Log file Viewer
#mongodb logs for Friday the 7th of September, 2012
[00:48:29] <crudson> fg3: I presume you can have multiple categories and multiple items in each category
[00:49:54] <crudson> fg3: to update a specific matched item in an array use the http://www.mongodb.org/display/DOCS/Updating#Updating-The%24positionaloperator
[00:50:36] <crudson> fg3: although I am not sure you can use it in an array nested within an array, that would be something to experiment
[00:52:28] <fg3> I lost the stack-exchange page where they chat about this -- I'm wondering if I have to change my structure
[00:54:43] <Guest55811> whats the cookie cutter use case example to choose mongodb over like mysql or postgresql
[01:00:46] <crudson> fg3: yeah, either fetch, change, save, or do in a very small db.eval() - both of course should be considered v carefully depending on the mutability/load of your db. Otherwise document restructure should be considered to make querying and updating easy, fast and atomic.
[01:21:48] <SpanishMagician> is it possible to sort result based on an element inside an embedded document?
[01:22:53] <SpanishMagician> ex I have a colletion of users and inside of each user there's an embedded document comments
[03:00:50] <phatduckk> hey all - quick question. In the logs does "locks(micros) r:63" mean LOCKED for 63 micros or WAITED on a lock for 63 micros?
[05:22:42] <_Tristan> Hello. How would I go about removing all documents from a collection where "expires" (a date objectything) is before now?
[07:02:15] <abhishek__> can someone explain me reason for this : https://groups.google.com/forum/?fromgroups=#!topic/mongodb-user/JddGQ8ooUcw
[08:00:41] <SpanishMagician> is it possible to sort result based on an element inside an embedded document?
[08:00:57] <SpanishMagician> ex I have a collection of users and inside of each user there's an embedded document comments
[08:20:17] <samurai2> what's actually sh.setBalancerState(false) or disable the balancer do actually? thanks :)
[08:28:18] <gigo1980> hi, i get an timeout in my mongo cluster some times, what can be the reasons ? i dont compact or repair the collections during this time
[08:33:36] <Axsuul> Not sure if I recall correctly, but was mongodb's global lock addressed recently?
[08:38:59] <Axsuul> kali: thanks, how do you guys get around heavy read/write applications involving mongo?
[09:30:09] <Albeneth> Hi. How do i do this (http://pastebin.com/R5RXXVZ3) in mongodb (with the same collections and fields names) ?
[09:40:20] <Gizmo_x> anyone have expirience with doctrinemongoODM reference query http://pastie.org/4678732 - self reference many to many
[09:41:20] <Gargoyle> Albeneth: Mongo doesn't have the concept of JOINs and for GROUP BY, you'll be wanting to checkout the new aggregation stuff in 2.2.
[09:49:43] <Gizmo_x> anyone have expirience with doctrinemongoODM reference query http://pastie.org/4678732 - self reference many to many
[09:53:42] <Gizmo_x> need help here http://stackoverflow.com/questions/12243929/mongodb-odm-docrine-self-reference-many-to-many-get-specific-data
[09:54:42] <kali> Vile: not efficiently. you can do it with application code, or by server side javascript (avoid that in prod)
[09:55:18] <Gargoyle> Vile: You mean update({someProp: 'foo'}, {$set: {someProp: 'bar'}}, false, true) type of thing. Don't think that's allowed.
[09:55:28] <Vile> basically i have a problem. i have large database with field format as http://pastebin.com/SxVySJNQ and I need to update _id.t fields
[09:57:17] <kali> The _id field is immutable. If you try to change its value on an update you will get a "Mod on _id not allowed" error from getLastError.
[10:02:26] <Vile> Gargoyle: basically, i need to take huge collection of records as in the paste and modify _id.t field
[10:03:43] <Gargoyle> Vile: You'll have to re-insert the while doc and delete the old ones. And have a think about how you have come to need to change your _id's !
[10:04:13] <stevie-bash> Hello, I want to run logrotate on the router, since kill -USR1 has a deadlock bug.
[10:04:38] <stevie-bash> can I send the echo also the the mongos process? Will it logrotate only the router?
[10:04:39] <Vile> there was a bug in external program that was reporting wrong values for _id.t field
[10:34:21] <jwilliams> for insert with continuous on error, is it required to have driver also support it? Or it would be default on at server side regardless of driver specifying "continuous on error" or not?
[10:50:18] <Vile> Guess I'll write a function that does find() and then iterates over cursor doing inserts into other collection
[12:25:31] <Gizmo_x> need help here http://stackoverflow.com/questions/12243929/mongodb-odm-docrine-self-reference-many-to-many-get-specific-data
[13:09:58] <gigo1980> hi all, is it posible to run only one map reduce action per collection at a time ?
[13:12:32] <Derick> for the 2nd valgrind to be useful, could you also export "export ZEND_DONT_UNLOAD_MODULES=1" ?
[13:15:38] <gigo1980> @jmar777 so the thing is i have an 24 core system where my mongo db is running
[13:18:45] <Gargoyle> Sorry Derick, Phone went just as you pinged back. I noticed still getting some segfaults again in my apache! :(
[13:20:53] <jmar777> gigo1980: hmm... i would think yes, but only with regards to m/r, since that can operate over individual shards in parallel
[13:21:03] <Gargoyle> Derick: http://pastie.org/private/lwzmyiwghhii42eytqbkbw - Is that berrer? Looks the same to me! ;-)
[13:21:12] <jmar777> gigo1980: for all other queries, you'd have to test it, but mongo already utilizes multiple cores pretty well
[13:27:56] <Vile> how can i check progress for the command like db.eval("function () { db.histories.find().forEach( function(v) { v._id.t = new Date( v._id.t.getTime() - 28800000 ); db.histories2.insert( v ); })}"); ?
[13:40:58] <Gargoyle> never been in the situation where I needed to wring every last bit of performance, and apache has always been stable! (And damn easy to install!)
[13:41:29] <NodeX> It's just weird because aot of people use Mongo for the speed, then throttle it with apache
[13:42:16] <Gargoyle> NodeX: Yeah, I'm aware that other webservers have been gaining ground over the last few years, not kept up with the progress properly.
[13:44:01] <NodeX> Gargoyle : what other things does the app do at the time of segfault ... i/e are you creating images or anything?
[13:44:30] <Gargoyle> Just looking to see if I can narrow it down to a single request like I managed to the other day.
[13:48:21] <Gargoyle> Slightly concerning that 5.3.12/13/14 are security updates, and 3 months down the line ubuntu is still not got them!
[13:49:33] <Gargoyle> Might schedule myself some time away from clients and do some digging and messing with other distros.
[13:50:09] <Gargoyle> Kept with ubuntu since 06.04 ish, just because of the sheer ease of installation.
[13:50:58] <Gargoyle> Slightly off-topic, what linux distros do people here use for desktop dev work?
[13:52:27] <Gargoyle> Again, been a while since I had a good play with a desktop linux. What's Mesa3D ?
[13:55:58] <Gargoyle> Just that latest vmware finally adds hardware support for 3D under linux, so it could yeild a more useable dektop under VM.
[14:44:23] <Gizmo_x> need help here http://stackoverflow.com/questions/12243929/mongodb-odm-docrine-self-reference-many-to-many-get-specific-data
[14:46:07] <Gizmo_x> Derick: how can i query a reference many to many as here: https://groups.google.com/forum/?fromgroups=#!topic/doctrine-user/tG_jHF8XdEw
[16:08:16] <Gargoyle> NodeX: This could be the longest drawn out conversation ever! But what do you use with php instead of apache?
[16:30:17] <veesahni> from the docs: "You may use sort() to return data in order without an index if the data set to be returned is small" .. this doesn't sound right.. limit() creates a small returned data set, but sorting needs to happen on the pre-limited() potentially large set.. am I missing something?
[16:34:05] <joshua> I have a shard cluster built, now wheres the best docs to start a 3 server configserver cluster... do you just start up 3 instances and point mongos to it, or is there more than that?
[18:52:37] <boll> Has anyone actually deployed 2.2.0 to production, and if so, how's it working out?
[18:56:42] <dstorrs> boll: I was actually coming here to ask about the deploy process. I am deploying it to prod right now. :>
[18:57:42] <boll> I deployed to staging the day it came out. It was a simple matter of upgrading the replica set one replica at a time
[19:01:05] <boll> We have quite a lot of writing going on, so I am quite excited about the global lock mostly being gone on writes.
[19:11:27] <dstorrs> boll: when you deployed it, did you just do a 'apt-get install ' ? or did you do manual installs?
[19:12:42] <boll> If memory serves, 2.2.0 successfully joined the 2.0.x instances in the running replica set
[19:20:52] <camonz> hi, is it posible to store and query a complex object, such an array of hashes inside a document? ex. https://gist.github.com/f16f02b496ce10b99ef3#file_new_format.json
[19:21:26] <camonz> also, i'm changing the keys to be information by itself instead of simply an accesor to a value
[23:09:54] <toddnine> What's the response to an insert operation? Wireshark doesn't show any response at all
[23:12:13] <wereHamster> if you want to make sure the write made it to disk you have to use getLastError
[23:14:44] <wereHamster> it's the clients responsibility to make sure the write succeeded. The server doesn't care.