Log file Viewer

[14:20:20] <ramnes> but your object is smaller

[14:20:38] <ramnes> and since you're going to have billions of this object, this is important

[14:28:56] <ramnes> your solution gives this: http://pastebin.com/Dggg90pY

[14:29:55] <ramnes> mine gives this: http://pastebin.com/0KkHaSAR

[14:30:28] <ramnes> now multiply this 9999999999 times

[14:30:30] <ramnes> :)

[14:31:45] <ramnes> but like I said, I'm not sure this is a good way to do, but I can't see why not

[14:31:54] <kali> one document per message ? really ?

[14:34:56] <ramnes> kali: how would you do? :)

[14:36:53] <kali> i would stack the message for one channel during... let's say 1 hour in one document

[14:37:55] <kali> at least, showing the state of a channel with some history is a local operation, not something accessing dozens of documents

[14:40:36] <ramnes> buffering data in a document array until document size < MAX_DOCUMENT_SIZE, ok, but then how do you query your data?

[14:42:09] <ramnes> and this way your document doesn't represent anything (a buffer, with an id?)

[14:43:30] <ramnes> s/</>/

[14:44:47] <kali> it does represent a one-hour-long log of a channel

[14:45:05] <ramnes> :D

[14:45:33] <kali> as to how to query... what queries do you have in mind ?

[14:50:44] <ramnes> your solution would be like this, am I right? http://pastebin.com/6BZ6AqNx

[14:53:46] <ramnes> if yes, how do you get every message of an author?

[15:06:47] <M-ou-se> kali: pretty much all queries will be for all messages between two time stamps of a specific user, and usually also for a specific channel

[15:07:17] <M-ou-se> kali: I probably also want to tag certain messages (such as highlights), which should be quick to query for

[15:37:51] <kali> ramnes: something like that, with a timestamp

[15:40:31] <kali> "all message between two time stamps of a specific user"... i would not have assumed that one to be really important. with my proposal, you need an index on messages.author + timespan of the main document, and process the docs with the aggregation framework

[15:51:39] <ramnes> it would be way slower

[17:34:28] <gaspaio> hi all

[17:35:05] <gaspaio> Is there a way to handle document level locks in mongodb+nodejs ?

[17:36:56] <gaspaio> i need to find a document and then update it with values resulting from some complicated computations (therefore i can't use findAndModify). how can I do this without risking a current update changing the doc in the meantime ?

[17:40:04] <kali> gaspaio: optimistic locking...

[17:40:43] <kali> gaspaio: http://docs.mongodb.org/manual/tutorial/isolate-sequence-of-operations/#update-if-current

[17:42:16] <gaspaio> cool, thanks.

[17:42:31] <gaspaio> kali:thx

[17:50:29] <gaspaio> kali: i'll use a lastModified field to check on concurrent updates, that should work.

[17:55:26] <kali> yep

[19:03:28] <sirious> trying to perform a removeShard, wondering why it would tell me that "localhost" is one of the dbs to move?

[19:45:51] <cbradley> Hey all... I have a quick question about comparing ObjectIDs using node

[19:46:55] <cbradley> From what I understand, I can compare gt/lt by using the first 4 bytes and the last 3 bytes, representing a timestamp in seconds since epoch and a counter, relatively speaking

[19:47:50] <cbradley> Just want to make sure that I'm doing this right in the javascript world...

[19:48:13] <cbradley> parseInt(oid.substr(0,8)) * 1000 // timestamp in ms

[19:48:49] <cbradley> parseInt(oid.substr(18, 6)) // incremental counter

[19:49:02] <cbradley> whoops, both of those should have a radix of 16

[20:24:48] <sirious> is there any way to fix the issue where you try to remove a shard but it has a jumbochunk and it won't move?

[20:24:54] <sirious> i tried bumping the chunksize but that didn't help

[21:22:36] <sirious> got it. disabling the balancer on that collection, temporarily bumping the chunks size, manually moving the chunk, resetting the chunk size, and re-enabling the balancer did the trick

[22:33:02] <Fandekasp> hi there

[22:33:44] <Fandekasp> What is the most simple way to join 2 cursors in mongo? I have 2 variables containing the results of a Collection.find(), and would like to join them and fetch the result

[23:43:17] <amitprakash> Hi, I am unable to start mognodb as a service http://pastebin.com/UGaghE6P , how do I resolve this?

[23:43:36] <amitprakash> I can run the same command as root w/o any issues?

[23:43:44] <amitprakash> s/?/.