pmxbot IRC Log Viewer

[00:01:50] <joannac> That doesn't make sense. Primary preferred means it will read from the primary if possible

[00:02:06] <joannac> so you'll be reading from the primary anyways?

[00:04:39] <joannac> I guess if you lose primary, you will still get reads, though maybe from out-of-date data

[00:04:48] <joannac> so if that's suitable, then sure

[01:50:35] <ehershey> /last -pub

[01:50:38] <ehershey> oops!

[03:44:18] <shkx> Hey guys, ive run int oan issue where I need to import a bunch of data from one cluster (lets say in our prod environment) to another (lets say dev).. There maybe some duplicate data like primary keys and etc. I had alook at the syntax for mongorestore and it states that if there is a duplicate record, it will skip that and continue, and will only perform an INSERT operation, not UPDATE. I also noticed you can use the

[03:44:35] <shkx> sorry, --drop, not --flag

[03:44:48] <shkx> my question is if you are using just mongorestore so that it will skip the duplicate info, does it inform you which data was skipped?

[03:45:28] <shkx> im writing a python script atm to parse the BSON data and check for any duplicate data but its probably just going to take a while to do that and id prefer it greatly if there is a way within Mongo to check for duplicate data

[04:05:52] <joannac> shkx: I don't think so, unless you check the logs

[04:31:40] <movedx> Hey guys. How can I go about filling a MongoDB installation with some example data? I need to do this from the CLI, without having to go into a shell. I'm doing this from Ansible.

[04:31:54] <movedx> I can write and push a shell script, etc.

[04:32:20] <shkx> joannac: no problem

[04:33:58] <joannac> movedx: erm, I don't understand the question. write a js script that loads data and use --eval?

[04:34:32] <movedx> joannac: Sure, I guess that works. I could write something in Python too... OK, I'll do that. Sorry for wasting your time - I should have given it more thought.

[04:36:50] <joannac> movedx: no problems

[05:35:08] <shkx> joannac: is it possible when using mongorestore to specify a new collection name

[05:35:20] <shkx> or would i just rename the files

[05:35:30] <shkx> before restore

[05:41:00] <joannac> shkx: http://docs.mongodb.org/v2.6/reference/program/mongorestore/#cmdoption--collection

[05:43:26] <shkx> thanks

[05:45:29] <shkx> joannac: specifying the full path when doing mongorestore doesnt seem to work

[05:50:33] <joannac> shkx: pastebin an example?

[05:51:23] <shkx> okay

[05:51:59] <shkx> joannac: one sec

[05:55:07] <shkx> joannac: http://pastie.org/10371721

[05:55:47] <joannac> shkx: erm

[05:56:40] <shkx> sorry i tried a few things as you can see :) i am not a mongo expert

[05:56:55] <joannac> 1. you should be obfuscating passwords. 2. --collection takes a collection name 3. the path should be the last argument

[05:57:47] <joannac> the last one you have is the closest. Instead of "--collection..." have "-c AutoPaymentInfo-bulk /mongo-backups/new/......"

[05:57:50] <shkx> obfuscated passwords

[05:57:55] <shkx> i figured it out thank you joannac

[05:58:00] <shkx> i will look into this obfuscation

[06:05:28] <shkx> joannac:

[06:05:30] <shkx> do you have any links

[06:05:36] <shkx> for obfuscated passwords

[06:09:38] <joannac> shkx: all i mean is, when I ask for an example, you should not share the password

[06:09:55] <joannac> i.e. instead of copying and pasting, replace the password with XXXX

[06:10:03] <shkx> oh

[06:10:13] <shkx> that wasnt the real password :) and you have no ip address...

[06:11:05] <joannac> oh. okay then

[06:11:10] <joannac> it looked like a password :)

[06:16:19] <shkx> lol yeah

[06:16:25] <shkx> thanks joannac :) you are really helpful

[06:19:44] <joannac> :) you're welcome shkx

[10:41:20] <yopp> hum

[10:41:29] <yopp> with this balancer thing

[10:42:00] <yopp> after 2.4 -> 3.0 updgrade some mapreduces are failed, because there a migration lock on sharded collection

[14:03:47] <atomicb0mb> hello ! Is there any document where I can read about the difference between upsert, update, and save?

[14:17:09] <sabrehagen> how can i monitor to see if mongodb is paging to disk?

[14:37:02] <danijoo> if I do `db.collection.remove()`, will it block the whole database or allow me to still write new entries while old ones are deleted?

[14:37:38] <danijoo> (I want to remove all entries of a pretty large database in production)

[14:42:03] <saml> danijoo, why not just drop collection?

[14:42:47] <saml> how big is your large database in production? 2GB?

[14:43:03] <danijoo> saml, ~50gb

[14:43:13] <saml> drop will take <1 sec

[14:43:27] <saml> remove is really expensive

[14:43:39] <saml> especially if you ahve lots of indexes

[14:43:40] <danijoo> yeah i think you are right

[14:43:46] <danijoo> i came up with that right after asking

[14:43:55] <danijoo> ill make a script for drop + immediatly recreate

[14:44:52] <saml> i hope your apps are written resilient to temporary inconsistency. or with eventual consistency of mongodb in mind

[14:46:26] <danijoo> yeah thats not a problem

[14:47:00] <danijoo> the db is more like a giant cache to limit queries against a 3rd party api

[14:50:44] <Sna4x8> When querying, why is it that the AND operations can optionally be an object--{name: 'foo', age: 2}--while the OR must be an array--{$or: [{<expression>},..{<expression>}]}? Is it just a short-hand convenience?

[14:51:34] <Sna4x8> That is, why can't one find({$or: {name: 'foo', age: 2}})?

[14:51:49] <fontanon> Hi! Is this the best irc channel to ask about nodejs mongoose questions ?

[14:55:00] <doc_tuna> fontanon: no, for mongoose, go to mongoose

[14:55:15] <fontanon> #mongoose ?

[14:55:20] <doc_tuna> people here use mongodb directly

[14:55:31] <doc_tuna> *cough* which is the way you should use it too *cough*

[14:55:48] <StephenLynx> lol mongoose

[14:56:21] <StephenLynx> yeah, most people here completely abhor mongoose. at least for the ones experienced with mongo.

[14:56:43] <StephenLynx> the only advice you will get is "don't use it"

[14:58:00] <fontanon> I respect your opinions, buts working good for me.

[14:58:24] <doc_tuna> try umm #mongoosejs?

[14:58:38] <fontanon> doc_tuna, empty

[14:58:39] <doc_tuna> not sure where their official chats are at

[14:59:15] <fontanon> doc_tuna, thnks I think it is

[14:59:24] <doc_tuna> gl

[15:02:32] <sabrehagen> for those who have mastered using the mongodb native node.js driver, do you any example code that is essentially the 'to-do' app for the driver, i.e. it does all common CRUD application tasks using the optimal design patterns?

[15:02:36] <saml> Sna4x8, {a:b,c:d} and $and:[{a:b}, {c:d}] might not be exactly the same since $and shortcircuits

[15:03:00] <saml> for {a:b,c:d}, c==d might be checked first

[15:03:05] <saml> maybe not. i'm bullshitting

[15:03:21] <StephenLynx> sabrehagen that is already done by the driver.

[15:04:25] <StephenLynx> the only thing you need is the documentation http://mongodb.github.io/node-mongodb-native/2.0/api/

[15:04:46] <Sna4x8> saml: Thanks, and I agree with that, but with AND you can use either option. With the former the user doesn't care about the order of evaluation.

[15:05:44] <sabrehagen> what's the commonly accepted counterpart to mongoose's populate functionality using the driver?

[15:05:45] <Sna4x8> With OR it seems like one is required to use the array notation but, just like AND, most (all?) of the time the order is irrelevant.

[15:06:10] <StephenLynx> what is the populate?

[15:06:15] <sabrehagen> sorry badly explained

[15:07:29] <Sna4x8> saml: Futhermore, OR is short-circuit evaluated as well, right?

[15:07:32] <sabrehagen> for a set of documents, you frequently want to expand an object id to properties of the referenced document, the notion of 'populating' the parent document with referenced data.

[15:07:47] <StephenLynx> wot

[15:07:58] <sabrehagen> for an array of documents, how do you do this efficiently with the driver?

[15:08:11] <StephenLynx> an array of documents is called a collection

[15:08:22] <StephenLynx> and I don't have to do anything about it

[15:08:36] <saml> Sna4x8, i don't know. maybe $or does things in parallel and accepts first check

[15:12:10] <sabrehagen> i was hoping you might have grown up if somebody asked you a serious question, but it looks like you don't seek to teach, rather you seek ostracise those coming from a simpler system.

[15:12:35] <StephenLynx> what

[15:12:45] <StephenLynx> I gave you a serious answer

[15:12:52] <StephenLynx> I didn't meant to be aggressive at all.

[15:13:06] <sabrehagen> seriously? that's your attempt to help?

[15:13:30] <StephenLynx> what did you expect?

[15:13:39] <sabrehagen> too much of you.

[15:13:44] <StephenLynx> lel

[15:14:01] <Sna4x8> Similar question: Is there ever a time where the order of AND or OR operations matters when querying? (Obviously, disregarding mixing the two operators.)

[15:14:22] <Derick> Sna4x8: i don't think so, unless you nest

[15:14:29] <diegoaguilar> Hello, I connect to a mongolab database from mongoshell giving user and password

[15:14:39] <diegoaguilar> one the same shell I try to "access" to a local database too

[15:14:46] <Sna4x8> Derick: That's what I thought, but I wasn't sure if I was missing a corner case or something.

[15:14:58] <diegoaguilar> with connect function and new Mongo.getDB

[15:15:11] <diegoaguilar> but that won't let me normally use local database (its giving permissions problems)

[15:15:30] <diegoaguilar> how can I access multiple databases

[15:15:33] <diegoaguilar> in same shell?

[15:15:44] <diegoaguilar> with its proper permissions for each?

[15:15:46] <Derick> are they on the same host?

[15:16:03] <diegoaguilar> nope

[15:16:04] <diegoaguilar> distinct

[15:16:13] <Derick> I didn't think the shell supported that.

[15:16:25] <Derick> I could be wrong though.

[15:17:33] <diegoaguilar> hmm

[15:17:46] <saml> new Mongo('localhost').getDB('yolo')

[15:17:59] <Derick> saml: wouldn't that lose the connection to the remote one?

[15:19:08] <diegoaguilar> saml, Derick

[15:19:16] <sabrehagen> can anybody in the channel shed some light on my populate question above? for those who aren't familiar with mongoose's populate: http://mongoosejs.com/docs/populate.html

[15:19:18] <diegoaguilar> I login to remote, it has credentials

[15:19:33] <saml> var l = new Mongo('localhost').getDB('yolo'); db is your remote. l is local

[15:19:44] <diegoaguilar> then I do new Mongo('localhost').getDB('yolo')

[15:19:52] <diegoaguilar> l would be local

[15:20:00] <diegoaguilar> and then I cant do a show collections on local

[15:20:11] <diegoaguilar> at it actually thinks upon the remotes localhost I guess

[15:20:14] <StephenLynx> ah, that? relations?

[15:20:18] <saml> l.getCollectionInfos()

[15:20:23] <saml> l.getCollectionNames()

[15:20:25] <diegoaguilar> btw the remote is Mongolab

[15:20:33] <StephenLynx> you just handle that in your application using fake relations.

[15:20:38] <StephenLynx> sabrehagen

[15:20:59] <StephenLynx> there is not tool for that.

[15:21:22] <StephenLynx> you just put a value that references a different document in a different collection and work knowing that.

[15:21:33] <sabrehagen> so you have any sample apps written in node that demonstrate that? e.g. an archetypal to-do app?

[15:21:48] <StephenLynx> I have multiple systems that does that.

[15:21:56] <StephenLynx> but there isn't a boilerplate to that.

[15:22:08] <StephenLynx> gitlab.com/mrseth/LynxChan

[15:22:41] <StephenLynx> I just have my boards having a field called boardUri and then threads belonging to that board also have a field called boardUri with the value of the board they belong to.

[15:23:08] <saml> sabrehagen, you're looking for a todo app example using node.js and mongodb?

[15:23:14] <saml> there should be plenty

[15:23:31] <StephenLynx> https://gitlab.com/mrseth/LynxChan/blob/master/doc/Model.txt

[15:23:36] <StephenLynx> my database model

[15:23:49] <StephenLynx> all fields marked with * are fake relations.

[15:24:06] <saml> but being node.js, it won't be simple. they will probably involve angular.js, react.js, gulp, grunt, expressjs, and other crap

[15:24:25] <StephenLynx> wrong

[15:24:38] <StephenLynx> I mean

[15:24:51] <StephenLynx> the community is dumb AF and usually use all that crap.

[15:24:55] <StephenLynx> but you don't have to.

[15:25:11] <StephenLynx> the driver documentation provides examples to everything using vanilla.

[15:25:26] <StephenLynx> it just have a bad habit of nesting callbacks, but for examples I think its passable.

[15:32:55] <doc_tuna> i have a node codebase that is express.js with some middleware and a swagger interface

[15:33:05] <doc_tuna> for building apis in

[15:34:21] <sabrehagen> StephenLynx: that codebase is fascinating.

[15:34:28] <StephenLynx> really?

[15:35:07] <sabrehagen> it looks like the code of a javascript veteran

[15:35:14] <StephenLynx> actually

[15:35:25] <StephenLynx> I have started using js for about a year only

[15:35:49] <sabrehagen> what did you write in previously?

[15:35:55] <StephenLynx> java, unrealscript

[15:36:09] <sabrehagen> how long have you coded for?

[15:36:11] <StephenLynx> obj-C

[15:36:14] <StephenLynx> 9 years I guess

[15:36:25] <sabrehagen> interesting.

[15:36:32] <StephenLynx> people are usually polarized about my code.

[15:36:40] <StephenLynx> either they think its great or that its shit :v

[15:36:53] <StephenLynx> I have worked towards improving code legibility this weekend

[15:36:58] <StephenLynx> splitting large files into multiple parts

[15:37:11] <StephenLynx> I had a file with over 2 thousand LOC

[15:37:20] <StephenLynx> I broke it into 6 parts now

[15:37:24] <sabrehagen> what module are you most proud of at the moment?

[15:37:39] <StephenLynx> a module in what sense?

[15:37:45] <sabrehagen> in the whole system. i.e. what has the most elegance as a result of abstraction?

[15:37:54] <StephenLynx> hm

[15:37:58] <StephenLynx> the front-end

[15:38:04] <sabrehagen> really.

[15:38:05] <StephenLynx> the way it handles the front-end project.

[15:38:08] <StephenLynx> you see

[15:38:14] <sabrehagen> that's a surprise. i thought it would have been back end for sure.

[15:38:16] <StephenLynx> this system it has to use HTML templates

[15:38:19] <StephenLynx> oh, its the back-end

[15:38:32] <StephenLynx> take a seat and let me explain how the engine deals with the front-end

[15:38:39] <StephenLynx> if you see on lynxhub.com

[15:38:49] <StephenLynx> you will see this project is an engine that runs a site.

[15:39:06] <StephenLynx> but the site doesn't require javascript, the engine generates the html

[15:39:27] <sabrehagen> right.

[15:39:29] <StephenLynx> so I need to find templates, fill in stuff and save a webpage.

[15:39:41] <StephenLynx> if you take a look into the engine, templates are nowhere to be seen.

[15:39:56] <StephenLynx> but you will notice in the readme, it mentions front-end projects

[15:40:19] <StephenLynx> you specify a path to the front-end, it looks for a json file in the directory

[15:40:27] <StephenLynx> and this json file indicates where is each of the templates.

[15:40:52] <StephenLynx> so you could completely change the layout of the site without even bringing it down.

[15:41:15] <StephenLynx> you just change the path to the front-end and the engine will reload the pages and start using the new front-end files.

[15:41:29] <StephenLynx> there is even a GUI to change settings that include the path to the front-end.

[15:41:54] <sabrehagen> cool. sounds fun. are you purposely running without javascript?

[15:41:58] <StephenLynx> yes.

[15:42:03] <StephenLynx> demographics.

[15:42:14] <sabrehagen> what do you mean?

[15:42:25] <StephenLynx> the people I am trying to appeal to would chew my liver if I were to require javascript for the site to be used.

[15:42:49] <nignog> Halo

[15:42:52] <StephenLynx> wat

[15:42:59] <StephenLynx> that fps?

[15:43:16] <StephenLynx> you can even post without cookies too.

[15:43:30] <StephenLynx> but its easier to post with cookies though

[15:43:34] <sabrehagen> you have to be careful, you're getting a little to modern with your design as it is.

[15:43:38] <nignog> Can I ask something about "populate" from mongoose orm

[15:43:41] <StephenLynx> no

[15:43:53] <StephenLynx> I mean, you can. but no one will answer.

[15:44:00] <StephenLynx> too modern?

[15:44:02] <nignog> Why is that?

[15:44:03] <StephenLynx> what do you mean?

[15:44:11] <StephenLynx> because people here abhor mongoose.

[15:44:17] <StephenLynx> specially me.

[15:44:20] <StephenLynx> but you can try.

[15:44:48] <StephenLynx> and afaik, mongoose is an ODM

[15:45:05] <StephenLynx> since it handles documents and not a relational database.

[15:46:22] <nignog> Well, it's likely about the structure of database itself

[15:47:24] <sabrehagen> how many chans are you part of?

[15:48:14] <StephenLynx> none.

[15:48:32] <StephenLynx> I am just running an instance of the engine so people can see it running.

[15:48:48] <sabrehagen> you're the writer of the software, no?

[15:48:51] <StephenLynx> yes.

[15:49:48] <sabrehagen> so surely your desire to build a chan from scratch came from being an avid user of chans?

[15:50:01] <duck_tape> anybody in here use ycsb to benchmark mongo ?

[15:50:02] <StephenLynx> oh, you mean user?

[15:50:08] <sabrehagen> yes

[15:50:09] <StephenLynx> I thought you mean as in on staff of a chan.

[15:50:14] <StephenLynx> I have used many.

[15:50:24] <sabrehagen> enough to feel as though you know the community?

[15:50:30] <StephenLynx> pretty much.

[15:52:02] <StephenLynx> on lynxhub I put my motivations on the about page

[15:52:18] <StephenLynx> http://static.lynxhub.com/about.html

[15:53:10] <sabrehagen> imagine you wanted to serve an api to be consumed by a web app as an alternative to the html version. do you have an internal api layer you could hook that into?

[15:53:13] <sabrehagen> yes, i read that.

[15:53:18] <StephenLynx> yes

[15:53:37] <StephenLynx> lynxhub.com/lynxchan/res/1.json

[15:53:38] <sabrehagen> i would imagine so

[15:53:51] <StephenLynx> and the controls also can be used by sending json parameters

[15:54:04] <StephenLynx> the only parts I am yet to make usable by json are the dynamic pages

[15:54:06] <sabrehagen> nice.

[15:54:12] <StephenLynx> but that is on the roadmap

[15:54:28] <StephenLynx> is just not a priority

[15:55:04] <StephenLynx> in the docs there is the specification of both the json api controls and the static json data.

[15:56:22] <sabrehagen> i started with the MEAN stack in december 2014 to build a web app. i had never done any web service nor web app development. i could tell having my system be dependent on mongoose was something i didn't want to have, but it was the only immediate way to get going.

[15:56:28] <StephenLynx> ugh

[15:56:37] <StephenLynx> MEAN is pretty much

[15:56:45] <StephenLynx> radioactive aids cancer

[15:56:58] <sabrehagen> i've written an internal api behind which mongoose is abstracted, meaning i could convert the access to the nodejs driver exclusively, or even another database entirely.

[15:57:27] <StephenLynx> yeah, I don't bother with that. my code depends on the native driver and thats it.

[15:57:46] <StephenLynx> I mean

[15:58:06] <StephenLynx> it could be changed on the db.js file so it would interpret the commands to be run somewhere else.

[15:58:28] <StephenLynx> but I don't intend to use anything else than I am using.

[15:59:21] <StephenLynx> some dude bitched at me for not abstracting the database and I had to explain him there is no other database that provides the tool I use from mongo

[15:59:39] <StephenLynx> the tools*

[15:59:57] <StephenLynx> the problem is that he really thought I was obliged to take his input

[16:00:25] <sabrehagen> have you run load tests?

[16:00:44] <StephenLynx> sort of.

[16:01:08] <StephenLynx> I have ran a benchmark where I would load a static html 1000 times with 200 or 50 concurrent requests

[16:01:10] <StephenLynx> 2ms per request

[16:01:13] <diegoaguilar> Hello, how can I tell MongoDB to write databases to a particular mount point?

[16:01:40] <StephenLynx> 50 to 70ms per thread reply without images

[16:02:26] <StephenLynx> it only broke because of an issue with the driver that I had to circumvent.

[16:02:32] <StephenLynx> but fixed that last week

[16:04:30] <sabrehagen> so that's 200 requests concurrently with 70ms response time. that's the kind of beauty you'd hope for from node and mongo together in their rawest form.

[16:04:56] <sabrehagen> i'd like to see how many concurrent users it would support before it exceeded 200ms.

[16:04:59] <StephenLynx> and that is for posting

[16:05:12] <StephenLynx> where it has to perform lots and lots of operations and checks

[16:05:42] <StephenLynx> and on a SSD

[16:05:44] <StephenLynx> no

[16:05:46] <StephenLynx> hdd*

[16:05:58] <StephenLynx> on a VM

[16:06:33] <Sna4x8> Shoot, with that number of requests (only 200 concurrent) we get comparable (70ms) performance using SQL. With a high level of abstraction, I might add.

[16:07:30] <StephenLynx> yeah, but I don't know what kind of operation you are performing.

[16:07:46] <StephenLynx> on readings I can serve a page in 2ms

[16:09:27] <sabrehagen> StephenLynx: do you consider a database call almost inconsequential? that is, it's just like a direct access memory operation on a json object?

[16:09:42] <StephenLynx> that depends on the operation.

[16:09:59] <StephenLynx> if its a small collection that is accessed frequently

[16:10:05] <StephenLynx> then the overhead is negligible

[16:10:35] <sabrehagen> what's an example of a collection size at each end of the spectrum and the type of query you might run on it?

[16:10:53] <StephenLynx> no idea.

[16:11:20] <StephenLynx> I don't have much experience running large and active databases.

[16:11:32] <StephenLynx> I just know from others that skips are slow

[16:11:53] <StephenLynx> and it becomes slower for the more you skip

[16:12:05] <sabrehagen> what method do you use to page through your data then?

[16:12:14] <StephenLynx> well, since it is a chan

[16:12:19] <StephenLynx> content is deleted often

[16:12:21] <StephenLynx> like threads

[16:12:41] <StephenLynx> so a board will never more than a couple hundred of threads

[16:12:54] <StephenLynx> so skipping through the results is not an issue

[16:12:57] <StephenLynx> not to mention

[16:13:04] <StephenLynx> that pages are generated on a queue

[16:13:07] <StephenLynx> and not on demand

[16:13:20] <StephenLynx> when inside a thread I don't have to skip or limit anyway

[16:13:31] <sabrehagen> you should do a writeup on your architecture.

[16:13:38] <StephenLynx> so skipping is not an issue for this system.

[16:13:40] <StephenLynx> however

[16:13:43] <StephenLynx> on the logs page

[16:13:48] <sabrehagen> a full blog post discussing your design choices would be fascinating.

[16:13:58] <StephenLynx> I intend to remove the skipping

[16:14:10] <StephenLynx> because the logs might achieve numbers.

[16:14:14] <StephenLynx> and they are not deleted

[16:14:22] <StephenLynx> I could do that.

[16:14:23] <StephenLynx> but

[16:14:33] <StephenLynx> I don't care, you know?

[16:14:46] <StephenLynx> I am not connected with the community.

[16:14:54] <StephenLynx> specially the js/node/web community.

[16:15:37] <sabrehagen> just document how all the components of the system interact. the software deserves it.

[16:15:44] <StephenLynx> that is already done.

[16:15:58] <StephenLynx> doc/Engine.txt

[16:16:22] <sabrehagen> just found that. it will be an interesting read.

[16:16:23] <StephenLynx> I just haven't updated it after I wrote, but the architecture haven't changed either.

[16:16:44] <sabrehagen> whs this written as a plan before you began coding?

[16:16:48] <StephenLynx> no

[16:16:49] <StephenLynx> after.

[16:16:59] <StephenLynx> I didn't planned ahead.

[16:17:08] <StephenLynx> aside from very few stuff

[16:17:23] <StephenLynx> like saving files on gridfs, having a queue to generate pages

[16:17:39] <StephenLynx> I did refactor a lot though.

[16:18:08] <StephenLynx> some parts became very messy with this organic growth, like the upload handling logic.

[16:18:15] <StephenLynx> and I realized a pattern too.

[16:18:25] <StephenLynx> the asynchronous recursion pattern.

[16:18:39] <StephenLynx> and applied it to the upload handling logic.

[16:19:24] <StephenLynx> if you have interest in the project, I got a channel on rizon.irg

[16:19:28] <StephenLynx> .irc*

[16:19:30] <StephenLynx> #lynxchan

[16:20:24] <sabrehagen> will keep it in mind.

[16:21:52] <sabrehagen> StephenLynx: do you have any cases where you have issues with concurrent updates to the same document?

[16:21:57] <StephenLynx> no

[16:22:04] <StephenLynx> afaik mongo handles that.

[16:22:16] <StephenLynx> and I don't have documents that are updated often anyway

[16:22:46] <sabrehagen> can two people ever edit the same document?

[16:22:56] <StephenLynx> threads are updated on every reply, boards are updated on every post, but it doesn't keep the operation from completing

[16:22:57] <StephenLynx> yes.

[16:22:59] <StephenLynx> they could.

[16:23:12] <StephenLynx> two mods could edit the same post.

[16:23:32] <sabrehagen> how do you handle the case of the first mod's save getting overwritten by the second mod's value?

[16:23:36] <StephenLynx> I don't.

[16:23:55] <StephenLynx> the value is completely overwritten by the second mod.

[16:24:21] <saml> implement versionnig

[16:24:24] <StephenLynx> no

[16:24:33] <saml> for cms, you need versioning

[16:24:38] <StephenLynx> is not an cms

[16:24:40] <StephenLynx> a cms*

[16:24:51] <saml> cms solves all business problems

[16:24:58] <StephenLynx> it doesn't solve mine.

[16:25:51] <StephenLynx> the post will say the name of the person that edited it the last time and the time of the edit.

[16:25:55] <StephenLynx> that is enough.

[16:27:34] <sabrehagen> what do you do by day?

[16:27:42] <StephenLynx> my job?

[16:27:51] <StephenLynx> software developer

[16:28:04] <StephenLynx> but lately I have been working on lynxchan because there is nothing for me to do

[16:28:20] <StephenLynx> I am the guy that gets shit done that no one else could.

[16:28:39] <sabrehagen> what software do you write for work?

[16:28:56] <StephenLynx> a real-time chat back-end, a C software that reads input from a serial port connected to lab equipment

[16:29:06] <StephenLynx> there was an android app for a convention when I got in too

[16:29:25] <StephenLynx> these are the 3 main projects I developed here in a little more than a year

[16:29:34] <sabrehagen> how many devs are there?

[16:29:37] <StephenLynx> dunno

[16:29:48] <StephenLynx> I am the only one at this office, but in the main office there are more

[16:29:48] <sabrehagen> 3? 30? 300? 3000?

[16:29:49] <StephenLynx> 4 or 8

[16:30:16] <StephenLynx> i more stuff than that, but small stuff

[16:30:21] <StephenLynx> I did more*

[16:30:28] <StephenLynx> like some web fe work on a system

[16:30:35] <StephenLynx> but I absolutely loathed

[16:30:40] <sabrehagen> are there any other good developers there?

[16:30:45] <StephenLynx> I don't think so.

[16:30:48] <StephenLynx> not from what I seen

[16:31:07] <StephenLynx> but I might have not seen something yet.

[16:31:11] <sabrehagen> are you the youngest?

[16:31:15] <StephenLynx> I think so.

[16:31:19] <StephenLynx> 26

[16:31:45] <StephenLynx> the guys I have seen are not the worse I have seen either.

[16:31:57] <sabrehagen> are you challenged by the job?

[16:32:04] <StephenLynx> somewhat.

[16:32:16] <StephenLynx> it was quite a challenge writing that C software

[16:32:25] <sabrehagen> in what respect?

[16:32:29] <StephenLynx> and a mild challenge to write the real time chat.

[16:32:42] <StephenLynx> the android app was just a matter of time restraint

[16:32:50] <StephenLynx> as in most small business

[16:32:57] <StephenLynx> you don't know what you will have to develop.

[16:33:19] <StephenLynx> so you are always having to learn something new and doing something you never did before

[16:34:36] <StephenLynx> the C software was mostly challenging because I had no idea on any of the involved aspects. I had no idea how raspPis were, I had no idea what was the interface the equipment used, I had no idea on how to use serial ports

[16:34:51] <StephenLynx> I just had some C++ knowledge, so learning what I had to learn from C was not hard

[16:35:29] <sabrehagen> tgis was the serial port comms?

[16:35:36] <StephenLynx> no

[16:35:38] <sabrehagen> *this

[16:35:40] <StephenLynx> ttyAMA0

[16:35:43] <StephenLynx> ah

[16:35:44] <StephenLynx> yes

[16:36:20] <StephenLynx> but in the end everything turned out great and I believe I secured my job on that project.

[16:36:32] <StephenLynx> I suggested him not using C, but my boss wanted C so

[16:36:48] <StephenLynx> now I don't think anyone else can maintain it because it is in C

[16:36:54] <StephenLynx> but oh well

[16:37:12] <StephenLynx> ¯\_(ツ)_/¯

[16:37:26] <sabrehagen> where was the software to be used?

[16:37:50] <StephenLynx> in raspPis connected to lab equipment to fetch the medical results

[16:38:42] <StephenLynx> it will run as a service, saving the content in files and sending them to feed the system

[16:39:07] <StephenLynx> so doctors will receive the results ASAP through a mobile app

[16:40:28] <sabrehagen> what language would you have used?

[16:40:56] <StephenLynx> python, because of the raspPi interface

[16:41:10] <StephenLynx> or so I heard it uses python

[16:42:14] <sabrehagen> with you experience in node now, would you choose to use node on the pi?

[16:42:47] <StephenLynx> no

[16:42:59] <StephenLynx> it takes about

[16:43:06] <StephenLynx> 15 minutes to compile on an i5

[16:43:11] <StephenLynx> and 9bm RAM idle

[16:43:42] <StephenLynx> about io.js

[16:43:52] <StephenLynx> haven't used node since i dunno, 0.10

[16:44:05] <StephenLynx> but I don't think it is much better by now

[16:45:02] <sabrehagen> that sounds like a nightmare.

[16:45:17] <StephenLynx> yeah, imagine compiling that on a raspPi

[16:45:44] <StephenLynx> I even considered at some point trying, but I didn't like the idea because of that.

[16:46:27] <StephenLynx> besides, if you are just going to use node/io as a general purpose scripting runtime environment, you might as well use something else meant to be used for that.

[16:46:55] <StephenLynx> you wouldn't use the single strongest point of it: handling high levels of concurrency.

[16:47:12] <StephenLynx> I tried bash

[16:47:34] <StephenLynx> but its extremely high-level behavior regarding serial ports screwed up everything

[16:47:35] <sabrehagen> how did that go?

[16:47:40] <sabrehagen> right.

[16:47:57] <Derick> try compiling MongoDB ;-)

[16:48:11] <StephenLynx> I believe node/io would fit the part where you have low-level access.

[16:48:24] <StephenLynx> but you would have to build from source

[16:49:07] <Derick> ehershey: you alright? :)

[16:49:35] <ehershey> fighting with irssi :)

[16:49:42] <ehershey> it's a losing battle

[16:49:45] <ehershey> 3.0.6 is out!

[16:49:50] <Derick> yay

[17:29:29] <Lujeni> Hello - There is a step to propose a new Third-Party Open-Source Tools on mongodb documentation ? Thx

[17:30:50] <StephenLynx> I don't think 3.0.6 is out on centOS repository from 10gen

[17:30:54] <StephenLynx> welp

[17:32:17] <christo_m> hello, im trying to do an autocomplete on movies and shows, should i be storing these as separate collections in my mongodb?

[17:32:27] <christo_m> would it be slower querying both collections?

[19:28:05] <Pinkamena_D> I am looking for an idea to solve this problem. I have a collection with docs with array values: [a, b, c, d]. I have an array of values and I want to query for documents where all of the document's array values are a match but there can be extra. For example [a,b,c] and [a,b,d,e,f,g] should not match but [a,b,c,d] or [a,b,c,d,e] will match.

[19:31:50] <StephenLynx> ok, so what you need is "give me all that contains AT LEAST all of these elements?"

[19:32:51] <doc_tuna> ehershey: irssi is the client of the future

[19:32:59] <doc_tuna> i'll be here, in the future, waiting for you

[19:34:42] <ehershey> the problem is it's also the client of the past

[19:34:50] <ehershey> so I'm stuck in the past

[19:34:55] <ehershey> maybe someday we'll be able to meet

[19:34:59] <ehershey> I guess this was it happening right now

[19:38:53] <Pinkamena_D> StephenLynx, yes

[19:39:53] <StephenLynx> http://docs.mongodb.org/manual/reference/operator/query/elemMatch/#op._S_elemMatch

[19:40:13] <StephenLynx> I think $elemMatch:{$in:[a,b]} would work

[19:40:18] <StephenLynx> wait

[19:40:21] <StephenLynx> I don't think so

[19:41:43] <StephenLynx> http://docs.mongodb.org/v3.0/reference/operator/query/all/

[19:41:44] <StephenLynx> here

[19:42:09] <StephenLynx> $all would do exactly what you need

[19:42:48] <Pinkamena_D> No, not quite, I guess I misspoke

[19:43:09] <Pinkamena_D> The "all" part needs to be the items in the document, not the items in the query.

[19:43:33] <Pinkamena_D> All of the items in the document need to match something in the query, but the query can have extra items in it

[19:43:49] <Pinkamena_D> I guess it is the opposite of normal usage.

[19:46:14] <StephenLynx> ah

[19:47:03] <StephenLynx> then I guess you need elemMatch

[19:47:32] <StephenLynx> it will try to do that, make sure every element matches something in the array

[20:17:26] <deathanchor> the forever recovering curse....

[20:17:37] <deathanchor> which it will never recover from.

[20:19:46] <Pinkamena_D> StephenLynx, thanks, I got it going with $elemMatch combined with $in.

[20:45:38] <daidoji> anyone got any one-liner queries I can use to find all unique keys in all documents containing a string?

[20:45:48] <daidoji> like all keys containing string 'foo' in key name?

[20:46:54] <StephenLynx> db.col.find({foo:{$exists:true}})

[20:47:11] <StephenLynx> wait

[20:47:20] <StephenLynx> I don't think you can do that.

[21:03:41] <synthmeat> heya, a quick question.. it's mongoose, but i doubt that's relevant... http://pastie.org/private/uwn3l9ohodlirdg4dvsoa

[21:04:04] <synthmeat> is there any way to get this not to overwrite all other { platform: { whatever } }?

[21:04:31] <synthmeat> this way, i successfully update 'facebook' bit of the object tree, but it deletes anything else in 'platform' beside it

[21:14:08] <synthmeat> nvm, got it :)

[22:06:58] <silasbarta> So … on mongo v3, how do you dump all collections in a database?

[22:07:11] <diegoaguilar> you should do a script

[22:07:29] <silasbarta> so I can’t just do mongodump?

[22:07:58] <cheeser> mongodump works just fine

[22:08:01] <silasbarta> I had it set up that way in 2.4 where a user had readWrite and userAdmin roles, and it worked

[22:08:02] <diegoaguilar> mongodump is per database

[22:08:06] <diegoaguilar> ohh u said all collections

[22:08:07] <diegoaguilar> sorry

[22:08:13] <diegoaguilar> I though database

[22:08:18] <diegoaguilar> yeah mongodump is fine

[22:08:21] <cheeser> you can dump all collections or specific collections

[22:08:34] <silasbarta> but I upgraded to 3, tried to dump the database, and it says I don’t have permissions

[22:08:38] <silasbarta> it worked perfectly fine before

[22:09:20] <silasbarta> the command was `mongodump --port='8501' --username 'useradmin' --password '<password>' --db 'central' --out outputfile`

[22:09:46] <silasbarta> that gave me all the collections in `central`. User roles were readWrite, userAdmin

[22:09:59] <silasbarta> Now the mongo db is upgraded to 2.4 and it says I’m not authorized

[22:10:20] <silasbarta> er, I mean upgraded from 2.4 to 3.0

[22:11:23] <silasbarta> Anyone know what’s going on?

[22:13:34] <silasbarta> Did anyone see my questions about dumping a DB?

[22:14:12] <diegoaguilar> silasbarta, dont u have to use --authenticationDatabase

[22:14:12] <diegoaguilar> ?

[22:14:46] <silasbarta> ah, that would make sense … to have one db for auth and the other for what database I actually want to dump

[22:15:33] <silasbarta> wait, that looks like it’s just for when I’m using an external authentication deal

[22:15:51] <silasbarta> but it looks like —excludeCollection is what I want so I don’t dump the user and so don’t need higher permissions

[22:15:56] <silasbarta> let me give that a twirl

[22:16:28] <silasbarta> that means “test it” for those of who you don’t know common English idioms or dislike them

[22:44:36] <silasbarta> hm, looks like the —excludeCollection is the way to go

Log file Viewer

Help | Karma | Search:

#mongodb logs for Monday the 24th of August, 2015