Log file Viewer
#mongodb logs for Monday the 7th of April, 2014
[00:54:57] <tpayne> Hello. I'm trying to setup mongo to support auth. When creating a user with readWrite access that my client uses, do i create it on the admin database or on the actual database?
[01:39:08] <joannac> however it may need extra roles as well, depending if you're in a replset, sharded, and what you're doing
[01:46:11] <joannac> Guest31177: "raw - driver returns documents as bson binary Buffer objects, default:false
[01:46:21] <joannac> Guest31177: from http://mongodb.github.io/node-mongodb-native/markdown-docs/queries.html
[01:51:08] <kevbook> Thanks, im trying to evaluate should my node server just respond with buffer object and save processing
[02:31:08] <tongcx> hi guys, what database can enfore all records have specfic fields with required types?
[03:52:51] <langemann> I have this setup for my mongodb, am I far off when it comes to best practice?
[03:53:15] <langemann> Or, the best way to save some data? This is part of the MEAN-stack so I'm retrieving and updating a lot.
[03:56:14] <langemann> I originally had them all split up in 4 different collections, tying them together using a relation-key which was set as their parent objectID - better solution?
[04:00:41] <langemann> I also was thinking about just having one collection and differentiating on the category-name, but I think that would give to little room to filter.
[04:06:50] <langemann> I guess I'll have to tie the data to another data-collection using some value then.
[04:13:35] <paulkon> what's the recommended max size for a shard before another shard should be added?
[04:31:27] <joannac> For a start, it would depend on whether you were on spinning disks, SSDs, network attached storage, etc...
[04:32:05] <joannac> You shard when your hardware isn't keeping up with your load, and it's easier to get more hardware than to get better hardware
[08:08:31] <arussel> when I install the package from 10gen, I have this line: daemon --user "$MONGO_USER" $NUMACTL $mongod $OPTIONS in the start function of /etc/init.d/mongod
[08:11:55] <faved> hey, quick question. if you have a data field that you want to set a TTL on using expireAfterSeconds. is there a way that you can roll that expire, so say its set to expire after 60 seconds, js there a call i can make @ 50 seconds that will reset it to 60 ?
[08:18:30] <Nodex> if you want your writes to scale to X/Y/Z then you can add that to your query or (I think) set it up as a default on your replica set
[08:24:50] <ollivera> Nodex, Okay, so it is possible to configure MongoDB so that if my primary data center goes down. I'll have automatic election of a new primary and failover. That's my question :)
[08:33:44] <Berg> hello im trying to sort my data using pymongo i have been able to sort player profiles into alphabetical list but now i need to limit that list to 20 profiles so as to put it in a web page sortlist=db.users.find().sort(keyword, method) that is the sort method
[08:40:26] <arussel> the 'daemon' command is not installed on my machine (aws ami), which package should provide it ?
[09:13:31] <mylord> starting out, is it a decent or bad decision, generally, to prefer databases over collections? ie, have only 1 collection per database, and have about 10 dbs, for concepts like tourmanets, scores, users, winners, turnyResults, etc
[09:14:41] <mylord> initially i thought since each database has its own write lock, it might be useful to have separate databases for performance, at least while data is small and possibly mostly in RAM?
[10:42:30] <k_sze[work]> Hello. I suspect that one of the members in our replica set ran out of disk space while trying to synchronize data over.
[10:43:54] <k_sze[work]> Can I just shut down that member, add more disk space and then start that member again?
[10:54:12] <aberrios> mongoNoob here. If I have a collection "jobs" and the documents have a field "status". And I have a process that grabs all jobs that have status "waiting" and updates the job to have a field "foo" = "bar" and "status" = "active". I'm worried that as the jobs collection grows this process is just going to get slowler and slower. How would others approach this problem? I suppose in a RDB i would have a table live_jobs and once the process has run move t
[11:01:30] <slikts> is there a way to set a callback for when a specific field is changed in mongoose?
[11:02:34] <slikts> I currently check for .isNew and hash the field then, but I have to remember to do it manually when updating it
[11:45:43] <kali> mxck: it's a good idea to start with the driver before stacking layers and layers on top of it
[11:47:00] <kali> mxck: adaptation between a scripting language like python and mongodb is less painfull than other combination (sql vs anything, or mongodb with a strongly typed language)
[11:50:17] <ncls> BostjanWrk-: yes I think a script will be easier to write than a query to do that !
[12:04:21] <mylord> is there a performance benefit, potentiall, to using multiple databases instead of collections? or performance loss, typically?
[12:10:43] <mylord> database keeps ‘separate files in filesystem’ for each database, so would that affect read/write?
[12:10:56] <kali> every database comes with its own write lock, but it also comes with about at least 128MB
[12:11:35] <kali> mylord: it's write-lock per database. but unless you are planning on building something really huge, it's not going to be a problem
[12:11:50] <kali> mylord: and it's not that simple about the 128MB because they will be mapped to memory
[12:13:08] <mylord> kali, but performance wise, database over collection, if 128 * num_dbs RAM isn’t a problem?
[12:14:26] <mylord> kali, in case of file-system usage, will database be much slower than collection, or negligable slower, or?
[12:16:01] <kali> slikts: well, at least you got some understanding of what was mongodb... people going straight for the ODM are so lost when the ODM fails them...
[12:16:03] <Nodex> slikts : you really should learn the raw uery language a little before using an ORM
[12:17:19] <kali> mylord: once again, it depends if you're planning hundred of databases or 5 of them
[12:18:02] <tazz> The use exists, password is correct yet i get "Error: 18 { code: 18, ok: 0.0, errmsg: "auth fails" } at src/mongo/shell/db.js:228"
[12:18:55] <mylord> about 10-20 DBs seems to make sense then.. don’t want more than about 1-3GB ram overhead..
[12:21:29] <Nodex> the only time I hit lock contention is when I am processing large numbers of files inside the same collection
[12:22:40] <Nodex> eventually I will just move the relevant parts to a queue when I can be bothered then it's never really a problem
[12:45:45] <moas> I try to set unique field on my collection via ensure_index (pymongo) but I've obtain None
[13:15:16] <asturel> anyone can help me? i try to rs.add but i get "assertion" : "set name does not match the set name host asturel.hu:27017 expects",
[14:08:13] <joannac> asturel: 1. make sure both mongods have the same --replSet argument 2. only your primary should be initiated
[14:08:34] <joannac> asturel: if you have no data, i would wipe one, start it with the right argument, then rs.add() it
[14:12:34] <moas> joannac, 0 value is an error but the modification (set to 1 or -1) doesn't fix my problem
[14:31:16] <asturel> on the primary i get this now for the secondary on rs.status "lastHeartbeatMessage" : "rollback 2 error findcommonpoint waiting a while before trying again"
[14:32:12] <asturel> but i cant drop Mon Apr 7 16:31:43 uncaught exception: drop failed: { "errmsg" : "not master", "ok" : 0 }
[14:36:29] <asturel> bah i did it but i get "lastHeartbeatMessage" : "rollback 2 error findcommonpoint waiting a while before trying again"
[14:39:40] <amitprakash> Hi, pymongo throws a InvalidDocument: Cannot encode object: {'sd': {}} .. How do I resolve this?
[15:29:51] <ShortWave> if I just wanted to start building a driver, would it be better to A: Reverse engineer existing drivers? Or B. start from scratch.
[15:31:21] <ShortWave> Not quite a full lua driver, but more like a customized lua-based client that could speak the driver language.
[15:33:53] <cheeser> what i'm doing for my kotlin driver is wrapping around the 3.0 java driver core.
[15:34:13] <cheeser> so my kotlin driver exposes a kotliny API but delegates underneath to the java driver.
[15:34:28] <ShortWave> no matter what I do, I have to write some kind of boilerplate translation logic to go from lua to the host application's networking interfaces
[15:35:31] <skot> lua is not jvm based. it is a c-style language and there are drivers already: https://github.com/moai/luamongo
[15:37:25] <skot> There may be interpreters in many languages but the main compiler creates machine code like c/c++/etc.
[15:42:02] <cheeser> hrm. i thought lua could run on the jvm. cloudera has a java app that use lua for configs/commands.
[15:53:02] <skot> Yep, there are interpreters out there, and the main distro has one too which is machine compiled.
[15:53:41] <cheeser> makes sense. i think the unreal engine uses lua so they'd need/want that native code in any case.
[15:54:09] <skot> It does well in a "scripting" environment for things like that, and games (which is generally c/c++)
[15:58:55] <saml> Mon Apr 7 11:58:36 Error: [aggregation framework] is not a valid database name src/mongo/shell/mongo.js:40
[16:05:08] <arrty> I'm tired of reading bad things about mongo. Can the downsides of mongo be mitigated easily?
[16:13:00] <Nodex> perhaps this might be more relevant to you http://docs.mongodb.org/manual/tutorial/manage-journaling/
[17:45:35] <Danny_Joris> I added replSet = rs0 and oplogSize = 100 to my mongodb.conf file in ubuntu. Though when I log in I don’t think it’s working.
[17:45:55] <Danny_Joris> when checking db.getCollectionNames(), I get this error: { "$err" : "not master and slaveOk=false", "code" : 13435 } at src/mongo/shell/query.js:128
[17:48:04] <Danny_Joris> how do I connect to a primary. I feel on osx it did automatically. but on Ubuntu I’m dealing with service stop/start etc
[17:54:24] <Danny_Joris> one machine - I need a replica set to push it to elasticsearch using the mongodb river
[17:54:31] <Joeskyyy> Because you should have a master that has an rs.conf() with all your hosts listed
[18:41:35] <mboman> I have an issue when I try to download a file from GridFS using python: The way I do it stores the whole lot in memory before writing to file and I need a more memory-efficient way to do it..
[18:43:59] <mboman> They way i do it now: https://github.com/mboman/vxcage-jobs/blob/master/utils.py : the get_file() function...
[19:24:55] <ShortWave> MongoError: exception: geoNear command failed: { ns: "drupal.offers", errmsg: "exception: assertion src/mongo/db/../bson/bsonobjiterator.h:79", code: 0, ok: 0.0 }
[19:32:05] <in_deep_thought> should starting mongodb be different on arch linux? I try mongod and it gives me ERROR: dbpath (/data/db/) does not exist. but then according to the arch linux wiki I try sudo systemctl start mongodb and it works
[19:32:11] <asturel> if i have a document like { something:{something2="x"}} how can i find it by something2 value?
[19:33:06] <cheeser> in_deep_thought: probably a permissions issue on that directory when start it as non-root/mongodb user
[19:33:34] <in_deep_thought> cheeser, so can I just use sudo systemctl start mongodb all the time without issue?
[19:34:13] <cheeser> most services tend to run as either root (ugh) or specific user accounts to isolate permissions
[19:34:43] <in_deep_thought> I read everywhere that you start it with mongod but I guess that doesn't guarantee I have the right permissions user info
[19:39:59] <in_deep_thought> I get Mon Apr 7 12:37:48.119 Error: couldn't connect to server 127.0.0.1:27017 at src/mongo/shell/mongo.js:145 when I try and connect to the shell with mongo. The start command ran without error. I can't tell what is going on in this error aas it doesn't give me much info
[20:03:13] <in_deep_thought> I would ask this on the mongoose irc but there doesn't seem to be anyone there: when I use the mongoose.connect() command in node does that switch the current mongo database to the one I specify, or do I need to manually switch it first with the mongo shell?
[20:18:21] <Joeskyyy> Depending on what you're storing you may want to try a different datastore, like Redis or something
[20:19:00] <Joeskyyy> Redis could be a problem too though, since that's all in mem, just depends on what you're doing again
[20:22:30] <BlakeRG> Hello all, i have a collection of documents with a 'birthday' field in each, whats the best way to pull out a list of age ranges for those documents if the date format is something like '01/10/1979'
[20:24:56] <BlakeRG> Joeskyyy: yeah, i realize that now.. but what kind of query would i want to run against the collection to pull the data i need out at this point
[20:25:29] <Joeskyyy> You'd need to write a script to use "/" as the separator, then sort on each block, would be how I would approach it
[20:26:54] <Joeskyyy> I'd imagine there's some JS way of doing it, but I'm no mighty JS programmer. I'd prolly look to awk/python to help me with that.
[20:31:34] <BlakeRG> thanks Joeskyyy will go that route, not too hard but i thought it might be easier
[20:32:49] <pasichnyk> hey, i'm having some issues detecting if an Update() with $addToSet actually inserted a record or if the record already existed. I'm using WriteConcern "Acknowledged" and after it finished checking the DocumentsAffected and "UpdatedExisting" properties (c# driver). However if my processing errors, and restarts, i'm getting duplicate downstream processing performed that should only happen if
[20:52:49] <slikts> am I right in thinking that the _id field will be indexed even if I do something like _id: String?
[21:31:07] <TylerE> If I recreate a collection using $out in an aggregate pipeline, which recreates the colleciton atomically, do any indexes stick around?
[21:43:48] <TylerE> I tried using concat and the date operators ($year etc) but $concat refuses anythign that isn't a string
[21:44:56] <hydrajump> hi is it possible to access the material /videos from this course https://university.mongodb.com/courses/10gen/M101JS/2014_March/about all in one go, e.g. before the course is complete?
[22:17:04] <asturel> i keep getting this: "lastHeartbeatMessage" : "rollback 2 error RS101 reached beginning of local oplog [2]" any idea?
[23:12:56] <andrewfree> Are there plans to add back the functionality of the identity_map for mongoid in rails 4