Log file Viewer
#mongodb logs for Monday the 11th of May, 2015
[00:48:00] <tejasmanohar> CastError: Cast to undefined failed for value "[object Object]" at path "referrals"
[00:48:04] <tejasmanohar> https://gist.github.com/tejasmanohar/085b373d4f8da9ea8d98#file-controller-js-L80
[00:48:20] <tejasmanohar> not sure how to track down this issue... i console logged both values and they both exist as shown in the error log i also attached to that gist
[00:49:48] <StephenLynx> GothAlice would recommend alternatives( MongoEngine in particular), I would recommend just using the driver.
[00:50:19] <StephenLynx> and since I just found out the connection you get from connect is a connection pool
[00:51:35] <tejasmanohar> StephenLynx: i console.logged the values of those variables so im certain enough that they are correct vals
[00:52:37] <StephenLynx> if you have bananas in a doc and camels in another doc in the same collection you get your bananas and camels and thats i.
[00:53:22] <tejasmanohar> StephenLynx: would the result of the raw query from mongo driver help here?
[01:36:58] <GothAlice> Fun aside: since MongoEngine is Python, not JavaScript, I would only recommend using the base JS driver or an async variant; async is nice if you can keep your code clean.
[01:37:40] <tejasmanohar> just got lazy for this project since its a waiting list app and once the waiit is done in 2 weeks from tmrw
[01:38:57] <tejasmanohar> with effort X you can get more readable code in promises than callbacks and `async` i ensure u
[01:39:25] <StephenLynx> you can't because you are adding more abstraction and moving parts to the code.
[01:44:39] <tejasmanohar> but oh god im hating mongoose rn, "MongoError: n/a" n/a is that really what mongo told you?
[01:45:45] <GothAlice> (Having "[object Object]" strings is a general JS silly thing, but Mongoose tends to draw it out. ;)
[01:48:12] <StephenLynx> you would still have to check if the query returned an error because the driver isn't designed around promises
[01:48:40] <StephenLynx> but now you have to implement a function elsewhere that will receive the two separate callbacks
[01:50:11] <StephenLynx> anyway, the point is that now you have a code more complex to work with promises
[01:51:45] <StephenLynx> that also enables you to further expand your logic by passing the callback around.
[01:52:27] <StephenLynx> what if you have a condition that you will perform a second asynchronous calls?
[01:55:41] <StephenLynx> and besides, callbacks are the most elegant way to deal with abstraction, if you ask me.
[04:56:37] <swhitt> does this look right for a nodejs implementation of a tailable cursor? https://gist.github.com/downslope7/9ec51905e4e8f2677224
[04:58:54] <swhitt> i inserted multiple records into the capped collection, nothing happens in the node console
[09:42:17] <abishek> could somebody advice on what could be causing this query to be less performant? http://pastebin.com/Jk5jMNNR . I have all the columns on the cocndition field indexed. what could be causing it to perform slow?
[09:56:24] <abishek> could you suggest how I convert that query to aggregate. I am newt o this and finding it difficult to figure out how to create the aggregate query
[09:56:48] <boboc> Guys i have a question, if i have a post with comments and choose to embed them, if i want to uniquely identify a comment and update it how can i do that? Would be better to have a separate document for comments?
[10:07:20] <abishek> nfroidure, when using aggregate how do I specify conditions in the reduce in the aggregate?
[10:08:39] <abishek> nfroidure, thats the global condition for the whole query. how about the if conditions on the reduce for the paste that I provided?
[10:31:49] <nfroidure> abishek, you can do a lot of things with aggregates once you master them. In fact, i never had to user reduces
[10:32:29] <abishek> yes, am right now figuring out how to master them. I have a lot of aggregates that I need to implement, so figuring out where to start
[10:45:36] <Cust0sL1men> I can migrate to wiredtiger without downtime if I just upgrade one replica set member at a time right /
[11:07:46] <bogn1> Hi all, is there anybody here that knows where mms-automation-agent stores its process configuration?
[11:09:23] <bogn1> removing mms-cluster-config-backup.json, the automation pid file and the /data dir doesn't help
[11:26:38] <bogn1> Please help, I'm going nuts. MMS is continuously restarting a process that I have removed with their web interface. Removing mms-cluster-config-backup.json, the automation pid file and the /data dir doesn't help either. Where can I find the automation agent's config?
[11:44:23] <ignasr> { "v" : 1, "key" : { "keyword" : 1 }, "name" : "keyword_1", "ns" : "skelbiu.keywordsLogTmp", "category" : 1 }
[11:44:32] <Cust0sL1men> https://slack-redir.net/link?url=http%3A%2F%2Fdocs.mongodb.org%2Fmanual%2Fcore%2Findex-intersection%2F&v=3
[11:49:00] <Cust0sL1men> if I have x = 1 & y = 2 || x = 1 & z = 2 - does it help to make x, y, and z a compound ?
[13:26:14] <swhitt> can someone help me with my tailable cursor / nodejs setup? the node process isn't picking up anything added to the capped collection. https://gist.github.com/downslope7/9ec51905e4e8f2677224
[13:30:52] <StephenLynx> dunno, never used that platform, you should consult the driver's documentation.
[13:36:47] <cheeser> lemonxah: you'd use $exists http://docs.mongodb.org/manual/reference/operator/query/exists/
[13:39:58] <cheeser> lemonxah: my scala is weak to say the least, but i think it'd look something like this: coll.findOne(MongoDBObject(“someField” -> MongoDBObject(“$exists” -> true))
[13:50:15] <AlienDrizzt> Hello. I have a question about de C API. Is is possible to make a call to a function in a filter collection, Ie I'm looking for C code that do something like that > db.historic20150511.find({}).count()
[13:55:11] <abishek> what is the issue here http://pastebin.com/vAzeP3yU ? I get this error message "exception: A pipeline stage specification object must contain exactly one field."
[14:03:53] <StephenLynx> I am not sure a $cond goes as a primary aggregation operator. can't you use a $match for that?
[14:06:00] <ignasr> hello, anyone here done a database compaction? How much it took with how much of data?
[14:12:36] <abishek> could I ping someone to discuss on some performance issue that I am having on Mongo?
[14:57:48] <freeone3000> I'm getting the error https://gist.github.com/anonymous/68552dcc2879eab4f0e9 when restoring from a disk-based backup. How can I tell mongo to just resync the bad records?
[14:59:43] <freeone3000> How can I get a reliable backup without taking an entire node offline, since apparently taking the disk image isn't going to give me consistent records?
[15:07:10] <erenburakalic> Hi everyone, is there anyone who DOES NOT ADVICEs to use mongoexport as backup solution? (DB IS ABOUT ~ 1-1.5GB). if yes why?
[15:09:33] <GothAlice> erenburakalic: https://gist.github.com/amcgregor/4fb7052ce3166e2612ab#backups
[15:10:30] <GothAlice> erenburakalic: I run a replication secondary in the office that is delayed by 24 hours, allowing us to recover not just from catastrophic failure of the main cluster "in the cloud" (we also have a non-delayed in-house secondary) but also recover from user error such as accidentally dropping a collection, or garbling a record.
[15:10:56] <GothAlice> We had one dev accidentally multi-update the same value across an entire collection once. That was fun.
[15:17:32] <SixEcho> so i have a unique+sparse index on .things.name… mongo will fail correctly when dup keys on different documents, but on the same document it appends dups to the array… { _id:x; things: [ { name: 'hi' }, { name: 'hi' }] } any way to make it fail in this case? bug? wad?
[15:17:55] <GothAlice> erenburakalic: Gist updated to include a delayed replicas and a lot more links into the documentation on the backup topics.
[15:36:50] <GothAlice> christo_m: You can, though there are few expert users (and many who strongly dislike mongoose) hanging around. I believe it has its own channel, though.
[15:38:10] <GothAlice> Ask, though, don't ask to ask. IRC is asynchronous, and you never know if someone can answer. :)
[15:49:09] <Cygn> Hello everyone. I want to use this query to fetch the id of one, the count of all entries and in addition i want to get the sum of the price of each sale. Fetching the id and the amount of sales works smooth and fine, but i can't get a value out of my sum, any hints? http://pastebin.com/rzEcTvnc
[15:52:11] <Cygn> StephenLynx: Sounds weird but it is, actually this comes from the business logic, where we have id's (out of the factory) to each item attribute. f.e. 10 stands for the price, 1 for the article id, 6 is the time of sale etc.
[15:53:07] <Cygn> StephenLynx: i could, but that would mean i would have to rename it back to the numbers everytime i give an export for the client.
[15:53:38] <Cygn> StephenLynx: okay, but in general this should work? so the field of each sale should be summed with that code right?
[15:54:49] <Cygn> GothAlice: It's not about my use, more about the namings they need, the really use "ten" as the naming when they talk about the price for example. Anyway, sure i could change it in my intern logic and rename it when i handover data to the client, but right now this is acutally okay for me.
[15:55:46] <GothAlice> Cygn: I'm saying the way you have structured this is bypassing any assistance MongoDB can offer in terms of efficiently querying or manipulating that data. How the data is presented to the user is pretty much irrelevant to how you store it, and the current method is actively defeating MongoDB's capabilities.
[15:56:37] <StephenLynx> you are losing much more by not doing two operations than you would by doing these operations.
[15:57:19] <Cygn> Okay, i am with you, i will most likely change this afterwards… anyway for now… i am really confused by my problem. $min, $max, $last, all working fine. As soon as i want to use $avg or $sum i get 0 …
[15:58:27] <GothAlice> Cygn: It's like this other trap people often fall into: {foo: 1, bar: 2, baz: 3, …} with an ever growing list of names that are completely dynamic. That won't fly. {fields: [{name: 'foo', value: 1}, …]} will preserve the ability to update individual values, while also letting you make use of indexes. (Index on fields.name == really fast lookup. No index on "foo", "bar", "baz", and friends means terrifyingly slow collection scans.)
[15:59:39] <Cygn> StephenLynx: Seems to be fine, even $min and $max present me the correct and expected values (as they are the highest/lowest price in my collection)
[16:12:51] <StephenLynx> if at least they were arrays where you consulted the indexes then it would make SOME sense.
[17:16:56] <mrmccrac> a db.repairDatabase() is supposed to reclaim disk space no longer needed i thought?
[17:31:05] <christo_m> im not sure how i can make the id field here and the user field be unique pairs in the database: http://pastie.org/10182760
[17:39:44] <Razerglass> hey is anyone availible to answer a quick question about saving a 2d array into my db?
[17:41:55] <greyTEO> Razerglass, it's never good to ask to ask. You will get much better results if you post your question.
[17:42:14] <christo_m> according to the mongo docs that should be enough http://docs.mongodb.org/manual/core/index-compound/
[17:46:03] <Razerglass> for when i declare my new schema does this look correct for saving a 2d array ({when i create my mongoose schema " about: [{ about1 : Array, about2 : Array}] })
[17:48:08] <Razerglass> then when i am setting the values up before saving im putting the 2 arrays in this format newDBentry({ about: {about1: array1, about2: array2})
[17:48:46] <Razerglass> for some reason when it saves in my DB its a empty array, i have something wrong with the formatting
[17:50:01] <christo_m> I basically want these fields to be unique pairings.. ie, i dont care if there are duplicate id or user entries, i only care that together when i look at id,user .. its a unique pair
[19:23:23] <xaxxon> how do I do a query to see if a particular value exists in a subarray in a document? Kind of like $in, but switched around
[19:25:56] <GothAlice> There are caveats if the nested values are complex objects (nested documents) instead of bare simple values (strings, integers, etc). Notably that when comparing documents the order of the keys is vitally important.
[19:27:43] <GothAlice> ^_^ 9 days on my own at-work project before our 1.2 feature release… it's a doozie.
[19:28:13] <GothAlice> After that, my "ruling the world" project is scheduled for completion in September. ;)
[19:29:02] <xaxxon> so what if it is {foo: [ {key: a, otherthing:asdf}, {key: b, otherthing: fdsa} ] }? and i want to know if key: a exists, but I don't know otherthing
[19:29:32] <GothAlice> Zelest: https://gist.github.com/amcgregor/07ded9226bfce1e7d5d4 < we're using distributed processing pipelines defined using a custom declarative schema. :3
[19:30:34] <GothAlice> Please keep language in-channel clean, thanks. (It's permanently logged… so it's not just for everyone else's sake!)
[19:33:27] <GothAlice> And, the coolest thing: $elemMatch and $ in the query projection let you get back not just the record with a "key" of "a", but lets you get back just the one "foo" element that matches.
[19:33:55] <GothAlice> See: http://docs.mongodb.org/manual/reference/operator/query/elemMatch/ and http://docs.mongodb.org/manual/reference/operator/projection/positional/#proj._S_
[19:34:42] <GothAlice> E.g. db.collection.find({"foo": {$elemMatch: {"key": "a"}}}, {"foo.$": 1}) — you'll only get back {foo: [{key: "a", otherthing: "asdf"}]}, not the full array.
[19:35:08] <GothAlice> (Likewise you can use $elemMatch in a query to update specific nested array elements.)
[19:52:06] <SixEcho> so i have a unique+sparse index on .things.name… mongo will fail correctly when dup keys on different documents, but on the same document it appends dups to the array… { _id:x; things: [ { name: 'hi' }, { name: 'hi' }] } any way to make it fail in this case? bug? wad?
[20:32:41] <abishek> could someone suggest a good tutorial to understand and master the aggregation framework in mongo, am actually new and I am right now using the group method to retrieve data which is not very performant and would like to change it to an aggregate. I am quite new to the aggregation framework
[20:40:38] <StephenLynx> just keep in mind that each stage will process the collection and create a new output to be processed by the next stage abishek
[20:42:35] <StephenLynx> and particularly, aggregation is what I use in almost any case that I query for documents.
[20:54:27] <GothAlice> I go the opposite route; standard queries are pretty much guaranteed to be faster and less server-intensive than aggregates. Thus things like http://s.webcore.io/image/142o1W3U2y0x tend to be a hybrid of both approaches.
[20:55:08] <GothAlice> (Aggregates are only used on the analytics sections of each detail view, and as the fundamental approach to reporting. Absolutely everything else uses standard queries.)
[21:06:26] <bros> I have a collection called orders. Each document has a subdocument called shipments. I have an existing aggregation going over orders, so I don't want to unwind shipments and spoil this aggregation, but shipments has a boolean field called voided.
[21:06:46] <bros> How do I include the size of shipments in the aggregation, but only the non-voided shipments? Do I have to unwind?
[21:41:25] <bros> I stored a float as a string accidentally. How can I cast it back to a float in an aggregation?
[22:52:53] <godzirra> So this is not a geojson point, but i'm not sure what's wrong with it: > db.incidents.find( { loc : { $near : { $geometry: { type: 'Point', coordinates: [ 36 , -115 ]}, $maxDistance: 10 } }})
[22:58:23] <StephenLynx> GothAlice so, which one generally is faster? a find, sort,skip,limit and toArray on a cursor or an aggregate that does the sort,skip and limit?
[23:05:02] <DrTachyon> Okay, so I have this query that runs perfectly well on a random db with dummy data
[23:05:42] <DrTachyon> The only difference is that there was tons of dummy data, while there's only two live entries.
[23:08:12] <StephenLynx> I don't know about it, just that is the main cause of people coming here with stuff not working. I work with io.js and the node.js driver.
[23:08:44] <DrTachyon> It's not Mongoose, since the query runs fine with dummy data.. I know, but you're like my last hope.. humor me..
[23:09:12] <DrTachyon> I have a collection, like a ledger, where the documents are transaction records
[23:11:12] <StephenLynx> but generally I can't see anything wrong regarding js. you pass a callback in exec and it will give you the error and the data.
[23:13:08] <DrTachyon> given "user" the query is supposed to match documents where "user" is either in the "debitAccount" or "creditAccount" key. It projects the "amount" value with the condition that if the "user" appears in "debitAccount", it multiplies the "amount" by -1. finally it groups the amount into a "total", and then executes a callback.
[23:15:43] <DrTachyon> Also, i'm going to connect through the console, to eliminate mongoose problems! :D
[23:16:16] <StephenLynx> if you are getting stuff from one deploy but not from the other, mongoose seems to not be the culprit in this case
[23:30:38] <godzirra> So this is not a geojson point, but i'm not sure what's wrong with it. Can anyone give some feedback? db.incidents.find( { loc : { $near : { $geometry: { type: 'Point', coordinates: [ 36 , -115 ]}, $maxDistance: 10 } }})