Log file Viewer
#mongodb logs for Monday the 18th of May, 2015
[04:08:47] <markholmes> hey all, i'm having trouble figuring out where homebrew installed mongo, and how i can start it
[07:33:46] <windows7_> i'm getting a error saying: couldn't add user: "email" is not a valid argument to createUser at src/mongo/shell/db.js:1004
[07:36:27] <joannac> windows7_: if you want to add random other fields, use the customData field http://docs.mongodb.org/manual/reference/method/db.createUser/
[07:42:21] <windows7_> can you shed some light on this statement joannac db.user.update({'email': 'guoqiao@gmail.com'}, {'$set': {'date_slug': '131126-4' }})
[08:18:07] <luc4> Hello! I'm using mongo on a web service running on openshift. It seems that after a few days of uptime I get something like this: http://pastebin.com/gzhDQjqH. Any idea why this might happen? It seems it fails to connect but... why?
[08:41:08] <sabrehagen> would really appreciate it if somebody could provide some input on my question here: http://stackoverflow.com/questions/30229872/node-js-mongodb-connection-in-master-or-forked-thread
[11:45:39] <MachineMan> I’ve got some problems with mongoexport. Trying to export as .csv but my non-_id columns are empty in the .csv. I’m not getting any errors from the commandline though
[11:45:44] <MachineMan> mongoexport --type=csv -d thesis -c ratings -f "_id, uid, action, date" -o usersexport.csv
[13:21:16] <cheeser> you can reorder with the aggregation framework. but typically key order doesn't really matter.
[13:36:19] <deathanchor> the cursor handler should be doing the work, be it a driver, or a javascript function you use in the shell
[15:09:12] <aendrew> Howdy folks! So, I have an array of objects in my document, “document.officers”, each child of which has an element, “name”. I’m wanting to find all documents that match /HERPA.*?DERPA/i, so my query’s like db.documents.find({‘officers.name’: /HERPA.*?DERPA/i}) — but this returns documents where either “HERPA” or “DERPA” is in one of the names, i.e, “officer.name: ‘HERPA LLAMA’” and “officer.name: ‘DUCK DERPA’
[15:09:13] <aendrew> returns a match, where I only want matches like “HERPA SOMETHING IN THE MIDDLE HERE DERPA”. Any thoughts?
[15:17:36] <aendrew> Lulz, noted. Any suggestions how I should approach the problem of people with indeterminate middle names?
[15:32:10] <StephenLynx> heh, the more you know. If you don't close a mongo connection in the node.js driver, the application won't close.
[15:32:28] <StephenLynx> and if you close the connection while a gridstore file is open, it will throw an error.
[15:33:40] <windows7_> db.user.update({'email': 'bucyrustech@hotmail.com'}, {'$set': {'date_slug': '131126-4' }})
[15:40:41] <saml> db.user.update({'email': 'bucyrustech@hotmail.com'}, {'$set': {'date_slug': '131126-4' }}, {upsert:true})
[15:53:33] <StephenLynx> 308 Helen Ave, Crestline, OH 44827" lol why do people put their home addresses on the internet like this?
[16:02:23] <StephenLynx> if you are that careless around the internet to the point people can find out your home address.
[17:54:23] <shlant> hi all. I am trying to set up a replica set with MMS. It shows on the Processes tab, but has no state. I also deployed users but when I actually check on the host, they don't exist…
[18:29:34] <bfrizzle> hi all, when mongod is configured to run in SSL mode, is it /fixed/ into mutual verification mode? or is there a way to configure mongod to require ssl connections without also requiring client certificates?
[18:30:14] <proteneer> there may be clock skew etc between the server that my DB is hosted on and the actual webserver itself
[18:37:20] <svm_invictvs> If I am writing to a collection, how do I know when that is written and replicated?
[18:44:06] <StephenLynx> so if I set batchSize of X, I will get X documents in toArray, then if I call to toArray I get the next X documents?
[18:45:57] <cheeser> probably depends on the driver but i think toArray() will pull in the all the results in to the array not just the current batch
[18:47:34] <cheeser> if you know/set the batch size and you incrementally pull off docs and stop before triggering the getMore
[18:50:00] <cheeser> i don't think you can skip once you start iterating the cursor short of just getting and discarding /n/ docs
[18:50:50] <StephenLynx> so you don't know any means to get a number of documents at once while not pulling them all from the cursor?
[18:51:41] <cheeser> at least, neither the C# not the java driver expose a method like that. i don't know that any of the others do either.
[19:15:18] <flicknick> we had an unclean shutdown and is trying a repair, but it segfaults straight away. I've searched through the docs, but am unable to figure out how to deal with this. Ideas appreciated.
[19:57:03] <dfosl> After every insert, update, or delete operation, MongoDB must update every index associated with the collection in addition to the data itself.
[20:00:00] <dfosl> you wrote when it needs to be moved on disk, if i update integer field from 10 to 2000 then it will not make whole structure larger which means no moving ?
[20:00:58] <dfosl> because i am building write heavy application, but it's write heavy only on un-indexed fields, but i have milions of elements inside
[20:01:22] <dfosl> and docs are saying that even if i update only unindexed fields, db will rebuild index
[20:02:16] <dfosl> "MongoDB must update every index associated with the collection " at the end there is "collection" not "field" :P
[20:06:02] <dfosl> cheeser: sorry but if you would write that only for insert operations that would be true, but in documentation it's true also for update operations
[20:07:17] <deathanchor> dfosl: perhaps if the document had to be moved because it ran out of padding from the update?
[20:07:22] <cheeser> i've already told you everything i know about it. if you find it confusing, file a jira against it.
[20:08:21] <dfosl> then it's specific case, if most of fields are integers or small strings then this not apply in most cases
[20:18:49] <dfosl> last question, without using indexes, on write heavy collection, best way to retrieve list of logs for example for specific user is just to make document for every user and just add logs to field which is array?
[20:19:27] <dfosl> then just retrieve whole array for specific user, which makes less traversal instead of just adding collection with all the logs
[20:34:34] <cheeser> you'll almost certainly need to move that document many times even with powerOf2 sizing
[21:55:45] <SpeakerToMeat> Sorry to ask this here, but I really trust the architectural view of mongoers.... Would you say a database of movies (cinema), would be more fit to a graph database than a document one, to conserve deduplication/normalization of linked items (like director, to movie(s)) rather than trying to implement rdbms like relationships between documents?
[22:01:11] <blizzow> Does anyone here know if the reactive mongo driver tries to access the arbiters of a replica set?
[22:17:27] <cheeser> blizzow: why would it? arbiters are non-data bearing members. there's nothing of interest there to an application
[22:20:26] <blizzow> cheeser: I'm not sure, but my application is trying to bang on my arbiter server. But the arbiter is not in any of my application config Uri.
[22:21:01] <joannac> blizzow: that sounds like a bug. one or 2 connections to keep an eye on stuff, sure. lots of connections - bad
[22:22:39] <cheeser> i guess if it's part of the seed list, it'll connect to discover the cluster topology
[22:26:26] <SpeakerToMeat> cheeser: I tought Java guys only believed in rdbms..... and enterprise beans
[23:00:37] <dreamdust> If I see huge spikes in query time for finds during periods of very low throughput… like randomly a find will take 14 secs for a collection with less than 500k docs. Indexes are in memory as well.
[23:00:55] <dreamdust> What would be a good place to start debugging? Does this indicate some kind of collection lock
[23:08:48] <itisit> how to clean all data of mongodb (db, user, replicaset, sharding, etc)? can I just remove all data from dbpath rather than run commands to drop database etc?
[23:23:17] <joannac> itisit: sure, shut down any mongod / mongos processes,a nd remove all the files