Log file Viewer
#mongodb logs for Monday the 27th of June, 2016
[08:03:34] <xarkes> Hello, I am having an issue with mongodb. I am trying to insert a document which contains many documents, here is a sample: http://pastebin.com/HdmqX6Sj
[08:03:58] <xarkes> And then I get a 'key too large to index, failing' which sounds false to me because no 'key' element seems to big
[08:04:20] <xarkes> I would have said perhaps the value field contains too many documents but ... not the key, anyway how am I supposed to do ?
[08:16:26] <xarkes> Okay according to this: http://www.blogjava.net/czihong/articles/370841.html "Documents which fields have values (key size in index terminology) greater than this size can not be indexed."
[09:06:14] <KekSi> i'm having trouble bootstrapping a replicaset in docker containers: i wrote a script that is executed when the (initial primary) container is launched
[09:07:14] <KekSi> it fires up the instance with replSet parameters and --fork, then runs the bootstrap (which calls rs.initiate setting up the other containers)
[09:08:36] <KekSi> i then want to restart it without --fork but (since i specified the only other instance as priority 0 (being just a read/backup replica that can never become primary))
[09:09:39] <KekSi> so "regular" db.adminCommand({shutdown:1,timeoutSecs:30}) yields me that there are no electable secondaries caught up and using force:true fails: DBClientCursor::init call() failed; Error: error doing query: failed
[09:10:48] <KekSi> does anyone have an idea of doing this another way? (maybe without starting with --fork and restarting but instead maybe doing the bootstrap/init for the rs in a different automated way?)
[09:19:08] <KekSi> i'm especially curious why forceful shutdown doesn't work in a scripted way but does when i do it manually
[09:40:07] <Keplair> I some Troubles with Mongodb, Currently My cluster is ok, it works fine, the Mongos Service for the Routers is pretty crappy, the script don't really work for me. I tried to copy the same procedure like the Mongod service but i don't know why there's some shits. I'm on Centos7.2 with MongoDB 3.2
[12:09:56] <jayjo> From a conceptual perspective, which is better for uploading data into mongodb... a very huge single file of multiple json documents, or many files of a small json document?
[12:18:47] <jayjo> Sorry I wasn't clear. The JSON document is about 150 bytes on average, and doesn't deviate much from this. I have files that are json documents separated by newlines. Should I concat them to one giant file to use mongoimport or make many different files with single jsons?
[12:23:55] <kurushiyama> jayjo If you need to access one, you only have to do a match and return it.
[13:49:10] <StephenLynx> storing json as a file makes sense if you have to serve the json file directly.
[13:51:51] <StephenLynx> not to mention when the data you are serving is not EXACTLY the data you are storing.
[13:54:39] <kurushiyama> StephenLynx Well, calculating such data is easy enough, imho and actually fast enough during runtime. I am not sure what you mean by exactly the same. JSON neither guarantees order of top level fields nor should one rely on it. But maybe I am just not getting the problem.
[13:55:17] <kurushiyama> StephenLynx Storing them in GridFS would be the option, then. And not as a standard doc, I guess.
[13:57:47] <kurushiyama> StephenLynx Which comes with other intricacies. Queryable as per JSON fields? meh. I'd rather store the metadata on insert or update (findAndModify comes to my mind) than have data which is totally queryable opaqued as an encoded []byte.
[13:59:49] <StephenLynx> some even let you handle them exactly the same way you handle actual files.
[14:00:43] <cheeser> all drivers should have that functionality now. https://github.com/mongodb/specifications/blob/master/source/gridfs/gridfs-spec.rst
[14:26:49] <Industrial> How do I query all documents that have the same 'deviceID', where there are two or more documents?
[14:52:41] <aeonFlux> I have a large json file (1m+ entries) that i'm parsing (checking if an ID exists in mongo, if not, insert the data). when i first start the script it parses 20-50 per second. after an hour that drops to 5-10 and after another hour it's barely 1 per second. looks like it starts to slow once mongo hits the full 16gb of ram in the server. thoughts on
[15:02:33] <direwolf> hey , i have a field which array of json i want to query over it how can i do it ..?
[15:06:37] <direwolf> db.playlists.find({"_id":ObjectId("571657f80d261af10a1a274e"),users:{"user_name":"pankaj","rating":"2" } }).pretty()
[15:13:10] <cheeser> direwolf: https://docs.mongodb.com/manual/tutorial/query-documents/#query-on-arrays
[15:13:21] <Ben_1> but it fails with the following message mongod.service: Control process exited, code=exited status=100
[15:14:15] <Ben_1> Insufficient free space for journal files, Please make at least 3379MB available in /var/lib/mongodb/journal or use --smallfiles
[17:05:28] <Ben_1> I don't know what's wrong and in mongod.log there is no log message for this exception
[17:09:18] <cheeser> if there's nothing in the logs, you might try starting mongod manually and watch the shell window where you start it.
[17:16:26] <Ben_1> cheeser: does that mean starting mongod manually won't use the default configuration?
[17:19:55] <cheeser> well, whatever that script says when running mongod, c-n-p that line to your shell
[17:22:45] <Ben_1> cheeser: now I can not start the daemon anymore Failed to unlink socket file /tmp/mongodb-27017.sock errno:1 Operation not permitted
[19:47:17] <jayjo> It seems like a nice tool - I made a few test connections and can't delete them though. Probably stored in a config somewhere?
[19:55:32] <troydm> is there a way to insert document if it doesn't exists or update it if it's there just based on one function
[20:54:47] <jayjo> sorry talking about Compass still... created test connections but can't remove them
[20:57:21] <cheeser> if you run mongostat are you seeing runaway connections? or are you just seeing them listed in a "recent connections" menu?
[20:58:28] <jayjo> Yes, the recent connections. I was just trying to get organized - I have a lot of these past connections
[21:03:30] <jayjo> Compass is pretty good to explore data. definitely depends on the use. And I do think this is the best I've found and its by mongodb, inc
[21:03:34] <cheeser> 1. it's not unnecessary. 2. you've never used it so how can you judge it? it's ok to not have, or at least not voice, an opinion on everything you come across