Log file Viewer
#mongodb logs for Friday the 16th of May, 2014
[02:34:50] <packagex> I am new to mongodb and need some expert suggestion or recommendation. My question is what is recommended: 1) A HUGE collection with lots of documents in it or 2) MULTIPLE collections with documents. In both the approach, my documents will be similar (saving event logs)
[03:24:56] <ranman> daidoji: http://api.mongodb.org/python/current/examples/bulk.html#ordered-bulk-write-operations 2nd example in
[03:26:12] <ranman> daidoji: or here http://api.mongodb.org/python/current/examples/bulk.html#unordered-bulk-write-operations
[06:57:29] <zhaoyulong> hello, can anybody tell me what if the new selected master in a replset has a small local time than the previous master?
[06:58:00] <zhaoyulong> hello, can anybody tell me what will happend if the new selected master in a replset has a small local time than the previous master?
[07:28:50] <zhaoyulong> @joannac what I means is: the newly selected master's local time is behind the preivous master
[07:28:50] <zhaoyulong> @joannac, when the new master was a slave, it get a row A, timestamp is {1000, 9}, then it becomes the new master, when some one insert a row, but it's time is behind the previous master, the new inserted row has a timestamp {900, 1}
[07:34:43] <king1989> i intend create a vm with 24Gb ram, 8vCPU and 500GB HDD ( RAID 10) on vmware esxi
[07:45:56] <zhaoyulong> @joannac, so the problem in indeed exists, what I want to know is how mongo do with this issule when sync between replset members, you know, time can't be excactly same between servers, differs always exist
[10:19:33] <richwestcoast> anyone know how i can fix this issue : https://pbs.twimg.com/media/Bnv3FocIIAASEOY.png:large
[12:16:26] <Pet0r> I'm using the latest PECL Mongo driver in PHP (1.5.2) and I'm getting this error trying to connect to my Mongo server - "Failed to connect to: localhost:27017: get_server_flags: got unknown node type"
[12:21:12] <Pet0r> it was connecting to the old master which is now been pulled out of the replica set
[12:52:18] <Derick> but yes, the index is used for finding things, and you're finding something to remove it
[14:53:18] <jiffe98> ladt time I tried it kept starting over and over and never finished, ended up going the rsync route
[15:14:10] <jiffe98> we replaced a drive in one of them which caused an IO problem and now replication is broken
[15:36:15] <agenteo> hi, when I run a find on a mongo client, is there a way to get an unescaped result? I currently see all my “ escaped. thanks
[15:53:44] <agenteo> ok I bypassed that by trying to export a query from mongo console to text file passing a .js file with the following query: db.collection.find({"article_type": "Article", "body": /<table/}, {"body": 1}).sort({"source_id_string": -1}).limit(1)
[15:55:11] <agenteo> the result is a json of a query I think… any idea why I am seeing that instead of the expected working result I see in the interactive console?
[16:53:35] <unitclick> Hey guys. I'm trying to use mongoose's findByIdAndUpdate to update a document with a complete JSON object, not a single parameter as seen in the documents. Is it possible? If so what's best practice for doing it?
[17:14:54] <daidoji> basically, in the pymongo documentation it seems to say it'll return a list of errors on a BulkWrite operation
[17:18:21] <skot> The shell docs you are reading are for a different thing, read this: http://docs.mongodb.org/manual/reference/method/Bulk/#Bulk
[17:19:12] <skot> The bulk inserts being referred to in your link are not the new bulk api commands, but the old bulk insert method which only returns a single error and doesn't use the new commands/api.
[17:21:11] <skot> feel free to click the feedback icon and leave a "this is confusing in the light of the new bulk api/commands"
[17:41:50] <daidoji> cause I'm not really sure how to run this code http://api.mongodb.org/python/current/examples/bulk.html
[17:42:14] <daidoji> like its not working for me because a lot of those methods they use here don't look like they're available to my MongoClient
[17:43:53] <daidoji> and it looks different than the API page http://api.mongodb.org/python/current/api/pymongo/bulk.html
[17:53:17] <daidoji> also, if I pass in a generator with insert, is there any way to "execute" every so often automagically?
[18:23:03] <daidoji> so is it kosher to blk_handle.insert( record_from_generator) and then call blk_handle.execute() once at the end?
[18:23:33] <daidoji> and the glories of memory savings in the generator will be realized through the driver, or do I have to manage that all msyelf?
[19:07:13] <ddod> Open question to anyone with opinions: If I were building something like a blogging platform, would it be better to store posts in their own collection as individual records or throw them in an array inside the user records?
[19:13:45] <houms> good day, I installed mongodb from 10gen repo on centos 6. we installed version 2.6.1-2, but for some reason mongod cannot be stopped using the service mongod stop
[19:16:20] <houms> possible this line killproc -p "$PIDFILE" -d 300 /usr/bin/mongod as the rm -f of subsys/mongod does seem to get run
[19:32:14] <BillCriswell> I'm semi familiar with MySQL, a little less with Mongo. I am wonder when would be a good time to store a field as an array of things. The idea feels awesome but wouldn't querying based on something like that be very slow?
[19:33:27] <BillCriswell> Like, say a task hasMany note, an array of "note" can be useful since I rarely have to get the notes outside of the task, but if I needed to ever query "notes" would that be best as it's own collection?
[19:34:23] <BillCriswell> cheeser: So I'm better off thinking about it as like relational for the most part?
[19:40:21] <BillCriswell> A record being a "document" feels weird to me right now. Sure if I read more I'll get it.
[20:41:57] <dgarstang> What does this mean when starting mongo? "Error parsing INI config file: unknown option nojournal". There are -ZERO- references via google
[20:49:16] <proteneer> we're thinking of using r2.large for the replicas, and something cheap like an m3.medium for an arbiter
[21:02:19] <tscanausa> depends on your use case. my application timing is everything so on the most critical items we have local ssds everything else is ebs
[22:01:30] <toddwildey> new Schema({..., array: [{ _id: { type: Schema.ObjectId, ref: 'OtherModel' }, num: { type: Number, default: 0} }], ...});
[22:02:39] <toddwildey> That would be fine, except I have no way of re-attaching those documents together
[22:04:12] <toddwildey> Is there any way to guarantee a tuple? Or do I need to change the way my Schemas work?
[22:19:18] <proteneer> {standard input}:37615: Warning: end of file not at end of a line; newline inserted
[22:19:25] <Guest36106> hi, i have a question.. i have about 1m documents in my collection, and i'm searching on a key 'statskey' which takes 2seconds even thoughi have an index on it {'statskey':1}
[22:27:54] <proteneer> and if I don't do sharding, then multiple CPU threads are not all that useful
[22:28:44] <toddwildey> As long as Mongo still has a global write lock, there is such thing as too much
[22:32:01] <proteneer> in a replicatSet, do I need to issue things like ensureIndex() to all the replicas?