Log file Viewer
#mongodb logs for Wednesday the 20th of March, 2013
[01:04:52] <fredix> I try to compile my app with mongodb 2.4 but I get this error external/mongodb/src/mongo/base/status.h:21: erreur : mongo/base/error_codes.h: No such file or directory
[01:28:26] <fredix> jzk: in fact scons generate the file : /usr/bin/python src/mongo/base/generate_error_codes.py src/mongo/base/error_codes.err build/mongo/base/error_codes.h build/mongo/base/error_codes.cpp
[01:29:07] <jzk> fredix: for me it was under the src/ directory, but now it builds into the build/ directory, so that has to be part of the include path
[02:12:30] <akamensky> Folk, i am having problem with getting a subset of embedded documents. Please check this question on SO http://stackoverflow.com/questions/15496071/finding-embedded-documents-in-mongoose-odm
[02:25:46] <bodik> it look that i cannd do a MR job on the top on sharded collection with output reduced into othe other sharded collection
[03:57:35] <akamensky> Folks, i am having problem with getting a subset of embedded documents. Please check this question on SO http://stackoverflow.com/questions/15496071/finding-embedded-documents-in-mongoose-odm
[08:43:38] <ianserlin> congrats on the newest release! looking forward to the new geospatial features
[09:48:18] <rev0> i'm using node.js +mongodb+mongoose. do you know why _id's are not generated for mixed schema types?
[10:49:21] <ron> millun: I have experience with morphia. while it's nice, it has its fault and not really under development anymore. I'd try looking at jongo as well.
[11:13:43] <millun> ron, cheers. didn't know jongo. had experience with morphia, but i figured it's dead either
[11:15:12] <millun> ok. last time i was checking jongo wasn't there. or maybe i was browsing another section of the site. maybe they assigned it to a framework category.
[12:30:00] <GeertJoh1n> Is there a convention for naming fields that contain a referencing ObjectId?
[12:30:55] <GeertJohan> I'm using userId now, but would like to stick to whats being used 'out there'..
[12:32:24] <kali> GeertJohan: the close thing to this is the dbref convention http://docs.mongodb.org/manual/applications/database-references/
[12:33:14] <kali> GeertJohan: but for the rest... you'd better avoid the "user-id" variant, it will drive you crazy when you'll need to query the database in the mongodb shell
[12:33:57] <GeertJohan> DBRef is more about the value of the field, but definitly worth checking out.. Now using manual references..
[12:47:12] <GeertJohan> Hmm, a small note DBRef support. The Go driver for MongoDB (named: 'mgo') does support DBRef: http://godoc.org/labix.org/v2/mgo#DBRef
[12:52:26] <CupOfCocoa> Hey guys, does anybody know how to dump multiple dbs at once with mongodump?
[14:42:16] <underguiz> hi, is it safe to run a mongos 2.4 instance with the --upgrade option while having mongod and mongos on the same sharded cluster?
[14:44:26] <underguiz> the documentation says "The upgrade to MongoDB 2.4 adds epochs to the meta-data for all collections and chunks in the existing cluster. MongoDB 2.2 processes are capable of handling epochs, even though 2.2 did not require them."
[14:49:01] <kali> underguiz: mongos 2.4 will refuse to start if you haven't run the upgrade before (unless it's the one running with --upgrade, obviously)
[14:50:32] <underguiz> but I'll keep the mongod proccess running for a while until I can find a safe upgrade window
[14:53:34] <underguiz> there's also this section of the doc: http://docs.mongodb.org/manual/release-notes/2.4-upgrade/#rolling-upgrade-limitation-for-2-2-0-deployments-running-with-auth-enabled
[14:53:47] <underguiz> That says "MongoDB cannot support deployments that mix 2.2.0 and 2.4.0, or greater, components. MongoDB version 2.2.1 and later processes can exist in mixed deployments with 2.4-series processes. Therefore you cannot perform a rolling upgrade from MongoDB 2.2.0 to MongoDB 2.4.0. To upgrade a cluster with 2.2.0 components, use one of the following procedures."
[14:54:21] <underguiz> if I understood it correctly, I can't upgrade it, since my mongod processes are all running 2.2.0
[14:58:01] <underguiz> Nodex, thanks, is there an option on mongos 2.4 to run in some kind of compatibility mode (like it was 2.2)?
[15:03:27] <GeertJohan> I have a collection containing documents that have a timestamp, reference_id and other fields.. What would be the best approach to get the latest document (based on the timestamp) for each phone_id?
[15:05:15] <GeertJohan> I could probably use aggregation/pipes with $group. But won't that leave me with just the fields I have named in the $group properties? Seems like a wrong approach...
[15:15:31] <Codewaffle> GeertJohan, If you figure it out, let me know - I'm sort of doing the same thing and $grouping feels clunky
[15:17:16] <GeertJohan> indeed, it seems that with $group you need to name each field.. I don't want to do that..
[15:39:49] <saml> i have an object, how can I create a new database /collection and insert the object?
[17:16:47] <lbjay> question about authentication: I can add users *before* enabling the --auth=true option, right?
[18:28:27] <Mmike> Starting database: mongodbstart-stop-daemon: unrecognized option '--interleave=all'
[18:29:51] <azylman> What happens if I have an expireAfterSeconds index on a field in a collection and some of the documents in that collection don't have that field? Will they just not ever expire, or could worse things happen?
[18:33:13] <bean> ah, can't really help there, only really have used it on ubuntu, which apparently has a different startup script
[18:34:03] <Mmike> But this is so serious issue that I'm having doubt's in the rest of mongodb, if a installer can't be fixed for so long
[18:38:23] <kali> Mmike: numa is highly problematic for mongodb, it is explicitely said in the production notes
[18:40:32] <kali> Mmike: the problem is numa and mongodb are not working well together... there is not a big interest in making the installer works as the real stuff will not be working anyway
[18:41:05] <Mmike> kali, what you're saying is - mongodb is not supported on debian, unless you patch the installer yourself?
[18:41:39] <Mmike> btw, I have several mongodb clusters that work pretty fine, once I applied patch to the startup script
[18:43:47] <Mmike> on Feb 25 14:29:34 [initandlisten] ** WARNING: You are running on a NUMA machine.
[18:43:48] <Mmike> Mon Feb 25 14:29:34 [initandlisten] ** We suggest launching mongod like this to avoid performance problems:
[18:43:48] <Mmike> Mon Feb 25 14:29:34 [initandlisten] ** numactl --interleave=all mongod [other options]
[18:45:31] <Mmike> what you're saying is 'there is low interest in getting mongodb to run on debian'
[18:49:17] <bobbytek> Should one be able to connect to a mongo db even if you are not authenticated?
[18:49:59] <ron> one should be able to do so unless one turned on mongodb's authentication mechanism.
[18:51:04] <bobbytek> we enabled auth but we can still connect via the shell, but cannot do anything
[19:01:15] <fjay> when a replica set is a member of a sharded database.. what interaction is there with the config servers?
[19:09:57] <fjay> i do see one of my shard replica set members having 6 open connections.. 2 to each config server..
[19:11:02] <fjay> trying to troubleshoot some connection reset by peer errors between replica set members.. thus causing random elections that should not happen
[19:12:32] <kali> ok. i've seen this kind of thing on ec2... the cause was changes in security group were sometimes making messing with firewalling rules they had no reason to mess for a short amount of time
[19:13:14] <kali> enough to be picked up as node failure by mongodb replica, not enough for us to understand what was happening for quite some time :)
[19:14:44] <fjay> one theory someone else has.. (which i call shenanigans on) is that we have too many mongos running and its causing contention w/ the config servers.. hence my question about config servers earlier.. but that has NO input on master elections for the shards
[19:16:04] <fjay> this is connection reset by peer between two members of the replica set.. and getting sometransport errors in the logs from the java driver..
[20:01:54] <hydrozen> Hey, just wondering, how do people deal with super long uuids (like those generated by mongodb or couch by default) in the url? It's passable when you only have one in your url, but as soon as you have nested resources… having super long ids make URLs pretty intense.
[20:04:43] <hydrozen> I guess all there is to do is generate a slug or anything else unique and shorter..
[20:05:26] <hydrozen> ./manage/routes/c4a7d2894daf712d8304ccde73036c98/services/c4a7d2894daf712d8304ccde73037c8f/edit
[20:06:10] <ron> ah, so you're worried about the aesthetics of the URLs. you want them to be lean and skinny. use photoshop.
[20:26:20] <kgriffs> I heard somewhere that keeping doc sizes under that limit helps with performance.
[20:27:16] <kgriffs> http://blog.mongodb.org/post/40769806981/checking-disk-performance-with-the-mongoperf-utility
[21:14:18] <Sep1> Hello guys, is there any way, how to get the next and the previous document from MongoDB with using RoR+ActiveRecord?
[21:33:27] <fjay> okay.. so i am seeing a ton of connections to my primary mongod's looks like roughly 18k per shard.
[21:53:45] <yfeldblum> after a replSetInitiate on the local mongod which is the only member of the replset (i'll add peers later), and calling isMaster, it reports the local node as a secondary ... what i'd like to do is to figure out if it's currently voting for a new primary and wait for that voting to be finished ... how can that be done?
[21:57:56] <yfeldblum> fjay, that's not the problem; the problem is i need to wait until the voting is finished
[22:00:07] <yfeldblum> fjay, i'm setting up a replset with a single node, and waiting for the state to stabilize to primary, and after the state is stabilized, do some more things
[22:03:57] <yfeldblum> fjay, i mean i'm sure i could, but that will require a lot more automation work
[22:55:39] <djangoobie> Hi I'm very new to mongodb and I'm using mongoose. When defining a schema with a field to a related schema I sometimes see people using ObjectID with a ref attribute and sometimes the related schema itself. Is there a difference between the two?
[23:58:50] <nemothekid> and can I run 2.2 mongos processes in parallel (so I don't have any downtime?)