pmxbot IRC Log Viewer

[00:03:02] <Havalx> !

[00:43:40] <MacWinner> is it possible for 2 different perl processes on the same server to generate the same MongoID? multicore machine

[00:44:13] <MacWinner> i mean ObjectId

[00:46:05] <Havalx> !

[00:47:46] <Havalx> References for Gen10's technology? I'ma read the documentation in a few moments.

[00:48:16] <Havalx> Any users here prefer sqlite3 over mongodb?

[00:51:12] <Boomtime> MacWinner: independent processes should not generate the same ObjectID according to the spec

[00:51:37] <kopasetik> Say i have a Mongo collection for Items and another collection for Categories. How do I reference a category (one category can have many items) in an item document?

[00:52:02] <cheeser> list the category IDs in your item docs

[00:52:02] <Boomtime> of the 12 bytes in ObjectID, 2 of them are intended to be a process ID that seperates processes on the same machine

[00:55:46] <kopasetik> @cheeser, basically embed as a sub-document?

[00:56:31] <cheeser> no. list the IDs in an array

[00:57:21] <MacWinner> cool, thank you!

[01:05:47] <kopasetik> @cheeser, thx

[01:36:41] <Havalx> with mongodb, is it possible to query/search the db for specific values and when conditions are met, return the object id/select the entire document given the id?

[01:37:44] <Havalx> I'm sure, yeah. I'ma read more documentation and memorize accepted commands.

[01:49:51] <bonhoeffer> how would i return the first element from a collection?

[01:50:05] <cheeser> db.collection.findOne()

[01:50:21] <cheeser> or

[01:50:27] <cheeser> db.collection.find().limit(1)

[01:50:44] <bonhoeffer> thanks!

[01:51:19] <bonhoeffer> ok, what if i wanted the keys from the result?

[01:57:51] <cheeser> you'd have the whole document...

[01:59:47] <bonhoeffer> i’m playing with variety

[02:04:40] <bonhoeffer> mongo handles dates well, correct? not just as strings.

[02:04:54] <cheeser> ISODate

[02:05:14] <bonhoeffer> thanks

[02:05:27] <cheeser> the types mongo supports: http://bsonspec.org/spec.html

[02:07:18] <bonhoeffer> thanks!

[02:09:32] <cheeser> np

[04:26:52] <Havalx> I'm using find_one() with a constructed JSON object.

[04:27:24] <Havalx> But it only returns the first entry in the db.

[04:27:59] <Boomtime> well, it found... one

[04:28:05] <Havalx> Yeah.

[04:28:40] <Havalx> Which it does; but I passed it values to search.

[04:28:50] <Havalx> Boomtime: Which it does; but I passed it values to search.

[04:28:54] <Boomtime> ok, so the issue is that it found the wrong thing?

[04:28:59] <Havalx> Yeah.

[04:29:13] <Boomtime> righto, can you paste your query?

[04:30:29] <Havalx> Boomtime: I fixed it.

[04:30:53] <Havalx> Boomtime: Thanks for the message!

[04:31:39] <Havalx> Boomtime: I'm update it in a min.

[04:31:49] <Boomtime> np, glad you found the issue

[04:37:38] <Havalx> Boomtime: I'm generating a document from variables; what is the formal syntax for it's construction using find_one() the docs show {"key": "value"}

[04:38:25] <Havalx> When substituting key, value with python variables I have: d = {x, y}

[04:38:57] <Havalx> Boomtime: does d need to be strings?

[04:39:54] <Havalx> Boomtime: and proper " ' format.

[05:19:20] <Havalx> Boomtime: http://pastebin.com/BtwCYptX

[05:45:26] <adsf> doubt many ppl are awake at this hour! :)

[05:56:50] <polydaic> Should I be using mongodb if I want to store a representation of a table

[05:57:18] <polydaic> or is it silly to use a NoSQL db in this case

[05:59:53] <adsf> probably not really needed :p

[05:59:57] <adsf> what is the purpose?

[06:00:33] <Boomtime> Havalx: what you pasted is pretty short.. can you try it in the mongo shell?

[06:01:10] <Boomtime> a query for {"x":7} will match any document that has a key called "x" with an integral value of 7

[06:01:33] <Boomtime> note that if x is an array in some (or all) documents then each value of the array is considered seperately

[06:01:57] <polydaic> adsf: I'm building something similar to TypeForm

[06:02:07] <polydaic> in usecase

[06:02:38] <adsf> are you replacing the table or just describing it?

[06:03:06] <adsf> Has anyone used a List<WriteModel<Model>> in a bulk update?

[06:03:12] <adsf> bulkWrite sorry, in java

[06:03:24] <polydaic> adsf: I'm describing it in arrays of objects

[06:03:40] <adsf> I have List<Model> but cant quite squeeze it into the WriteModel<Model> list

[06:04:12] <adsf> so you can recreate the database elsewhere?

[06:04:18] <Boomtime> Havalx: can you try your query in the mongo shell?

[06:08:13] <Havalx> Boomtime: It runs.

[06:08:50] <polydaic> asdf yep

[06:08:58] <polydaic> adsf:

[06:09:13] <polydaic> but I'm like 80% done with mongodb

[06:09:13] <Havalx> Boomtime: I passed {"x": 7} into insert.

[06:10:56] <adsf> polydaic: i suppose it could work, i mostlysee people doing it in xml

[06:11:30] <polydaic> wait why would you use xml over *SQL

[06:11:35] <polydaic> *over a db

[06:11:45] <adsf> i mean just to store a representation of the schema of the db :p

[06:12:16] <polydaic> hmmmmm

[06:12:25] <polydaic> adsf: so what you mean is I should store the schema in mongo

[06:12:44] <polydaic> and then update the postgresql db from that

[06:12:48] <adsf> you can store the schema where you want :) Long as youare happy getting it out

[06:13:08] <adsf> you could eport the table schema from postgres as sql

[06:17:53] <polydaic> also

[06:18:00] <polydaic> is mongoose really terrible?

[06:18:05] <polydaic> i've heard bad things about ti

[06:23:35] <gemtastic> It depends on what you're gonna do with it I suppose

[06:23:48] <gemtastic> Right tool for the right task

[06:42:02] <adsf> if i am using the new models in mongo 3, is it best to work on the iterator, or move to something else?

[06:42:12] <adsf> currently im in java and i convert it to a list of models

[07:20:12] <adsf> anyone konw why update_mongo.add(new UpdateOneModel<Mongo_Game>(collection.find(game).first(), game)); wouldthrough filter cannot be null?

[07:20:21] <adsf> i debugged and there is definitely an object there

[08:11:07] <adsf> can never find any examples of updating an existing document from an object queried with a codec :(

[08:31:17] <skiold> morning, what is the default value of :connect for Mongo::Client.new with the ruby driver (v2)?

[08:31:27] <skiold> as in http://api.mongodb.org/ruby/current/Mongo/Client.html#constructor_detailshttp://api.mongodb.org/ruby/current/Mongo/Client.html#constructor_details

[10:25:23] <aps> While trying to start mongos, I get following error:

[10:25:24] <aps> "ERROR: error upgrading config database to v5 :: caused by :: could not write initial changelog entry for upgrade :: caused by :: failed to write to changelog: could not verify config servers were active and reachable before write"

[10:26:29] <aps> I can connect to config members. what is the reason for this error?

[10:33:15] <arussel> I have a list of document {"a": "a"},{"a":"b"}, {"a":"c"} and a map (a -> 1, b -> 2, c -> 3) outside of mongo. Is there a way in an aggregate to have a replaced by 1, b by 2 and c by 3 so the list becomes {"a":1}, {"a":2},{"a":3} ?

[10:59:15] <synthotronic> Am following the instructions at http://docs.mongodb.org/v2.2/tutorial/install-mongodb-on-red-hat-centos-or-fedora-linux/ to install version 2.2.2, but it says "Package mongo-10gen is obsoleted by mongodb-org, trying to install mongodb-org-2.6.10-1.x86_64 instead", then tries to install 2.6.10-1. Anyone know the best way to actually install 2.2.2?

[11:17:19] <synthotronic> (I just grabbed the packages directly from http://downloads-distro.mongodb.org/repo/redhat/os/x86_64/RPMS/ and installed with rpm command.)

[11:27:49] <cheeser> why such an old version?

[11:48:55] <adsf> cheeser: never managed to get the bulk write working. So I copped out and updated in a loop

[11:50:18] <adsf> also if i do an update instead of a replace i get a Invalid BSON field name _id :(

[11:59:47] <adsf> /wc 13

[11:59:49] <adsf> fail

[12:40:42] <cheeser> adsf: sorry to hear that. if you post your failed attempt to the mailing list, jeff and I can work it out.

[12:45:47] <adsf> cheeser: just the mongodb-user one?

[12:50:24] <matthavard> Is there a way to run mongo so that it just writes to a local file like sqlite?

[12:51:25] <matthavard> Ah I guess there's this: https://github.com/hamiltop/MongoLiteDB

[12:52:35] <matthavard> Oh wait that's pure ruby

[13:01:02] <StephenLynx> but

[13:01:06] <StephenLynx> mongo just writes to a local file.

[13:01:30] <StephenLynx> pretty much any databases does that.

[13:16:34] <cheeser> adsf: yeah

[13:16:49] <cheeser> matthavard: what are you trying to do?

[13:17:31] <matthavard> cheeser I just want to use mongo for a small app that I just started making. I don't want to have to run the whole server and everything

[13:17:43] <matthavard> Like for the same purposes one would use sqlite

[13:18:29] <cheeser> what is "the whole server" to you?

[13:18:35] <matthavard> StephenLynx: There is a huge difference in the way mongo writes to a file and the way sqlite writes to a file. Let's not be pedantic here

[13:18:35] <cheeser> you want to embed mongo in your app?

[13:19:01] <matthavard> I don't want to have a mongod process running the whole time

[13:19:02] <StephenLynx> I am not being pedantic. That is LITERALLY what happens.

[13:19:13] <StephenLynx> if you don't know how sqlite works, don't call me pedantic.

[13:19:15] <cheeser> mongo isn't really set up to embed in applications.

[13:19:27] <matthavard> cheeser: that is what I am gathering

[13:19:30] <cheeser> but you can start/stop mongo with your app

[13:20:15] <matthavard> think I'll just dump to a json file

[13:20:37] <Derick> you might as well use sqlite then

[13:20:49] <Derick> as that's optimized for local on-disk manipulations

[13:20:59] <Derick> and handles locking a bit better than your app...

[14:51:29] <Absolute0> I have been using postgresql for all my projects. What will I be gaining by switching to mongo. What will I be losing?

[15:20:30] <amit001> Hi all project pymongo is managed by mongoDB?

[15:28:20] <cheeser> yes. http://docs.mongodb.org/ecosystem/drivers/python/

[15:41:10] <proteneer_> my upstart fails

[15:41:14] <proteneer_> it just says fail

[15:41:17] <proteneer_> and i can’t find the fucking log file anywhere

[15:41:21] <proteneer_> as to why its failing

[15:44:30] <ehershey> upstart is a pain

[15:44:41] <ehershey> are you using the ubuntu package?

[15:44:55] <proteneer_> yes

[15:45:34] <proteneer_> /usr/bin/mongod --config /etc/mongod.conf

[15:45:38] <proteneer_> works perfectly fine

[16:20:13] <danijoo> hey, i need to somehow reduce the size of a mongodb that is used for caching. all entries have an updatedAt field. If I set a TTL index, will this immidiatly delete all entries that are expired?

[16:23:50] <cheeser> yes, give or take a minute.

[16:24:02] <cheeser> but it won't necessarily reduce your on disk footprint

[16:24:53] <danijoo> its a 100gb database and i could shrink the entry count to about 25% of that. so i think it will :)

[16:25:25] <cheeser> disk space isn't necessarily released back to the OS

[16:25:44] <danijoo> i was on the wrong docs page. the 3.0 explains how it deletes pretty good

[16:25:50] <danijoo> cheeser, it doesnt? oh..

[16:26:15] <danijoo> can i force it somehow? Any keyword to google?

[16:26:37] <cheeser> http://blog.mongolab.com/2014/01/managing-disk-space-in-mongodb/

[16:27:24] <danijoo> awesome thanks.

[16:27:51] <danijoo> im moving to a new server right now. this is the only time where i can do this maintaince work without risks of production

[16:28:40] <danijoo> so.. the strategy would be to apply the TTL, and then do a db.repairDatabase()

[16:30:23] <cheeser> that's one option

[16:30:36] <cheeser> probably your best one. though that requires downtime.

[16:30:54] <danijoo> thats fine. i already dumped all data from server1 to server2

[16:31:10] <danijoo> i can now do full maintance on the new collection before switching the frontend to use it

[16:38:10] <allisthere2love> Hi, can somebody help me with user authentication in a 3 mongod, 1 arbiter, 3 config servers and 2 mongos cluster?

[16:40:36] <allisthere2love> I've created an application user in the primary mongod, but it does not exists in mongos. For what I understood, I have to create the user in mongos, but this user created in mongos isn't 'synced' to mongod

[16:40:40] <allisthere2love> is it correct?

[16:40:54] <allisthere2love> the users database are independent in mongod and mongos?

[16:41:31] <cheeser> that's correct, i believe

[16:42:54] <danijoo> oh dammit. just realized my updatedAt fields are NumberLong, no dates :/

[16:43:26] <allisthere2love> thanks, Cheeser. I was thinking there was some problem in the fact that I've created the user X in mongos and can't authenticate with it in mongod. But as I was writing, I think I've understood it :)

[17:31:44] <danijoo> is there an easy way to replace a NumerLong field opf every document in a collection with the corresponding date?

[17:47:32] <joannac> danijoo: write a script and do it as a bulk update?

[18:50:38] <greyTEO> is there a limit or recommendation on how many indexes to have?

[18:51:59] <danijoo> i'd say as less as possible and make sure they all fit into your memory

[18:52:30] <greyTEO> that is what I was thinking.

[19:08:15] <greyTEO> im having a problem querying fro documents where a field doesn't exists. From my understanding, $exists does not use an index?

[19:08:51] <greyTEO> since I am not using a sparse index, I tried 'field': null, but still seems to be the same slow query.

[19:10:26] <greyTEO> collection has about 2.5 million records.

[19:24:26] <cheeser> null is not the same as not $exists

[19:30:13] <GothAlice> Thus constructs like {$or: [{published: {$exists: 0}}, {published: null}, {published: {$lt: now}}]}

[19:50:05] <greyTEO> I thought the index stored a null value for documents that did not have the field for the index.

[19:50:18] <greyTEO> when sparse index was not being used.

[19:50:43] <greyTEO> needless to say it didn't work...

[20:11:06] <ngl> Really quick question:

[20:11:16] <ngl> Can I not do this?

[20:11:17] <ngl> unitsCollection.find({serial: /[a-z]/}, 'serial', function(err, units) {

[20:11:49] <ngl> I know I can in mongo, but with a mongoose model?

[20:12:11] <ngl> The regex part is what I'm asking about.

[20:18:19] <ngl> Nevermind. That is not where I'm breaking.

[20:36:36] <tejasmanohar> whats the best way to keep mongodb in sync with stuff published to postgresql?

[20:39:30] <cheeser> use something like flume or sqoop. maybe mongo-connector though i think that tends to be for outbound data.

[20:46:43] <tejasmanohar> ah

[20:46:48] <tejasmanohar> yeah thats all ive seen... outbound data

Log file Viewer

Help | Karma | Search:

#mongodb logs for Thursday the 2nd of July, 2015