Log file Viewer
#mongodb logs for Thursday the 7th of July, 2016
[00:33:04] <edrocks_> any using mgo(golang driver) bulk api to do inserts expierence many documents not being inserted?
[06:31:19] <bambanx> guys how i can keep running mongodb, i am using webfaction is shared hosting, i used fork method but am not sure if i am doing well
[06:33:08] <bambanx> that looks like this /home/myuser/web/mongo/mongodb-linux-x86_64-3.2.7/bin/mongod --auth --dbpath /home/myuser/web/mongo/data/ --port 34012 --fork --logpath /home/myuser/logs/user/mongo_db.log
[06:33:53] <bambanx> and for stop the only way i founded is : killall -15 mongodb, kill pid is not working at all
[06:40:22] <Boomtime> check the mongod.log to see if the signal is received - it prints all signals received
[06:50:08] <Boomtime> mongodb stores data exactly as it receives it, there is nothing in the world that can tell you if it's "correct" because only you decide what is correct
[06:50:33] <Boomtime> if you can query your data, pretty much at all, then mongodb is doing it's job just fine
[06:51:46] <Boomtime> @bambanx: that is a logistically impossible request - "check your data" is meaningless - what do you want to check? be specific
[06:52:21] <Boomtime> if you just want to know if stuff is accessible and general stats like size and counts etc, are there then the answer is cloud manager
[06:53:59] <bambanx> if i have an app for store : title and post for example, and for some unknown events , my database is saving data in a wrong way or the relations betwen user and posts are wrong, all that shpuld be tested on the app itself so no related to database?
[06:54:40] <Boomtime> the database doesn't know what your app intends to do with the data - how could it?
[06:55:16] <Boomtime> if you want to check that all the data is readable then you can run validate -> https://docs.mongodb.com/manual/reference/command/validate/#dbcmd.validate
[07:09:59] <Boomtime> there really shouldn't be any reason for the process to exit though - also, you asked about ways to terminate the server, you could also use the shutdownServer command from the mongo shell with appropriate permissions
[08:59:47] <sandstrom> I've got issues with stacked fsynclock making it impossible to unlock using `db.fsyncUnlock()` Has anyone had this issue?
[09:46:48] <krion> hi guys, how can i investigate a chunk migration that result in a "data transfer error" ?
[11:11:54] <Mmike> when using mongo cli - is there a way NOT to display some field from some document? I have a document that has some 'log' field which holds like 10k documents
[11:38:09] <deathanchor> Mmike: db.collection.find({ search : "forthis" }, { "want.this.only" : 1 })
[11:39:49] <Mmike> now just to figure out how to do it when I have this field in array of documents :)
[11:41:23] <deathanchor> suppose you have sub : [ { log : "longthing" }, { log : "otherlong" } ], just do this
[11:58:17] <Bengt__> What kind of websites/ web apps benefits from using noSQL rather then a SQL database?
[12:47:12] <mroman> Can you count in {foo: [8, 7]},{foo: [9,1]},{foo: [8,4} how many times each number occurs without doing an unwind?
[12:48:02] <mroman> Currently I do an unwind {$unwind : foo},{$group : {_id : '$foo', count: {$sum : 1}}}
[13:39:45] <cslcm> hi folks, what's the difference between a multikey index and a compound multikey index?
[13:43:04] <cslcm> If I wanted to search for documents with A = hashed index OR A in multikey index ordered by a date index, is that doable in an efficent way?
[15:09:49] <deathanchor> cslcm: you can pastebin the explain, but the important stuff is the nscanned stuff, also look at the explain() docs online
[15:18:31] <deathanchor> hmm... I haven't done much troubleshooting queries in 3.2 yet. mostly working in 2.x world.
[15:21:00] <deathanchor> cslcm: if you query through mongos and it spans multiple shards, the sort will happen on the mongos side instead of on the mongod server.
[15:21:21] <deathanchor> basically it will query the mongod servers, get results, then merge and sort on the mongos.
[15:22:54] <deathanchor> this becomes super painful when you do stuff like skip() with a sort() on mongos as it will query mongod, sort on mongos, and then mongos skips to wherever
[15:23:49] <deathanchor> I don't know if 3.2 fixes the problem but in 2.x it would have mongos do the sort/skipping even if your query had the shardkey only hit data on one mongod.
[15:25:30] <deathanchor> it might do the sort on mongod server, but then has to merge sorted lists on the mongos
[15:26:55] <spleen> I would like to check if my mongo are sync via optimeDate value. Do you know a good way ? a nagios check to do that ?
[15:27:06] <deathanchor> it's funny, mongos is described as lightweight, but sometimes it does hog up some ram for these cross cluster sort/merges.
[15:31:44] <spleen> deathanchor, arr.. i have to install python on my prod servers.. is there a bash check to do that ?
[15:35:42] <deathanchor> you can install python on your monitor server and have it connect via mongo to your prd servers
[15:37:35] <deathanchor> yeah then configure your check to run the command local on the montior server with the -H -P options to connect to where it should
[15:37:47] <spleen> question : what is the difference between optime & optimeDate via rs.status() command in Mongo ?
[15:41:10] <GothAlice> deathanchor: On your earlier comment about light-weightness; all database services can be reasonably expected to consume all available resources on their respective hosts. They don't play well with others, and this is a good thing, to keep data in RAM cache as much as possible. One of the difficulties with sorts/merges is that they can't easily be streamed, you must gather the results first, and that takes RAM.
[15:48:52] <spleen> deathanchor, on my monitoring server i need wish paquage just for connect to mongo ? mongo-cli something like that ?
[16:02:22] <Mmike> Hi, again! :) How do I copy documents across collections? I tried this, but it won't work: db.collectionB.insert(db.collectionA.find({"_id":"myNeededId"}))
[16:02:31] <Mmike> but I'm getting: can't save a DBQuery object at src/mongo/shell/collection.js:134
[16:05:45] <kurushiyama> Mmike http://stackoverflow.com/questions/27039083/mongodb-move-documents-from-one-collection-to-another-collection/27041518#27041518
[16:11:43] <spleen> The best information to know if mongo rsm are sync or not according to you, is optime -> ts when rs.status() or timeDiff when db.getReplicationInfo() ?
[16:18:20] <spleen> kurushiyama, optime is the same thing of OptimeDate but in timestamp mode (then i exec rs.status())?
[19:13:58] <Mmike> I tried several tools but they all claim that my document is flawed: http://jebo.me/pas/1g
[19:14:30] <Mmike> I created stateServers.json using bsondump agains the .bson file created by mongodump
[19:18:50] <jayjo> I have asked a similar question here before - is there a way to have a user defined function or VIEW in mongodb? My very particular use case right now is to return a result (like views, or clicks) for the last 10 days at any given day. Can I have a stored query with a dynamic aspect somehow, or is it entirely dependent on the api I access the DB with?
[19:28:35] <Doyle> jayjo, no, that doesn't sound like a db function. That sounds like a 3rd party analysis tool
[19:30:26] <cheeser> you could always dump these documents in to a new collection with a TTL index...
[19:32:05] <Mmike> Doyle, that wont't work for some reason - I dod: --eval "printjson(db.mycollection.find())"
[19:36:15] <Doyle> For .find() operations, you need to call an operation on the result object to print the documents, such as toArray() or shellPrint(). e.g., mongo userdb --eval "printjson(db.users.find().toArray())" - http://stackoverflow.com/questions/4837673/how-to-execute-mongo-commands-through-shell-scripts
[20:54:38] <cschneid_> I'm looking at the drivers in ruby - is moped deprecated / for older versions - and mongo-ruby-driver newer? How are they related/different?