pmxbot IRC Log Viewer

[01:47:37] <hahuang62> joannac: lol...

[01:47:45] <hahuang62> joannac: but it's so hard to make a ticket :p

[01:47:47] <hahuang62> jk

[01:48:11] <hahuang62> joannac: I think it's because a lot of people are used to getting instant responses in IRC, irrelevant of the quality of response :p I'm gonna file a ticket. Thanks

[03:38:24] <Jonno_FTW> anyone in?

[03:38:27] <Jonno_FTW> I have a problem

[03:39:15] <Jonno_FTW> I need to conditionally add a field to my documents

[09:31:35] <Lujeni> Hello - can i set as downtime througt mms to avoid a fake alert ? Thx

[09:41:49] <lxsameer> what is a single file size limit in GridFS? 16Mb ?

[09:50:28] <lxsameer> is there any file validation for gridFS ? for example content_type validation

[11:17:51] <mazzy> I get this error when I try to use mongoimport to import a json file

[11:17:52] <mazzy> invalid character '\x00' in string literal

[11:19:48] <Chepra> Hey, running 2.4, can we kill a index creation?

[12:00:04] <no-thing_> hello, how to make meta server see created users ?

[12:03:16] <no-thing_> if i add a user on a database when connected directly to mongod, it does not appear on configsrv mongod (though created dbs show themselves just fine in mongo meta server)

[12:15:42] <mazzy> the default path for journal is

[12:16:11] <mazzy> the following /data/db/journal right?

[12:50:58] <sharejs> Hello everyone

[12:51:57] <sharejs> It seem I've found a bug in $geoWithin operator http://stackoverflow.com/questions/29166054/mongodb-geowithin-box-bug

[12:56:00] <Domenique> hi

[12:59:11] <cheeser> i don't see any jira's open against $geoWithin

[13:00:25] <sharejs> cheeser: do you think it's not a bug?

[13:06:44] <cheeser> i don't use geo so i don't really know.

[13:08:23] <joannac> sharejs: you sure it'

[13:08:32] <joannac> s not picking the opposite box to you?

[13:09:09] <sharejs> joannac: no

[13:14:16] <joannac> sharejs: try switching -29 and -52

[13:16:29] <sharejs> joannac: unfortunately it does't work

[13:23:42] <joannac> sharejs: oh

[13:24:09] <sharejs> joannac: what?

[13:24:33] <joannac> sharejs: $box uses planar geometry?

[13:28:09] <sharejs> joannac: I don't understand what do you mean? I have 2d index on location field

[13:32:47] <mazzy> I have deployed mongodb in a container

[13:33:05] <mazzy> I have logged in and I have realized I don't have admin collection just local

[13:33:08] <mazzy> could be an issue?

[13:33:23] <Derick> you mean database

[13:33:39] <mazzy> yes sorry

[13:33:57] <Derick> afaik it's empty if you have not done anything yet

[13:34:21] <joannac> sharejs: 2d index is on a plane. there is no box with a left hand side at 165, and the right hand side at -175

[13:35:13] <mazzy> even if I add some db it isn't show

[13:35:27] <sharejs> joannac: but google Maps shows right bounds: http://jsbin.com/bomogojopa/1/edit?html,output

[13:37:50] <joannac> mazzy: how do you add the database?

[13:38:08] <mazzy> through mongoimport

[13:38:20] <mazzy> since I have data stored in json files

[13:38:29] <joannac> mazzy: what's the mongoimport line you use and the result? pastebin please

[13:39:19] <Derick> sharejs: 174 is not between -175 and 165

[13:40:31] <mazzy> just a simple mongoimport http://pastebin.com/A5RnyXcb

[13:40:48] <sharejs> Derick: but google Maps shows right bounds: http://jsbin.com/bomogojopa/1/edit?html,output

[13:40:55] <mazzy> in my case db name, collection name and file name are different

[13:40:57] <Derick> sharejs: that is not relevant

[13:41:04] <hydrozen> Hey, is it normal that this file is empty in the official source for ubuntu? http://repo.mongodb.org/apt/ubuntu/dists/trusty/mongodb-org/3.0/multiverse/binary-i386/Packages

[13:41:05] <sharejs> why?

[13:41:27] <hydrozen> im thinking thats why im unable to install it right now on my ubuntu box even if I follow the instructions

[13:42:20] <joannac> mazzy: and what's the output?

[13:43:40] <mazzy> 2015-03-20T13:43:09.128+0000 imported 53 documents

[13:44:02] <Derick> sharejs: 174 is not between -175 and 165

[13:44:24] <mazzy> @joannac: the database is correctly created along with the collection

[13:44:47] <joannac> mazzy: okay... so what's the problem?

[13:45:22] <mazzy> I don't have any problem...I have asked if not having admin database could be an issue

[13:45:46] <sharejs> Derick: so is it GoogleMaps bug? I'm getting bound from it

[13:46:01] <mazzy> is it normal or it can generated error ?

[13:46:14] <joannac> mazzy: yes it's normal if you don't have auth on

[13:46:15] <Derick> sharejs: it does something odd - you use a 2d planar query, so if 174 is not between -175 and 165, it won't be found

[13:46:42] <mazzy> ok perfect @joannac

[13:46:54] <joannac> hydrozen: i don't think there are packages for 32bit?

[13:46:55] <mazzy> I thought that it was a misconfiguration

[13:47:04] <hydrozen> yeah that’s what I just realized

[13:47:24] <sharejs> Derick: how could I resolve my problem? What I'm doing wrong?

[13:47:25] <mazzy> thank you very much @joannac

[13:47:32] <Derick> joannac: wiredtiger doesn't work on 32bit at all

[13:47:55] <Derick> sharejs: which index did you create?

[13:49:38] <sharejs> @Derick: { "location" : "2d"}

[13:50:19] <Derick> you need to change that to "2dsphere"

[13:51:22] <Derick> actually, I don't think you can actually do this boundary

[13:51:30] <Derick> because it spans the meridian

[13:51:34] <sharejs> @Derick: I'm followed official docs http://docs.mongodb.org/manual/reference/operator/query/box/#behavior "Only the 2d geospatial index supports $box."

[13:51:50] <Derick> yes, but you can replace a box by a polygon with the corners yourself

[13:52:01] <Derick> *that* is the non-deprecated way

[13:52:51] <sharejs> @Derick: how should I rewrite my query?

[14:48:49] <wc-> hi all, im about to mess around with mongod log verbosity, i figured i should stop by here and ask this before i do: is there a good overall global verbosity setting that will just display log entries when 1. something actually breaks 2. slow queries?

[14:52:42] <atatus> hello, in mongodb is there any performance improvement to choose aggregation over normal cursor operation. I’m using mgo-golang and both the below commands return the same output

[14:52:42] <atatus> pipeline := []bson.M{{"$match": query}, {"$project": fields}, {"$sort": sort}, {"$skip": skip}, {"$limit": limit}}

[14:52:42] <atatus> c.Pipe(pipeline).All(&l.Data)

[14:52:43] <atatus> or

[14:52:44] <atatus> c.Find(query).Select(fields).Sort(sort).Skip(skip).Limit(limit).All(&l.Data)

[14:52:55] <atatus> Is there any reason why I should lean to use pipeline

[14:56:14] <StephenLynx> afaik, they don't actually return the same output, using io.js aggregation returns me an actual array while find returns me a cursor.

[14:56:52] <StephenLynx> and it is easier to perform operations like sort and skip using aggregation when you are working asynchronously.

[14:57:54] <StephenLynx> try something, atatus. can you print the type of objects in go?

[14:58:00] <atatus> StephenLynx: WRT to find returning the cursor, does that mean the last All operation is a second call to the mongo

[14:58:04] <StephenLynx> try it in both outputs.

[14:58:16] <StephenLynx> WRT?

[14:58:38] <atatus> WRT - With respect to the statement that find returns the cursor

[14:58:55] <atatus> could be that the mgo driver performs another call to the mongodb server

[14:58:58] <atatus> when the All method is called

[14:59:01] <StephenLynx> no, it won't perform aditional queries.

[14:59:08] <StephenLynx> afaik

[14:59:48] <StephenLynx> but I am not sure, that is indeed a good question.

[15:00:15] <StephenLynx> GothAlice would know it for sure.

[15:00:25] <atatus> Aggregation does seem easier, but just wanted to know what exactly goes in the background

[15:00:47] <StephenLynx> you just need to watch out of big result sets with aggregation.

[15:00:57] <StephenLynx> for*

[15:01:58] <atatus> I use a limit of 20 in both cases, so big results would not reach my layer.. but whether the aggregation query within mongodb consider this limit of 20

[15:02:03] <atatus> and then handle it more efficiently

[15:02:52] <StephenLynx> yes, the aggregation pipe has a bunch of optimizations.

[15:03:01] <StephenLynx> for these combined operations

[15:03:21] <StephenLynx> http://docs.mongodb.org/manual/core/aggregation-pipeline-optimization/

[15:05:09] <StephenLynx> so thats a another plus for aggregations when you have multiple operations besides a simple match and project

[15:08:28] <atatus> StephenLynx: Okay, thanks. I will use aggregation going forward

[16:51:09] <mdavid613> hey all…has anyone done a mongorestore with x509 authentication here?

[16:51:14] <mdavid613> successfully? ;)

[16:59:20] <wc-> hi all, does anyone know if it is possible to get less logging than verbosity: 0?

[16:59:27] <wc-> just errors and warnings for example

[17:49:13] <ggoodman> I'm in the process of upgrading a replicaset from 2.6 to 3.0.1 (2 monods and one arbiter). When the secondary (now 3.0.1) had finished replicating, I asked the primary to step down.

[17:49:45] <ggoodman> The now primary 3.0.1 (wiredTiger) primary started handling live queries and fell flat on its face

[17:49:59] <ggoodman> is there an additional index build step that I might have missed?

[17:53:38] <Joel> I have three collections total, each collection is a tenth of a gig. I have two prealloc.(1|2) files, a gig each, and a j._0 file at a gig. Any idea on how I can reduce these? Or see how they got to be so big?

[17:53:45] <ggoodman> specifically, count queries were suddenly running for minutes

[18:02:26] <ggoodman> ah, perhaps a driver issue :S

[18:08:16] <Joel> ahh, I had smallFiles set, but I hadn't stopped mongo, removed the old ones, and fired it up again

[18:34:49] <girb1> userAdmin auth fails even if give the correct username and password

[18:34:53] <girb1> any reason ?

[18:49:45] <ggoodman> Yeah, definitely seeing a huge degradation in performance with a mongodb 3.0.1, wiredTiger migration... can anyone help?

[18:53:46] <StephenLynx> try to pinpoint what operation you are having issues with

[18:54:25] <StephenLynx> then look for open issues. If I remember correctly there was a couple of things that were broken on 3.0

[18:57:46] <ggoodman> I'm having issues with count queries from what I can tell (I'm struggling to figure out how to debug this)

[18:58:23] <ggoodman> I can see that counts seem slow b/c db.currentOp() is showing them as the only ops consistently over a second

[18:59:20] <ggoodman> perhaps related: what would consequences of indexing on a sparse field be?

[18:59:54] <greyTEO> what is the fastest way to read the mongo oplog?

[19:00:49] <greyTEO> using the capped collections seems to have a lag time when hundreds of operations are happening at a time

[19:01:15] <StephenLynx> ggoodman do you know what is the query that is slow, then? could you paste your application code?

[19:02:23] <ggoodman> Application code is buried in layers of nightmares, coffee-script and mongoose

[19:02:45] <ggoodman> and don't forget regret

[19:03:29] <StephenLynx> :v

[19:03:47] <StephenLynx> yeah, then maybe that would be a good starting point.

[19:04:59] <ggoodman> StephenLynx: but the question for me is that this code was producing particularly slow results until I migrated one node to 3.0.1 and used the wiredTiger engine

[19:05:19] <ggoodman> this code was *not* producing

[19:05:41] <StephenLynx> yeah, but you don't know which query exactly is causing that.

[19:05:58] <ggoodman> right now I have 3 nodes in the replicaset 1: arbiter, 2: 2.6.8 (primary) 3: 3.0.1 (secondary)

[19:06:18] <ggoodman> well from db.currentOp, I can see that count operations are slow

[19:06:44] <StephenLynx> by count you mean a group with an $inc?

[19:07:04] <ggoodman> equivalent of db.collection.count({query: stuff});

[19:07:45] <ggoodman> so I generally execute two queries for collections: 1 to get a page of data, 2 to get the number of records

[19:08:16] <StephenLynx> have you tried performing the same count query in the shell?

[19:10:26] <ggoodman> it shall be done

[19:11:12] <ggoodman> yes, extremely slow

[19:11:42] <StephenLynx> http://stackoverflow.com/questions/7658228/mongodb-count-is-very-slow-how-do-we-refine-work-around-with-it

[19:12:16] <StephenLynx> this guy suggest an index

[19:14:34] <StephenLynx> and according to http://docs.mongodb.org/manual/reference/command/count/ you can specify an index to be used

[19:26:06] <ggoodman> Hi StephenLynx thanks for the links. Your questions triggered me to look deeper into the queries used vs indices present and I hope to have some improvements :D

[19:26:20] <StephenLynx> np

[19:26:48] <StephenLynx> I always use the pure driver so if something goes wrong, I can easily spot and fix.

[19:27:08] <StephenLynx> layers of abstraction looks like a good idea until they bite you in the ass big time

[19:44:35] <ggoodman> StephenLynx: I would tell my noob self the same thing ;-)

[19:44:56] <ggoodman> gotta make mistakes to learn though!

[19:44:58] <StephenLynx> and since you are using the node.js driver, have you heard about io.js?

[19:47:39] <fewknow> ggoodman: if you use explain().n it is faster than count()

[19:48:39] <ggoodman> fewknow: that is odd... haha... unfortunately my piles of abstraction (read: regret) will prevent me from using a hack like that

[19:54:58] <StephenLynx> mongoose is cancer in code form :v

[19:55:19] <StephenLynx> and you can always use the pure driver when you wish.

[19:55:43] <cheeser> i feel that way about javascript in general.

[19:55:44] <cheeser> :)

[19:57:22] <StephenLynx> when it comes to scripted languages, I think javascript is quite good.

[19:58:02] <StephenLynx> you can go with just minimalistic original implementations, there is strict mode

[19:58:08] <StephenLynx> and has a very C-like syntax.

[20:02:45] <ggoodman> one new, well-placed index and this is flying

[20:03:13] <ggoodman> actually never seen the site so fast... o.O

[20:04:44] <t355u5> hello, I experience a rather strange behavior when running one of the tests. this is with 3.0.1.

[20:05:28] <cheeser> ggoodman: good to hear :)

[20:05:52] <t355u5> repl_coordinator_impl_test fails with "StepDownAndBackUpSingleNode Expected: getReplCoord()->getMemberState().primary() @src/mongo/db/repl/replication_coordinator_impl_test.cpp:1118"

[20:06:08] <ggoodman> http://plnkr.co if you're into front-end

[20:06:40] <ggoodman> The real test will be next Wednesday peak ~10am

[20:06:54] <cheeser> what's that?

[20:06:57] <t355u5> but, if I run it like this: repl_coordinator_impl_test >out.txt ; no error SUCCESS - All tests in all suites passed

[20:08:14] <t355u5> does anyone have an explanation for this?

[20:21:43] <tubbo> haha is plunker just like forrst for devs who only use bootstrap? ;)

[20:21:59] <tubbo> j/k this looks pretty neat.

[20:26:59] <t355u5> if I compile with --wiredtiger=off, does that mean that wiredtiger will not be supported or that it is not activated by default?

[21:14:32] <t355u5> anyone has an idea why the test fails when not redirected?

[21:35:50] <greyTEO> is it possible to have multiple threads polling the oplog? Im pretty sure I will get duplicate data though.

[21:38:27] <epx998> is it possible to import 2 schemas from 2 different mongo primaries into a new single mongo primary?

[22:04:31] <epx998> shouldnt this work?

[22:04:37] <epx998> test0:SECONDARY> test0:PRIMARY> db.upgradeCheckAllDBs()

[22:04:39] <epx998> 2015-03-20T17:01:54.480-0500 ReferenceError: PRIMARY is not defined

[22:04:41] <epx998> test0:SECONDARY> Fri Mar 20 17:01:37.662 TypeError: Property 'upgradeCheckAllDBs' of object test is not a function

[22:04:43] <epx998> 2015-03-20T17:01:55.718-0500 SyntaxError: Unexpected identifieru

[22:08:51] <fewknow> epx998: use admin ?

[22:09:32] <epx998> so just use admin - then run the command?

[22:09:45] <fewknow> I think

[22:09:55] <fewknow> haven't done it in a while, but I think yo have to be in the admin db

[22:09:59] <epx998> same error

[22:10:33] <epx998> wait, how do i connect a 2.6 shell to 2.4

[22:11:03] <fewknow> same way

[22:11:05] <fewknow> mongo

[22:11:13] <fewknow> mongo --host

[22:12:20] <epx998> im new to this, 3rd day. so from a 2.4 mongo, i run that command against a 2.6 mongo?

[22:36:56] <epx998> i just keep getting that error regardless meh

[23:04:32] <lost1nfound> Hey, any of you guys know if there's a good mongo web interface that isn't rockmongo? rockmongo was working well for us, but after upgrading to 2.6.8 (from 2.4 where it worked fine), it seems pretty broken, and has no active development anymore...

[23:05:14] <lost1nfound> issue driving me nuts now is that I cant query am objectid anymore like "{ "_id": ObjectId("550c94d46e44273c597b23c7") }"

[23:05:56] <lost1nfound> says its invalid json.... im sure thats a rockmongo issue not a mongodb or driver issue, (well, maybe driver, but 1.5.5 still does it, as well as 1.6.5), but such query works fine at the shell

[23:06:33] <lost1nfound> phpmoadmin is ugly as fuck, genghis doesnt even allow json queries, cant really find any modern up-to-date viable alternatives

[23:10:04] <Serializator> Can someone help me with a MongoDB Java Driver problem?

[23:19:29] <cheeser> way to hang in there Serializator

Log file Viewer

Help | Karma | Search:

#mongodb logs for Friday the 20th of March, 2015