Log file Viewer

[00:00:07] <skot> That is always a challenge for each language.

[00:00:30] <cheeser> especially one in transition like Go is.

[00:00:48] <skot> It is still so active, or unstable?

[00:01:49] <skot> I thought things were more stable after 1.2/3

[00:03:18] <niemeyer> skot: You're certainly right.. any project with a maintainer that cares deeply about the project will eventually see someone frustrated for trying to do something that isn't aligned with the project design.

[00:03:57] <niemeyer> skot: That's still much better than the alternative, though..

[00:05:04] <niemeyer> skot: I don't recall the particular events you're alluding to.. if you feel I wasn't as helpful as I could have been at some point, I apologize.

[00:05:29] <kurushiyama> niemeyer But I am right that in some things you oriented yourself on stdlibs sql?

[00:05:35] <niemeyer> kurushiyama: Absolutely

[00:05:52] <niemeyer> kurushiyama: Wait, not sql, no

[00:05:56] <niemeyer> kurushiyama: The stdlib itself

[00:06:01] <niemeyer> kurushiyama: sql was a late comer

[00:06:15] <niemeyer> kurushiyama: In some cases, it was the sql package that looked at mgo

[00:06:59] <kurushiyama> niemeyer Hahaaaa.

[00:07:36] <kurushiyama> niemeyer btw, even more thanks for gopkg.in ;)

[00:08:03] <niemeyer> kurushiyama: Glad you like it, and thanks for trying to make your own packages offer a stable API.. that's helpful for everyone

[00:33:48] <skot> niemeyer: no need for an apology. I didn't mean my comment to be negative either, just descriptive.

[00:35:21] <skot> niemeyer: glad to see you online; it has been a while -- hope to you around again, and soon.

[00:35:52] <skot> but I must go offline for a while to change venues.

[03:29:28] <ehershey> we should get a bot to do that

[06:53:01] <moogway> hi, could any kind soul shed a light on what is the difference between insertedCount and result.n in insertOnewriteOpResult? Thanks

[06:54:48] <moogway> insertedCount supposedly is "The amount of documents inserted" while result.n is "the total count of documents inserted"... but both of them are the same in value in almost all the write operations i've made

[06:56:46] <moogway> anyone?

[06:59:06] <Boomtime> @moogway: is this a bulk insert op or something?

[06:59:43] <moogway> @Boomtime: this is for any insert operation

[07:00:27] <moogway> http://mongodb.github.io/node-mongodb-native/2.1/api/Collection.html#~insertOneWriteOpCallback

[07:01:03] <moogway> actually this one - http://mongodb.github.io/node-mongodb-native/2.1/api/Collection.html#~insertOneWriteOpResult

[07:15:36] <Boomtime> @moogway: that just looks like you're seeing the generic server response as well as the translated response - the 'n' field is the generic count from the server saying how many docs were involved

[07:16:15] <Boomtime> 'n' is often present but changes meaning depending on the operation - 'insertedCount' is unambiguous

[09:05:37] <mkozjak> is there an official systemctl script for mongo-connector available?

[09:28:06] <Ange7> hey

[09:28:29] <Ange7> how can i make db.collection.update({}, {$set: {version: new MongoDate()}})

[09:28:38] <Ange7> doesn't work, update only the first doc

[09:28:48] <Ange7> i want update all document in collection

[09:30:13] <Derick> Ange7: look for "multi" option. It's in the docs

[09:31:03] <Ange7> Ok, thank you Derick

[09:32:04] <Ange7> i hope it work with PHP Driver :D

[09:32:36] <Ange7> Option multi not working with the PHP Driver

[09:33:04] <Ange7> Ok it's work

[09:33:08] <Ange7> multiple not multi

[09:38:21] <Derick> Ange7: sounds like you're using the old driver. You really shouldbe using the new one, with the new library.

[09:38:30] <Derick> which handles all those intricacies for you

[12:52:33] <dashu> hey guys im generating mongo querries in java but the querries behave really strange. some give me data and some dont. despite that there is data available in the mongodb. here is the DBCursor Object my querry is in.

[12:53:07] <dashu> the querry is always completely the same just the objectid changes

[12:53:23] <dashu> http://pastebin.com/XXQingLT

[12:53:41] <dashu> im using mongo 2.6.11

[12:57:08] <dashu> can someone maybe help ?

[13:10:47] <rekoil> i'm trying to install mongodb on freebsd 10.3 (well pfsense 2.3 really) but it's apparently not available from ports

[13:10:52] <rekoil> anyone know why this is?

[13:13:24] <kurushiyama> rekoil Iirc, BSD is not supported

[13:13:33] <rekoil> fair enough

[13:13:40] <rekoil> i'll ask in #freebsd instead

[13:13:48] <rekoil> it has worked in previous versions

[13:13:52] <rekoil> just not 10.3/2.3

[13:13:52] <kurushiyama> rekoil OSX, Win, some Linux derivates and Solaris is.

[13:14:03] <rekoil> thanks anyway

[13:14:10] <kurushiyama> rekoil maybe the package maintainer just gave up...

[14:06:10] <dashu> :(

[14:07:33] <kurushiyama> dashu tsup?

[14:17:56] <dashu> i dun know :/ this strange bug is driving me crazy

[14:18:49] <kurushiyama> what exactly?

[14:22:50] <dashu> i have written this generic querry builder in java that you can give as many filters you want to search for in my mongodb. and i thought it works but my mongo behaves strangely. i send a querry with an objectid and two dates

[14:22:55] <dashu> to my mongo

[14:23:08] <dashu> and sometimes it gets data and sometimes it doesnt

[14:23:33] <dashu> some ids just seem not to get any results

[14:23:45] <dashu> but there is data for those ids in the mongo

[14:24:09] <dashu> http://pastebin.com/XXQingLT

[14:24:27] <kurushiyama> dashu Well, it is safe to assume that MongoDB would handle queries on multiple fields correctly.

[14:24:38] <dashu> this is the dbCursor object

[14:25:39] <dashu> with the querry i send to the mongo

[14:26:05] <kurushiyama> well, the first thing I see is nested consecutive ANDs, which is sort of useless.

[14:26:27] <dashu> yeah useless but shouldnt it still work ?

[14:27:46] <dashu> can i somehow see on which shard the data of a specific id lies ?

[14:33:04] <dashu> i think it is maybe the $eq that is messing everything up

[14:33:32] <dashu> because i used a mongo 3.0 tutorial but i have to use mongo 2.6.11

[14:33:55] <dashu> and there came new $eq functionality

[14:33:59] <dashu> with 3.0

[14:34:07] <dashu> i think

[14:34:07] <kurushiyama> you can identify the shard via the shard key and sh.status()

[14:34:22] <kurushiyama> dashu Have to run, maybe later

[14:34:33] <dashu> ok cu ;)

[18:23:05] <hyades> I am having a collection which has 2 keys - say Day, Hour. I am having a compound index on {Day:1, Hour: 1}. However, doing a db.col.distinct('Hour', {Day: 'some day'}) takes a huge amount of time. Anyway I can speed things up?

[18:43:36] <dman777_alter> how do I show parse indexes? db.collection.getIndexes() doesn't show them

[19:10:56] <MacWinner> for wired tiger snappy vs zlib, is the performance impact mainly on insert/update? or also on read?

[19:11:46] <StephenLynx> hm?

[19:12:02] <StephenLynx> what exactly are you referring to?

[19:12:43] <MacWinner> when deciding between using snappy vs zlib for compression, I'm trying to figure out where the downside of zlib is since it compresses better for my data set

[19:13:08] <StephenLynx> it doesn't matter.

[19:13:08] <MacWinner> i don't have a heavy load now, but just planning for the future

[19:13:20] <StephenLynx> you are compressing on the application layer.

[19:13:26] <StephenLynx> mongo just stores the data.

[19:13:54] <MacWinner> i thought wired tiger storage engine is doing all the compression logic.. not my application

[19:14:09] <StephenLynx> wired tiger does use some compression.

[19:14:25] <StephenLynx> but I don't think you can tell it how to do it.

[19:14:53] <StephenLynx> but it will serve you the same data.

[19:14:57] <MacWinner> i think you can: https://www.mongodb.com/blog/post/new-compression-options-mongodb-30

[19:16:05] <StephenLynx> ah

[19:16:09] <StephenLynx> I didn't know about that.

[19:16:14] <StephenLynx> nevermind what I said.

[19:36:49] <CaptTofu> hi all!

[19:37:13] <CaptTofu> I am presenting at Percona Live in a few weeks and my demo is automating a mongodb sharded cluster

[19:37:38] <CaptTofu> it works flawlessly, but I'm not a mongo expert and want something that I could perhaps run to show that the cluster is working as expected

[19:38:01] <CaptTofu> what do people here suggest? What test data lends itself well to showing off replication and sharing?

[19:39:42] <jmmk> I am experiencing performance problems with an ETL process on a large collection. We need to do a full collection scan (not feasible to fit everything in memory) and read every record and send it elsewhere. The problem we are seeing is that the disk reads are incredible high (using EBS Provisioned IOPS and we are maxing out our 4000 IOPS). Is there anything that will reduce the number of disk reads required to scan the collection?

[19:52:21] <mylord> how do I get the _id of the inserted record?

[20:34:48] <rectalogic> I have a unique compound index and cannot get mongodb to use it with a regex prefix query - it will use a non-unique compound index - is this a limitation of unique indexes? sample explains: https://gist.github.com/rectalogic/07195f36fb4466b346a8f8cce559ee9b

[22:34:15] <FatalNIX> TEst

[22:57:48] <kurushiyama> test

[22:57:58] <kurushiyama> FatalNIX Successful

[22:59:19] <FatalNIX> kurushiyama, you look familiar

[22:59:31] <FatalNIX> okay I guess not

[23:19:59] <kurushiyama> FatalNIX From where?

[23:21:35] <FatalNIX> You had a similar nick to somebody I saw earlier