Log file Viewer
#mongodb logs for Thursday the 7th of February, 2013
[02:52:43] <taf2> i have a collection with a record that i've been recording a timestamp of events within
[04:30:12] <mrandall> How do i rename a collection and update all of the references to use the correct new name?
[06:06:01] <cinvoke> Hi , im following this tutorial and i can't seem to connect to mongo. i dont think i have mongo installed correctly. whats the easiest way for me to tell? http://docs.mongodb.org/manual/tutorial/write-a-tumblelog-application-with-django-mongodb-engine/
[07:10:25] <HHELLD> Hi, All. Is it possible to disable listening on TCP port in mongod? (leaving only unix socket)
[08:35:14] <HHELLD> … ah, bind_ip can be a string to be a unix socket. I wonder if that was even intended.
[11:08:14] <Nodex> http://www.dailymail.co.uk/sciencetech/article-2274388/MI5-install-black-box-spy-devices-monitor-UK-internet-traffic.html#axzz2KD6UaAd2
[13:48:28] <synchrone_> but with cmd.exe , mongofiles -d db search <path part> gets me messed characters
[14:12:11] <BadCodSmell> Can mongo use an index like this: index = {a:1,b:1,c:1,d:1} db.abcd.find({a:123,c:321}).sort({c:1, d:1})
[14:12:26] <BadCodSmell> I have a case where explaining this shows that it is not using the composit index
[14:28:30] <ExxKA> Hey Champs. I am going over old log files from the channel to find an answer to my question, but I can not seem to. How is replication different from a RAID 10 setup in concept? Aren't they both about data preservation and uptime?
[14:51:23] <BadCodSmell> Can mongo use an index like this: index = {a:1,b:1,c:1,d:1} db.abcd.find({a:123,c:321}).sort({c:1, d:1})
[14:55:45] <sfa> I think you need to create two separte indexes on for{a:1, c:1} and the other for {c:1, d:1}
[14:57:43] <BadCodSmell> Shouldn't it be a nested index? It should work with and portion in the find and the rest in the sort
[15:07:16] <ExxKA> BadCodSmell.. Have you tried doing more than 1? So one for your long find, and one for the sorts?
[15:12:34] <BadCodSmell> adding another though removes that ability to use the composite index at all
[15:15:34] <ExxKA> I asume it is because it chooses the shortest index that fits the criteria, but I am not skilled enough with mongo to answer that question
[15:17:16] <BadCodSmell> so it should just be jump to a.b then foreach(c as contents)foreach(contents as item)send(item)
[15:21:01] <BadCodSmell> Do I have to do something stupid like create a single big text field to merge?
[15:21:29] <kali> {a:1,b:1,c:1,d:1} db.abcd.find({a:123,c:321}).sort({c:1, d:1}) you meant { a:..., b:...} in the selector right ?
[15:45:19] <vlad-paiu> Hello. I seem to be having an issue with the Mongo C driver. I'm setting the MONGO_SLAVE_OK flag when performing reads, but the find queries are still directed to the primary server. Is this behavior not supported yet ?
[15:58:35] <vlad-paiu> anybody with knowledge of the mongo C driver ? I took a peak at the MONGO_SLAVE_OK flag.. and it doesn't seem to be used anywhere in the C driver
[16:23:11] <algernon> vlad-paiu: fwiw, there's an alternative C driver, which does use MONGO_SLAVE_OK.
[16:24:05] <algernon> vlad-paiu: other than that, as far as I understand, you should be able to set the flag yourself when sending commands to mongodb, on a per-command basis. So the driver won't use it much more than defining the constant
[16:29:04] <vlad-paiu> algernon: thanks for the answer. I see the official driver takes the flags passed on a per command basis and just passes it in the mongo cursor.. but I see no sort of routing logic change being done whether the SLAVE_OK flag is set or not.
[16:29:28] <vlad-paiu> algernon: which alternative C driver ? :) I'd be happy to use it, if it supports this functionality
[16:30:39] <bartzy> Lots of inserts (with binary data up to 1-2MB) - and then almost all of it is deleted.
[16:31:00] <algernon> vlad-paiu: routing logic isn't done by the official C driver, iirc. the flag just says that if whatever is on the other end is a slave, it's ok.
[16:32:09] <algernon> vlad-paiu: if you have questions wrt my lib, I'm happy to help (but got to rush home atm)
[16:32:12] <Nodex> if you're deleting a alot and things span these files you'll get large disk seeks for chunks
[16:32:47] <bartzy> Nodex: Not necessarily gridfs. The binary data is up to 2MB ... so why use gridfs ?
[16:38:00] <vlad-paiu> One more small question... if I set the SLAVE_OK flag to a query directed to a mongos.. will the mongos know to direct the query to a slave ?
[16:46:30] <bartzy> The storage space stays the same - but when I then insert the same amount of files again - will mongo allocate more files - or just use the existing ones.
[16:47:43] <Nodex> you should consult the docs for that answer, there is arguments for new allocations and arguments for reclaiming space
[18:40:21] <baegle> Hello, not sure how to craft this query in google and haven't been able to figure it out from the docs. Is there some way, instead of writing the value "6" that I could write an object value AND reference a property in it for the same effect? Example: db.collection.find({x:{frank:6}.fank});
[18:50:32] <baegle> I'm VERY new to MongoDB, so please bear with me. The shell is reporting that find is not defined. Do I have to do db.collectionX.find?
[18:55:56] <salentinux> Hi guys, is there way to get results order by ascending discance when use $near or $within?
[19:08:38] <baegle> kali: I'm confused, I'm not sure your example does what I was trying to do, but I'm having trouble deciphering it
[19:09:24] <baegle> kali: if I do db.coll.find({"x.frank":6}, {"x.$.frank":1}); what does the 1 represent? Where does that come in?
[20:24:51] <Leeol2> How would you query for documents where X: 'X', and if there is more than one match then also check Y: 'Y', and etc. The goal would be to get a single document, by compare as few fields as required.
[21:45:35] <jmpf> kinda confused on config servers - you have to cp the entire dbpath over first before you convert your replica set to sharded replica set? - ours is 300G - that means we have to rsync 300G to 3 servers??
[21:47:09] <jmpf> kali: looking @ http://docs.mongodb.org/manual/tutorial/manage-sharded-cluster-config-server/ it seems pretty clear there but then I read about only metadata needs to be on config servers
[21:48:12] <kali> jmpf: make sure you understand every single word of http://docs.mongodb.org/manual/core/sharded-clusters/ begore going any further
[21:51:17] <jmpf> kali: yeh, I've been reading it but I'm still confused - the dbpath on our replicated mongods doesn't appear to be just metadata - that's what I don't understand? it's not clear whether or not we have to cp that over
[21:53:21] <kali> jmpf: honestly, you're so confused i feel reluctant to give you a yes/no answer. you haven't grasped the basics stuff, you'll shoot yourself in the foot.
[22:16:51] <_sri> what's the easiest way to force a command reply with QueryFailure flag set and $err value? :)
[23:13:55] <bartzy> When inserting a document , will it be in RAM immediately if I "find" it , or only after the first find ?