Log file Viewer
#mongodb logs for Friday the 21st of December, 2012
[20:30:48] <Smarknov> i would like that the client could be away but his "block-caused" pending query would still be executed when its unlocked
[20:59:37] <SegFaultAX> I'm working on building out a real-time analytics system using mongodb as the datastore and I have some general high level questions.
[20:59:57] <SegFaultAX> 1) Is there an open source system that I can use as a reference implementation for this sort of thing?
[21:01:03] <SegFaultAX> 2) What are some common ways to shard time series data? By object (eg users, posts, etc.) or by granularity (minutely, hourly, daily), or perhaps by some combination thereof?
[21:01:24] <SegFaultAX> 3) Are there any technical downsides to creating lots of collections in mongodb?
[21:03:23] <meghan> SegFaultAX you may want to check out http://www.10gen.com/use-case/business-intelligence or http://www.10gen.com/presentations/mongosv-2012/mongodb-analytics-github or http://docs.mongodb.org/manual/use-cases/pre-aggregated-reports/
[21:15:38] <weeb1e> I cannot get this ruby driver to get mongodb to unset 2 keys that were set on one or two documents
[21:16:13] <weeb1e> Which due to the keys containing "-" are breaking the ODM which in turn is breaking all my applications using these databases
[21:17:06] <weeb1e> ModelName.collection.update({}, { '$unset' => { 'screenshot-name' => 1 } }, multi: true)
[21:18:14] <weeb1e> Nevermind, it seems the ODM was caching keys somewhere and the query did work, probably long ago :|
[21:59:08] <tracker1> bit of an issue.. I have the following index... {"fieldname":-1,"field2":-1} ... but to preserve order from node, I saw somewhere using array syntax for the sort... [["field1":-1],["field2":-1]] should I define the index that way? when I pass the sort in via array syntax it doesn't sort right, but the order isn't preserved via node's object...
[22:15:24] <Arnout> Hey, quick question since the docs aren't clear about this. When using db.setProfilingLevel from the shell, does it operate on the current 'used' database? (ie if I enter 'use mydatabase' before, does it operate hte profiling on that only?)
[22:16:16] <Arnout> Reason I am asking is because it looked like profiling didn't work on our cluster, but looking at the UI of mongoqp it seems like you can change profiling levels per-database, not per server
[22:30:35] <fommil> Hi all – I am thinking of using a sort in an aggregation but I'm wondering if it's the most efficient way to meet my objectives. What I really want to do is to group documents by some key and then pluck out the document (for each group) that has the highest value of another field. What is the best way to do this?