Log file Viewer

[00:19:47] <looopy> are there any performance downsides to using gridfs for say....images?

[00:22:14] <skot> Nothing special for images. It all depends on how you server them up...

[00:24:17] <looopy> skot: ok....i'm just allowing users to add a profile image and any additional shots they want...it's for a project where people manage their fitness level so i'm not expecting anything too huge. though i guess if it does occur then i'd move images to another server all together...amazon or somewhere

[00:31:30] <skot> yeah, or just stick a CDN in-front.

[01:52:21] <iamjarvo> im doing this app that does calling. basic. i am going to record who is on the call and maybe other details about the call. i feel like mysql would work but mongo would be good. a document for each call sounds good in my head vs a ton of records

[01:52:31] <iamjarvo> your input is much appreciated

[01:58:10] <deoxxa> i would have thought the NSA would have its own database format

[01:58:26] <deoxxa> one where you have to fill out a ton of paperwork to perform a transaction

[04:32:53] <qawsedrf> anyone alkive

[04:34:25] <chandru_in> i'm just getting started with MongoDB. I understand that when setting up replica sets, all nodes rollback to the state of the newly elected primary when the current primary goes down. Am I correct in understanding that this will lead to a loss of data unless I ensure replication to at least 2 nodes before updates succeed?

[04:40:04] <skot> That doesn't sound right, can you describe what state you want?

[04:40:19] <skot> Only the most up to date node can be elected primary

[04:40:26] <skot> (or the avail nodes)

[04:41:48] <chandru_in> skot, isn't is possible that the most up-to-date node may not have all the updates of the primary which went down?

[04:41:58] <skot> If you want to make sure that all acknowledged/safe writes are available after any primary failure then you should do majority verified writes

[04:42:30] <skot> chandru_in, yes, but there is no rollback when the primary is elected since the new primary has the most up to date writes

[04:42:54] <chandru_in> ah yes

[04:43:20] <chandru_in> thanks

[04:43:23] <skot> When the old primary, which has those unreplicated writes (assuming your don't wait for replication with your write concern) comes back online (if it does) then it must rollback.

[04:43:58] <skot> And those unreplicated writes are written to a bson file and the local database for you to merge/apply manually

[04:44:35] <skot> http://www.mongodb.org/display/DOCS/Replica+Sets+-+Rollbacks

[04:44:55] <chandru_in> So if the original primary can be brought back up, I can pretty much avoid any data loss

[04:45:18] <skot> yes… but… at that point you have the potential for conflicting writes

[04:45:48] <skot> So depending on your application it may or may not be possible to merge that data back into the new primary state

[04:46:23] <skot> The goal of the system is to never allow multiple writes to more than one primary

[04:46:50] <skot> basically no multi-master replication and the accompanied conflict/merge resolutions required

[04:46:57] <chandru_in> thanks skot

[04:47:05] <skot> (at least no need for automated rules)

[04:47:47] <skot> also, remember that you can set the write concern and replication level for each writes, the collection, the database or server/connection

[04:48:41] <skot> ttfn

[05:00:15] <qawsedrf> ttfn?

[05:00:25] <qawsedrf> i run mongod everytime. i want to run it as a daemon, how ?

[05:01:01] <qawsedrf> on startup, like apache does

[05:30:06] <mexlex> can reduce return tuples?

[09:22:54] <carsten> whois pmxbot

[09:23:06] <carsten> who installed this stupid bot here?

[09:24:58] <Derick> carsten: what happened?

[09:25:17] <Derick> carsten: did it msg you when you got in here?

[09:26:31] <carsten> -pmxbot- PRIVACY INFORMATION: LOGGING IS ENABLED!!

[09:26:32] <carsten> -pmxbot- The following channels are logged are being logged to provide a

[09:26:32] <carsten> -pmxbot- convenient, searchable archive of conversation histories:

[09:26:32] <carsten> -pmxbot- #dcpython, #mongodb

[09:27:01] <Derick> indeed

[09:27:13] <Derick> I don't remember anybody saying anything about that

[09:27:15] <carsten> it's rude logging without explicit permission

[09:27:30] <Derick> yes, I agree

[09:29:53] <Derick> jaraco: I see you put it in here. Please remove the bot.

[09:30:36] <Derick> actually