Log file Viewer

[00:15:56] <cheeser> sewardrobert: is there an errant lock file in the old directory?

[02:41:22] <iio7> I am a newbie at both MongoDB and HHVM. I have done some testing using MongoDB with vanilla PHP on Apache, but want to try using HHVM. Anyone with experience who knows which MongoDB driver is best for HHVM? The "mongo-php-driver" or "mongofill"?

[04:28:09] <FrozenFire> Can anyone offer some good resources about bulk loading that I might not find on Google easily?

[04:28:17] <FrozenFire> I'm not seeing much in the way of optimization tips.

[04:28:46] <FrozenFire> Specifically I'm looking to optimize our backup loading process.

[05:37:06] <pagios> hi all, any idea if amazon is providing mongodb as a service in their cloud?

[05:37:26] <pagios> instead of paying for a vm, pay for the mongodb service directly via api?

[05:54:07] <masterken> Hi Mongo Community

[06:03:21] <masterken> WiredTiger is a big mess. Do any one else feel the same ?

[06:03:44] <Zelest> Why so?

[06:04:13] <masterken> Memory usage grows over time

[06:04:25] <masterken> Check no. of open issues related to WiredTiger on Mongo Jira

[06:05:17] <masterken> All the customers who moved to Wired Tiger looking at its data compression, speed and row level locking got a big shot on their head

[06:06:07] <masterken> It seems immature for 10gen to rollout WiredTiger before doing proper testing Zelest

[06:06:36] <masterken> Those tests are very easy to figure out if they have tested at all

[06:07:18] <masterken> We working on around 400 GB of data on normal operation found 10's of bugs and reported to Jira

[09:27:40] <lars_bauer> moin

[09:30:10] <lars_bauer> search in af text field and update rows with given text? db.task.update("5" $text text, ...

[09:54:42] <Naeblis> Hi. Can anyone suggest a good GUI tool for MongoDB on OS X ? I've tried RoboMongo but I can't get it to connect to remote databases.

[11:56:53] <mbwe> Could somebody tell me what the standard size of mongo's oplog is?

[11:59:16] <Hassaanz> Hi is there a way I can use aggregationto create a document containing subset of an array from a collection contain the parent array?

[12:28:56] <remonvv> Hi all, quick question. Under what circumstances would a mongod instance decide on a global or database level lock even though the query is shard targeted? We are using an older driver (Java 2.6.x).

[13:42:51] <deathanchor> remonvv: chunk splitting/migration does full shard lock

[13:43:18] <deathanchor> remonvv: also new DB, new collection, basically anything that modifies the config metadata

[13:44:02] <remonvv> It's not doing either though, although the balancer is active it is no chunk splits or migrations happen. We just get fairly consistent slow query reports like this : locks:{ Global: { acquireCount: { r: 2 } }, Database: { acquireCount: { r: 1 } }, Collection: { acquireCount: { r: 1 } } } 768ms

[13:44:09] <deathanchor> remonvv: also depending on your version of mongodb, locks act different within the DB.

[13:45:08] <remonvv> it's 3.x, WT

[13:45:09] <deathanchor> remonvv: db.currentOp({ active : true, waitingForLock : false }) to see what is actively going

[13:45:19] <deathanchor> ah sorry I haven't moved to 3.x yes

[13:45:31] <deathanchor> yet, but others here have more exp with it

[13:46:16] <remonvv> Alright

[13:57:47] <dcrosta> i have pymongo (3.x) connected to mongos (3.x) connected to a sharded cluster (also 3.x). how can I tell if mongos is properly respecting my NEAREST read_preference? we see some query & latency stats that suggest it might not be.

[14:00:32] <dcrosta> or, more generally, how can I tell through pymongo what host served a request (a find or findOne)?

[14:18:27] <deathanchor> dcrosta: the object properties get set I believe either in the result object.

[14:22:09] <dcrosta> deathanchor: properties on the collection object? or connection?

[14:27:58] <deathanchor> on the resultset

[14:31:45] <scrawl1971> New to IRC, please bear with me. I'm having a problem with a recent upgrade to 3.0.6 on a CentOS VM. Anyone up for discussing it?

[14:52:44] <scrawl1971> exit

[14:55:56] <StephenLynx> centos 7?

[14:56:06] <StephenLynx> I use 3.0 on centos 7 vms

[18:34:14] <jayline> I am trying to figure out this query problem. I need to fetch all documents from last 5 months, which have entries. So this is not necessary previous 5 months from now. These months can be anywhere in past. So documents may have been created in last year, but are still the latest ones. Any tips for this?

[18:52:19] <coudenysj> jayline: i'm not quite sure what you need

[18:52:25] <coudenysj> "I need to fetch all documents from last 5 months", but they can be older?

[18:55:07] <jayline> coudenysj: End result is monthly lists of documents and skipping the empty months. 5 latest of these monthly lists

[18:55:54] <jayline> coudenysj: if no entries found from last month, keep finding from previous etc

[18:56:05] <coudenysj> you will need a clever query loop, like this:

[18:56:26] <coudenysj> first loop: query last 5 months, filter out empty months and count how many you are missing,

[18:56:55] <coudenysj> second loop: query last the missing amount from before 5 months, filter empty, etc...

[18:57:05] <coudenysj> do that until you have 5 months

[18:57:43] <jayline> coudenysj: that sounds bad. Would it be better just to query months one by one

[18:57:55] <coudenysj> no way

[18:58:33] <coudenysj> in my case, in the best scenario, you have all your data in one run

[18:58:48] <coudenysj> one by one would always call 5 queries and counting

[18:59:03] <jayline> coudenysj: hmm, I understand it now

[18:59:18] <jayline> coudenysj: that is the best option so far

[18:59:20] <coudenysj> there is no solution to write a query for this logic

[19:00:48] <jayline> coudenysj: really good point. I most cases the first query probably is enough. Thanks!

[19:00:58] <coudenysj> yw

[19:07:13] <BigOrangeSU> Hi all, is it possibel to configure the mongo oplog to contain the whole record and not just the udpated fields?

[19:11:57] <asteele> hi can anyone help me with a query i am trying to run using aggregates: I have a collection of payments, i want to group them all up by the user (Field: _channel), sum up all of the payments for each user (field: convertedAmount), and then order them by that summed up amount to get users with top overall summed payments, if that makes sense

[19:20:48] <louie_louiie> hi asteele, this would be how I would do it of the top of my head: use $search or organize _channel's by group

[19:21:44] <louie_louiie> do a update() on the user field totaling all the payment variables

[19:22:00] <louie_louiie> then put a index on a total sum and order by highest to lowest

[19:25:31] <asteele> an update? that involves actually modifying the data?

[19:26:38] <louie_louiie> one could output it into another field like 'total' or something... it would not involve messing the the raw data

[19:32:56] <louie_louiie> @BigOrangeSU maybe pipe the output into a log file using SHELL piping? bumping for someone with more knowledge on the subject

[22:05:54] <tejasmanohar> is there a way to write a mongo query that excludes a certain _id from results

[22:06:02] <tejasmanohar> or excludes values with a certain key: value

[22:23:23] <melissamm> can someone tell me how to tell if a collection is empty with Mongoose?

[22:23:36] <melissamm> I'm using a MEAN stack

[22:23:41] <melissamm> so I'm using Mongoose with MongoDB

[22:23:44] <melissamm> thank you

[22:24:26] <Derick> You can do a query without criteria and see what you get back. You can run a count() on the collection. « Plain MongoDB, this is not a Mongoose support channel.

[22:34:39] <daidoji> tejasmanohar: $not

[22:35:07] <daidoji> or $ne

[22:35:10] <daidoji> http://docs.mongodb.org/manual/reference/operator/query/ne/

[22:35:19] <tejasmanohar> ahhh

[22:35:22] <daidoji> http://docs.mongodb.org/manual/reference/operator/query/not/

[22:35:22] <tejasmanohar> i saw not, but ne makes sense now

[22:35:29] <tejasmanohar> i want value, not expression ;)

[22:35:35] <daidoji> word

[22:39:20] <tejasmanohar> no more .filter( :)