Log file Viewer
#mongodb logs for Thursday the 19th of June, 2014
[01:44:11] <raphael> { "v" : 1, "key" : { "updated_at" : 1 }, "ns" : "cloudworkflow_production.temp_tasks", "name" : "updated_at_1", "expire_after_seconds" : 1800 }
[01:44:39] <raphael> and yet I see old documents that haven't been deleted a few hours later (TTL is 1/2 hour(
[01:44:41] <raphael> and yet I see old documents that haven't been deleted a few hours later (TTL is 1/2 hour)
[01:45:29] <raphael> { "_id" : ObjectId("53584b8bedaf89e7c200039d"), "updated_at" : ISODate("2014-04-23T23:23:55.268Z") }
[01:46:04] <raphael> I'm running out of ideas, perused the usual source of information but couldn't find anything that would explain what I'm seeing (or not seeing)
[03:09:48] <umquant_> Any experienced time series mongodb users have some wisdom to share with someone looking to move from mysql to mongo for real time sensor data
[05:56:45] <ns5> Hi, my mongodb can't start, could anyone help? This is the mongod log: http://ur1.ca/hk6tm
[06:20:51] <ranman> also on the partition that the journal is stored on? that looks like it had an error trying to allocate a file :/ which would be an issue around space or file descriptors
[06:26:27] <joannac> my hypothesis is that is you moved all the files for that database into a different folder, mongod would start
[06:43:53] <ns5> joannac: no hidden files. I moved the db files to another folder and mongodb can start
[08:13:48] <arussel> in an aggregate, is there a way to cast from int float to string to be used in a concat ?
[10:01:22] <greybrd> hi guys. the number of open client connections to a mongodb server is always constant ( max used ). it is not actually closing and connection are still active. any inputs to close them after a db operation is done.
[10:06:06] <greybrd> I'm actually using only one MongoClient object as a static variable and I'm sharing it across the threads, against my previous logic, which will instantiate new MongoClient in every thread.
[10:07:05] <greybrd> even the single MongoClient object as a static variable has constant connections. when opened it's not getting closed.
[10:11:43] <greybrd> Nodex: yeah. I came to know about that. is there a way I can avoid this pooling or close the connection in the pool of my choice.
[10:22:10] <greybrd> Qalqi: sorry for my ignorance. can you tell me how or any Internet link which explains this?
[10:26:34] <fontanon> Hi everybody, I've a newbie question with indexes ... If I'm making queries like db.mycol.find({FieldA:'valueA'}).sort({FieldB:-1}) which would be the best index to create? I'm in doubt between creating an index for FieldB or creating a compound index for FieldA/FieldB
[10:32:46] <Nodex> greybrd : I am not sure but it's probably not the best idea to go outside the pooling internals of Mongodb
[10:33:43] <greybrd> ohhh.. I see. so the I shouldn't be bothering about the active open connections?
[12:52:05] <apeeyush> Hi! I have a collection.. I run an aggregation query on it that may take minutes (large data).. Will I be able to insert new documents or perform other queries..?
[12:57:13] <apeeyush> Also, I will be mainly inserting and querying my data.. (No updates at all). Any benchmarks of using MongoDB vs SQL(may be postgres) for indexed or non-indexed data...
[12:57:43] <spacepluk> hi, I'm trying to use mapreduce with mongoose/nodejs. Is it possible to use closures for map/reduce functions?
[13:09:44] <arussel> cast would be nice, cast to string for concat and cast to int as a 'cheap' floor
[14:17:10] <MANCHUCK> why would my primary take up 3 times the amount of disk space as the secondaries ?
[14:31:02] <rspijker> MANCHUCK: the only way to actually reclaim disk space is to do a repairDatabase
[14:43:45] <rspijker> “However, if you trust that there is no corruption and you have enough free space, then repairDatabase is the appropriate and the only way to reclaim disk space.”
[14:44:47] <MANCHUCK_> "If you are trying to repair a replica set member, and you have access to an intact copy of your data (e.g. a recent backup or an intact member of the replica set), you should restore from that intact copy, and not use repairDatabase.
[14:51:47] <kali> it means that repairDatabase can be used to try to salvage whatever can be salvaged if you're that desperate. but it won't break anything if your database is sound and you just want to reclaim space
[16:40:26] <fontanon> Hi everybody, I've a newbie question with indexes ... If I'm making queries like db.mycol.find({FieldA:'valueA'}).sort({FieldB:-1}) which would be the best index to create? I'm in doubt between creating an index for FieldB or creating a compound index for FieldA/FieldB
[16:59:44] <fontanon> Nodex, why did you asked me about the mongodb version? was it a significative enhacement or changes in mongodb 2.6 related with index?
[17:01:25] <kali> fontanon: in 2.4 and before mongodb can only deal with one index when dealing with one query
[17:03:13] <kali> fontanon: the -1 will be marginally different. with this compound index, mongodb will just perform a lookup in the index then scan the index, and it will find the document it needs to send you back in the right order
[17:04:24] <kali> fontanon: 2.6 can handle less ideal cases in a less catastrophic way than 2.4, but if this query is important for you use case, just create the "right" compound index. you'll save a few polar bears
[17:56:53] <Nodex> pre-allocation of a file for the journal so it can be in the FS inodes next to the next one
[18:01:51] <NaN> I just created /var/run/mongodb/mongod.pid with sudo because log tells me wasn't present
[18:03:32] <NaN> how can I be sure it's not running? kill all the mongod and try run the service again
[18:10:30] <kali> but strikes in Toulouse can be strong (in general, not sure about air control specifically)
[18:10:54] <NaN> is mongod.lock the only file I can remove in order to "reset" a clean start for the mongod service, or may I remove another file?
[18:14:49] <kali> NaN: in that case i would start with a chown where these wrongly attributed files are
[20:31:16] <AliRezaTaleghani> but when I cheeck the shared collection! it's self status! it dosen't show archive shard status
[20:31:35] <harttho> Anyone know what would cause a field to become NaN? The field starts at 0 and is only ever incremented with $inc
[20:37:41] <joshua> AliRezaTaleghani: That looks like all your chunks are on one shard. Is the balancer enabled?
[20:39:45] <AliRezaTaleghani> all was perfect unless that the reserved space on SSD disks with primary Shard000 was not released to OS
[20:42:03] <AliRezaTaleghani> we restart the shard and balance! there was some thinks that hint me to config db
[20:45:24] <AliRezaTaleghani> I'm not sure! what happed if I wait for Balance to finish it's process! maybe after an scanning round! anything get done! nicely but I;m not sure :-/
[20:45:29] <joshua> It lists them at the top in shards, so the cluster should be aware of both of them but I think you might have something else going on.
[20:47:25] <joshua> If you want them balanced with even amounts of data you would need a different shart key. But maybe it wasn't designed to split it up for performance. It looks like its putting older data on one shard
[20:50:29] <AliRezaTaleghani> yep via the Tag based sharding I'm just moveing old docs to an slower shard
[20:50:39] <joshua> If I am understanding it correctly, anything older than that date goes on the one with the archive tag, shard0001
[20:51:52] <harttho> Anyone know what would cause a field to become NaN? The field starts at 0 and is only ever incremented with $inc
[20:52:58] <joshua> AliRezaTaleghani: Maybe try doing this with -1 and 1 and see your date ranges: db.bi-reporting.find({},{"dateCreated":-1}).limit(1)
[20:55:04] <joshua> something like this (my syntax might be off) db.bi-reporting.find({"dateCreated" : {$gt: ISODate("2014-06-01T00:00:00Z")} })
[20:56:30] <joshua> I'm heading out, but that might help you figure out what your data looks like which those tags are using to split it.
[21:19:35] <mst1228> hi, i'm using Mongo with Node and have a question about doing queries based on the return of another query
[21:20:08] <mst1228> i definitely understand that there are no joins in Mongo, and if it was up to me, my db would be organized differently
[21:20:29] <mst1228> for various reasons, i have these two seperate collections, both storing user info
[21:20:54] <mst1228> one is 'user' collection, this is the detailed info for a user (name, username, email, etc)
[21:21:27] <mst1228> another collection, 'peoplemeta', has meta data associated to a user. things like tags and previous employers
[21:22:07] <mst1228> I want to get all the documents in the 'people' collection, then add all other those users detail info to the return
[21:22:52] <mst1228> but there is one unique field in 'people' collection that will match a field in the 'user' collection
[21:23:41] <mst1228> sorry for the setup, the question is, do i just have to loop through the first query return and do an individual query for each document in it?
[21:55:38] <tinuviels> Hi there, fast, easy question, I'm Reading this: http://docs.mongodb.org/manual/tutorial/query-documents/
[21:56:23] <harttho> Anyone know what would cause a field to become NaN? The field starts at 0 and is only ever incremented with $inc
[22:00:57] <tinuviels> joannac yes, if exist, or even better would be: if contain at least one element
[22:04:58] <tinuviels> let's fallow this example: 'producer.company': 'ABC123', I want to find all documents, that contain something in "company". All documents will have "producer" but not all will have company.
[22:09:28] <joannac> tinuviels: I gave you 3 examples above. should all of them be returned for your query or not?
[22:10:38] <tinuviels> joannac - got it now, sorry. Ideal would be to count only first one, so when it's not empty
[22:11:43] <tinuviels> however I'm almost 100% that in this DB if document contain "company" then it also contain something inside
[23:01:13] <notaninja> is it possible to change field values of a document reference after it has been populated?