Log file Viewer
#mongodb logs for Monday the 2nd of July, 2012
[02:55:34] <freezey> seems that my db keeps getting *** glibc detected *** mongod: invalid fastbin entry (free)
[04:33:12] <jwilliams> when using skip, limit function, is it possible the docs read size would be smaller than limit specified? for example, spcifying limit 10000, but only 5000 result are actually read.
[04:40:01] <ron> siva: data modeling can change between different nosql database. normally, you should model your data to make querying easier.
[04:42:15] <siva> Ron, Thanks. I'm just planning to use mongodb as it seems to be easier and it can work with C#
[04:43:30] <siva> :( Example, I would just go to have a model Person->Skills->Technical->Competency Level
[04:44:29] <siva> Since I'm just a starter and have explored the basis of MongoDB using Mongly not sure the right way to adapt it
[04:45:13] <ron> again, don't think about how you'd like to store it, think about what kind of queries you'd like to run.
[04:49:58] <ron> siva: have you gone through this: http://www.mongodb.org/display/DOCS/Schema+Design ?
[06:01:08] <calvinfo> has anyone here had a corrupted collection which makes the cursor run in a loop?
[07:41:25] <skython> Hi, is it possible and a good idea to deliver mongodb bundled with a node.js application? I'm looking for a way to build a ready-to-use installation for my app which is using mongodb
[07:52:47] <wereHamster> skython: make a server image, complete with mongodb, node and all other dependencies
[07:54:04] <skython> You mean an image for a virtual maschine ? I would prefer a native way of installation for each platform
[07:56:26] <skython> making an installation wizard for windows, providing a package repository for debian and a dmg for OS X
[07:57:40] <skython> I just need to know if mongodb is easy to bundle without a seperate installation
[08:00:19] <wereHamster> skython: see for yourself. Download the tarball and look inside it. It's not rocket science.
[09:20:28] <stevie-bash> Hello, I just started to use mongotop and I wonder why the local.oplog.rs has a read perfomarnce of 2000ms
[09:48:13] <W0rmDrink> when I turn on full journaling on ext4 shouldl I disable journalling on mongo ?
[09:52:01] <gigo1980> hi i have an problem i have some documents with microtime (unixtimestamp), how can i add to each document an date ? entry. is there an query posible, that can do this
[09:53:38] <W0rmDrink> so mids - ext4 full data journaling wont journal mmap updates is the assumption ?
[09:53:57] <calvinfo> if I am missing a datafile, is there any way to force mongo to continue without it and boot anyway?
[09:54:55] <W0rmDrink> mids - if I run mongo without journaling - how much data can be lost - given I keep sync to 60 seconds ?
[10:50:37] <W0rmDrink> When mongodb updates data files - it does this by just changging the mmaped location ?
[12:01:27] <jondot> hi all, anyone knows how best to handle using unicorn (ruby) and mongodb? the sample https://github.com/mongodb/mongo-ruby-driver/blob/master/test/load/unicorn/unicorn.rb.template encourages sharing one global variable
[12:27:20] <W0rmDrink> it was just the guy said the journal is written before DB - and was not sure how exactly they manage that - since reads are llike read commited - but he soon after explained they mmap the data files 2x
[12:30:25] <FerchoDB> Hi. What is the best way to query the union of an array element of a collection of documents? I mean, I have a collection of documents and each one has an array of Something. I want to query all the "Someghings", but I don't want an array of arrays, but just one array of "Somethings"
[12:37:52] <skot> I think you want $all/in probably: http://www.mongodb.org/display/DOCS/Advanced%20Queries#AdvancedQueries-%24all
[12:40:55] <FerchoDB> You're right, I wasn't clear. I have a collection of Customers. Each Customer document has an array of Orders. I need to query all orders, but instead of having an array of arrays of orders, I just need one array containing all the orders
[12:53:50] <FerchoDB> that's what I thought. I'm still with 2.0.6, but maybe the "MongoDB's New Aggregation Framework" in 2.1 will also do the trick
[13:00:42] <augustl> considering to use mongodb queries with regexps instead of a full text search engine. But that means adding indexes for all fields present in my documents, and the documents typically have many different fields. Specifically, there are groups of documents with similar fields, but the fields differ a lot from group to group
[13:02:34] <deoxxa> augustl: regexes are going to be craaaaazy slow if you have a more than insignificant number of documents
[13:03:37] <augustl> that is, I'll have a gazillion documents in the collection, but the searches will always be scoped on a "foreign key"
[13:03:59] <augustl> so assuming the query planner in mongo is clever enough to filter on that "foreign key" first, it'll be relatively few documents
[13:04:49] <mw44118> i set up a replica set with one primary and one secondary. then the amazon outage killed my primary, so now my secondary is the new primary
[13:05:31] <mw44118> here is my problem: how do i configure my app that talks to mongo so that the app keeps working even after an election?
[13:06:25] <augustl> deoxxa: also just realized that indexes are pretty much irrelevant if it does a full regexp compare, obviously :)
[13:09:21] <deoxxa> yep, much cheaper than spending a couple of weeks restructuring things to work with slightly less hardware
[14:10:39] <arkban> mw44118: think of it like your load balancer, if your primary box disappears and it has mongos, you can't connect to the secondaries
[14:32:20] <Bartzy> What book should I read (a complete novice on MongoDB, need to evaluate it for our production (millions of users) app in the company):
[14:32:22] <Bartzy> http://www.amazon.com/The-Definitive-Guide-MongoDB-Computing/dp/1430230517/ref=pd_rhf_cr_shvl5
[14:32:35] <Bartzy> http://www.amazon.com/MongoDB-Definitive-Guide-Kristina-Chodorow/dp/1449381561/ref=sr_1_1?s=books&ie=UTF8&qid=1341238810&sr=1-1&keywords=mongodb\
[14:32:43] <Bartzy> http://www.amazon.com/MongoDB-Action-Kyle-Banker/dp/1935182870/ref=sr_1_2?s=books&ie=UTF8&qid=1341238810&sr=1-2&keywords=mongodb
[14:34:03] <NodeX> Kristina Chodorow is a 10gen employee and maintains the PHP driver so her book would be gospel
[14:36:28] <Bartzy> I prefer condensed books. "MongoDB: The Definitive Guide" is not updated, or just not as updated as MongoDB in Action ?
[14:42:06] <NodeX> my advice is dont use something fat and bloated in the form of frameworks if using php
[14:43:18] <FerchoDB> What is the syntax to do this, i.e compare two string against the same field? {City : /ondon/ , City: /b/}
[14:45:50] <kchodorow_> Bartzy: TDG is fine for general intro stuff, but it was written before sharding & replica sets were released
[14:48:05] <Bartzy> I have some design question. i.e. is mongo is a good fit, before I dive in into it. Hope it's fine to ask that kind of questions here.
[14:49:05] <Bartzy> I'm part of a team that is developing a pinterest-like feature from its application photos (user-created photos). Each "pin" has likes, views, and comments
[14:49:33] <Bartzy> each pin has view counts, and if you click on it, it also shows the viewers list, a comments count, and the last 3 comments (if you click you get all of them).. same for likes
[14:50:32] <Bartzy> We actually have 500 million "pins" (photos), but only 150 million of them are shared with the world, and those are the ones we will show on the pinterest-like feature.
[14:56:54] <stevie-bash> Did I miconfigures my repset if the local.oplog.rs has read performance of 1767ms in mongotop?
[15:03:04] <Bartzy> NodeX: So when should one use a RDBMS such as MySQL instead of something like Mongo? Only for transactional stuff like financing and payments?
[15:41:05] <Baribal> Hi. We're having a strange problem with pymongo: pymongo.son suddenly just *vanished*. Anybody ever heard of that?
[15:46:54] <NodeX> one I had to reboot with a 10tb raid that had never fsck'd, it took hours to come back up lolol
[15:56:17] <devastor> Hi all, I got some "DR102 too much data written uncommitted" warnings and backtraces after initial sync when it was doing the replSet initialSyncOplogApplication stuff in mongo 2.0.4. It continued and completed ok after those, though. Is there a risk that some data didn't get written properly or anything like that?
[16:02:56] <Baribal> neil__g, I'd be all for it. I know a few buildings which have a rather simple and easily detectable geometric shape, and whichs bombardment would further the cause of open sources.
[16:06:17] <JoeyJoeJo> I have a function that returns the value of a text field. When I call that function using the button next to the text field, I can get the value of the text field. If I call that function from another function I get the error "cannot get value of null". Why does that happen?
[16:54:36] <Bartzy> also, creating an index on an existing collections means locking that collection? the entire DB ?
[17:19:13] <tystr> I'm trying to wrap my head around map/reduce….so I have a noob question: I need to sum some numbers in two differen collections then divide one sum by the other…
[17:32:18] <kali> tystr: you perform both sum() separately, and then perform the division in your application code
[17:34:31] <kali> tystr: that said you can use a group instead of a map reduce, it is slightly more compact
[18:33:22] <FerchoDB> I don't know why it is not working, it is returning all the documents in collection. (I'm doing tests with MongoVUE)
[18:33:51] <FerchoDB> and if i do the "query" on its own, it returns only the filtered objects. Maybe it's something with MongoVUE
[19:22:32] <JoeyJoeJo> In pymongo I want to find all documents in a collection but db.collection.find() with no arguments returns an error. What am I doing wrong?
[19:47:34] <halcyon918> if I have an index defined like {accountId: 1, status: 1} and I look up a record but JUST "accountId" (ignoring status for now), will the index still be used, or should I have a second index for just accountId?
[19:49:26] <elarson> I'm trying to fix a race condition in an application and I'm considering using find_and_modify in the pymongo driver rather than using update. Is this a horrible idea? I'm concerned this will cause more locking or something similar that could slow down other operations
[19:51:08] <elarson> the docs don't suggest that findAndModify adds any extra locking, specifically the global lock, but I'm curious if that is really the case
[19:53:47] <Bartzy> also, creating an index on an existing collections means locking that collection? the entire DB ?
[21:36:44] <JoeyJoeJo> I'm trying to do a geospacial index for a large collection. I did db.collection.ensureIndex({'loc':'2d'}) and it's just sitting there with three periods instead of a prompt. Does that mean it's working?
[22:39:08] <dstorrs> how do people normally manage backups? We're going to have something like 5-6TB to start with, and we're adding about 10G per day
[22:40:04] <dstorrs> I'm running into issues like "S3 / CloudFiles / etc can only store 5G per object"
[22:48:25] <esad> halp! I'm getting Invalid BSONObj size: 1601332595 (0x7365725F) first element: gent: ?type=97
[23:27:29] <cgriego> Is there anywhere with download links to older copies of the deb packages? Such as 2.0.2?