Log file Viewer
#mongodb logs for Monday the 2nd of June, 2014
[00:29:45] <Lawhater> I am trying to install mongo on my computer.. I have the zip file but I dont know where to go from there.
[00:54:59] <SirFunk> anyone know if with the mongo-org package on ubuntu if it is possible to disable the mongodb server from autostarting
[01:18:54] <proteneer_osx> root does not include the ability to insert data directly into the system.users and system.roles collections in the admin database
[05:36:24] <narutimateum> what are alternatives to join in mongo? i need the stores in here http://i.imgur.com/rFiqKbU.png to have info of these http://i.imgur.com/1ZlfLmS.png
[05:38:51] <lqez> it would be better record the ObjectId of referenced objects, not just _id integer value.
[05:41:32] <lqez> anyway, mongodb (currently) does not support any join operation(as you know), so we have to
[07:30:45] <Viesti> Hi, I was reading about bulk updates: http://blog.mongohq.com/bulk-updates-for-all/
[07:32:09] <Viesti> we'r using mongoimport at the moment, after a quick glance it seemed to me that it doesn't use the bulk import API, could someone confirm if I'm right or totally wrong :)
[07:56:41] <Ponyo> In order to get real high availability out of a 3 node replica set, do I need to run the mongos router on each node as well?
[07:59:19] <Ponyo> So when I connect to the secondaries using the mongo client and try to write, i get the error they aren't the primary, how does one use a replica set appropriately if only one can do writes at a time and the others aren't forwarding the writes to the master?
[08:00:11] <Ponyo> that is, if I can't use round robin to achieve fail over, then how does the application client know which is the master accepting writes?
[08:43:32] <rspijker> Ponyo: mongos knows how data is distributed across shards. It’s a router that does aggregation of results etc.
[08:44:04] <rspijker> in a replica set, data is not distributed. The same data is on all of the nodes (normally), so when one fails another can take it’s place without problems.
[08:44:36] <rspijker> shards is partitioning, so each shard (which is usually a replica set) contains a part of the data
[08:45:31] <rspijker> you query against the mongos and it knows where to go and get the different data. It also handles the underlying querying for you. So it will ask the relevant shards for the data, then stitch it together and return it to you as if you had just queried a single DB
[09:21:12] <AlexejK> With a single mongod, is it possible that when one of the DBs is getting heavy writes (or doing MapReduce), the reads from others are affected? I know that its Single Write-Multiple Read locks but I'm observing this problema nd trying to see if this is somewhere on Mongo side or my webapp (performance is affected even when i do an operation through console)
[09:21:45] <Ponyo> I read that a replica set can't have more than 12 members, does this include config servers? I also read that I need 3 config servers per shard, so does that mean the total max size of a replica set is actually 15 servers including it's config servers?
[09:22:23] <rspijker> Ponyo: config servers are not part of the RS, furthermore there are 3 per CLUSTER, not per shard
[09:22:50] <Ponyo> Config servers store metadata for a single sharded cluster. Each cluster must have its own config servers.
[09:22:56] <Ponyo> http://docs.mongodb.org/manual/core/sharded-cluster-config-servers/#sharding-config-server
[09:23:15] <Ponyo> someone might wanna work on the wording of that then, I find it confusing as all getout
[09:24:42] <kali> AlexejK: heavy load will affect performance, regardless the locks. for instance, if you're mapreducing a whole collection, you're impacting the cache distribution and that will degrade the read performance
[09:25:52] <AlexejK> kali: got it. So to mitigate that, what is the best thing to do? shard + high IOPS?
[09:27:15] <kali> AlexejK: then, fish for low hanging fruits: queries with missing indexes (the log will show you slow queries) or map/reduce that can be replaced by aggregation framework
[09:28:43] <kali> Ponyo: you will find a form in the low left corner to report a documentation page that you find confusing
[09:39:49] <AlexejK> kali: Thanks, we looked into that a while back but deffo worth looking into, thanks
[11:41:49] <gancl> http://stackoverflow.com/questions/23993558/node-js-mongodb-cursor-each-get-empty-cant-run-in-sequence Hi! How to let mongodb cursor.each(function (err, docs) run in sequence?
[12:29:58] <kali> gancl: not sure i understand the question, and the code formatting is off in your question
[12:51:49] <Repox> Hi guys. I'm trying to add a replica with rs.add(). Isn't it possible to add the username and password for authentication?
[13:07:51] <jackw411> anybody got a guide for mongo on ec2 and being able to connect remotely? having difficulties!
[13:37:50] <hactor> Folks if you have an AGPL source, and it's interfaced to an Apache License source, and the latter is interfaced to you code, aren't you still bound by the AGPL?
[13:41:37] <rspijker> hactor: that’s fairly off-topic… not that I mind particularly, you’re just not very ;ikely to get a proper answer
[13:54:40] <Ephexeve> Hey guys, I wonder what's the best way for me to check if the user exists in mongoengine before loggin in? Is it like this? http://docs.mongoengine.org/tutorial.html#accessing-our-data
[14:00:04] <hactor> rspijker, it's the scheme used by MongoDB, I don't think it's been proven in court.
[14:09:12] <jackw411> is there anything i need to do to make mongo accessible externally on a port on a server?
[14:09:39] <hactor> However from the AGPL it follows the drivers are bound under the terms of the AGPL.
[14:10:07] <kali> jackw411: check out 1/ the firewalling rules in your security groups, 2/ the presence of a "listen" directive in your mongo configuration
[15:44:13] <Repox> Hello. Some input on the error I get here? http://pastie.org/private/9c2n6txumvv6otc85jgmjq - I really can't get it to work.
[15:55:21] <rspijker> and you might have some timing issues across servers or something causing the rollback to fail
[17:12:49] <kali> mikehelmick: http://docs.mongodb.org/manual/reference/operator/aggregation/out/#behaviors
[17:22:00] <mikehelmick> kali: When aggregating a collection — if I try to aggrgate it again it returns a different result
[17:25:23] <mikehelmick> kali: I have 1 document, and 30 subdocuments in that document.. when i limit the result of the aggregation, then do..
[17:26:07] <mikehelmick> That cursor only has the number of subdocuments that I previously limited in the aggregation
[17:35:30] <mikehelmick> kali: https://gist.github.com/michaelhelmick/c732d68af1f1f1b356e9 I've updated the gist
[17:52:05] <kali> yes. http://docs.mongodb.org/manual/tutorial/manage-the-database-profiler/#profiling-levels
[17:52:59] <kali> mikehelmick: i suggest you run a count() from the shell, not for python, run the script, and check the count() again from the shell
[17:59:20] <modcure> I setup a test shared environment... there is data on shared-a but not shared-b.. not sure why shared-b has no data in it. thoughts?
[18:02:08] <kali> run a db.collection.stats() and .getShardDistribution() and gist the output somewhere
[18:11:30] <modcure> kali, i did this early.. sh.shardCollection("cloud-docs.spreadsheets", {username: 1, _id: 1})
[18:13:49] <kali> ok, are you in the right database when you run db.spreadsheets.stats() ? not in the "config" database ?
[18:17:23] <kali> modcure: your collection is too small. you'll start to see things happening when it reach the chunk size (it should be 64MB)
[18:20:30] <modcure> kali, about the sharding... mongodb uses the shared key to perform a range paritioning... about the chunks.. each chunk is consider a new partition?
[18:39:00] <Repox> When I try to do rs.addArb() I get the message "replSetReconfig command must be sent to the current replica set primary." What does this mean and how do I fix it?
[18:47:09] <synth_> one of our servers crashed that had a mongodb set up on it, is there an easy way to import all data from the old database which we have on another drive to a new installation of mongo?
[18:52:40] <joshua> synth_: the mongorestore command has a --dbpath option to import from a data directory
[19:12:54] <MacWinner> hi, had a quick question regaridn the steps outlined here: http://docs.mongodb.org/manual/tutorial/convert-standalone-to-replica-set/
[19:13:30] <MacWinner> in the section, "Expand the replica set", step 1, it says start 2 standalone instances.. do we need to start these standalone instances with the --replSet flag?
[19:21:31] <MacWinner> if the initial server has 500M of data in its collections, and you add the second item in the replica set, will the primary node be stalled until the replication is complete?
[19:23:52] <joshua> You should be able to continue using it while it syncs the data, but if you want to speed it up you can possibly copy the data directory before you start.
[20:00:21] <cpennington> what's the best way to use a c-backed version of the SON class as the document_type argument when connecting using PyMongo?
[20:00:43] <cpennington> the SON class you import from bson/son.py is implemented in python, and has a number of things that slow it down significantly
[20:17:55] <smatt706> Hello, I'm trying to adjust a centos/rhel 6 to increase the ulimit for open files and processes. I edited /etc/security/limits.d/90-nproc.conf per http://docs.mongodb.org/manual/reference/ulimit/#unix-ulimit-settings however this doesn't seem to get read unless there is also an adjustment in /etc/security/limits.conf