent of each other
[16:00:12] <Nodex> I think he means without actually installing mongod
[16:00:30] <redsand> mongod doesnt have to be installed on the same box
[16:00:50] <redsand> but it must be installed, otherwise what is its value? its an aggregator/router for mongod's
[16:01:49] <Nodex> that's what he;'s asking LOL
[16:01:53] <Nodex> tfi fridya
[16:01:56] <Nodex> friday*
[16:07:01] <jiffe98> so if I have a sharded collection but sharding is broken, can I rebuild it?
[16:11:33] <GRMrGecko> This isn't obvious to me… But how can I run this in the perl driver? "db.posts.runCommand('text', {search: "hands"})" I've tried $collection->runCommand and $collection->run_command
[16:12:03] <GRMrGecko> I've also tried using find.
[16:26:12] <jiffe98> most repo's I've installed mongodb from don't break it out
[16:26:21] <jiffe98> you have to install mongod and mongos
[16:27:18] <jiffe98> it makes sense to break it out though
[18:06:27] <GRMrGecko> How can I do a full text search in perl?
[18:08:02] <ron> how is that related to mongodb?
[18:44:14] <ehershey> I assume they meant a full text mongodb search
[18:44:16] <ehershey> with the perl driver
[18:47:17] <GRMrGecko> ehershey: Yes
[18:47:41] <GRMrGecko> I know how to do db.posts.runCommand('text', {search: "hands"}) in the mongo console.
[18:47:56] <GRMrGecko> But I can't find out how to do this with the perl driver.
[19:28:09] <Tcon> Hello, is anybody out there bored enough to help me out with some optimization? :)
[19:34:12] <damurphy> I want to run the smoke test via scons , how can I specify --smoke-db-prefix=XXXX so that I keep it in my jenkins directories
[19:34:36] <leifw> damurphy: I think the scons option is --smokedbprefix=XXXX
[19:37:25] <leifw> damurphy: yep http://github.com/mongodb/mongo/blob/master/SConstruct#L231
[19:41:32] <jiffe98> so if I have a sharded collection but sharding is broken, can I rebuild it somehow?
[20:13:00] <GRMrGecko> Can anyone help me with text search in perl? db.posts.runCommand('text', {search: "hands"}) works in the console.
[20:29:01] <diminoten_> how do I use $elemMatch against an array of strings
[20:29:03] <diminoten_> so like {'foo':['bar', 'baz']}
[20:29:04] <diminoten_> how do I search for the foo array containing the 'bar' value?
[21:09:44] <aaronshaf> With the npm mongo native module, I get Error: Cannot find module './connection/repl_set' -- any ideas?
[22:27:55] <jiffe98> if I have a sharded collection but sharding is broken, can I rebuild it somehow?
[22:28:46] <redsand> i've only dumped it to disk, then re-setup sharding and re-imported
[22:28:54] <redsand> had issues working with existing sets
[22:28:59] <redsand> jiffe98: ^^^^
[22:31:11] <jiffe98> I see
[22:31:26] <jiffe98> well that's unfortunate but doable
[23:00:28] <pcmaniac> can anybody who already used sharding help me with this? https://jira.mongodb.org/browse/SERVER-9480 thanks