Log file Viewer
#mongodb logs for Tuesday the 10th of June, 2014
[00:57:37] <joannac> default is 27017. you can pick whatever you want, but don't clash it with something else
[03:26:11] <abonilla> does anyone know how to configure mongodb.conf so that it starts as part of a shard or replica?
[03:40:29] <abonilla> joannac: I now see an example file, so replSet = xxx also works in the config file.
[03:45:22] <javeln> i'm using '{_id:{\$oid:"\$uuid"}}' as my query but i keep getting 'invalid ObjectId [$uuid]' as an error
[03:47:04] <javeln> when i did '{_id:{\$oid:"<validoid>"}}' it didn't throw an error, but it then it only update that 1 dock
[08:07:32] <rspijker> so, either make sure dots never appear in field names, or make a little wrapper that replaces dots in field names with some unicode character not likely to show up in your actual content and replace it back on read
[09:00:54] <Nopik_> hi, i'm experiencing some weird situation with mongo, after inserting the record, i do query for it, and it is missing, although if i retry query e.g. second later, it gets returned. it happens both on my devel machine (just single mongod) as in production (where there are 3 servers + sharding). writes are done with { w: 'majority' }, although if I do { w: 3
[09:00:54] <Nopik_> } it works. It is not a replication lag, as I send my queries to master. It is 2.4, though, not 2.6 yet. Have anyone seen something similar?
[09:50:03] <Industrial> I'm not going to bother anymore with people that go MEEH it something I dont like
[09:51:05] <Industrial> dont be a retard over the first thing you see that you can write off as something you dont like
[09:55:10] <Industrial> Hi. I have a Java/Python/Ruby/Perl/Lua/whateveryounameit problem with mongodb
[09:56:06] <Industrial> for clarity and because people are generally unable to read this one language called coffeescript, I've taken the liberty of converting all code I write every day into something you might understand
[10:00:05] <Industrial> (The annoying thing about this all is that probably I'm not going to get any help anymore because of my outburst earlier, but I was the one asking a question, the one getting a "NO I DONT LIKE WHAT YOU LIKE SO UR STUPID" in his face)
[10:00:06] <Nodex> children like you crack me up, you seem to think that help is a god given right and that everyone should bend to you to help.
[10:06:33] <Industrial> well I've got some things to think about, maybe one for you could be to think about wether the best reply to any question asked wether you agree with the philosophical backings of the programming language involved with a question on IRC would be "meet my demands or receive no help" or not.
[10:11:04] <Nodex> LOL this is why you're childish - because you still seem to want to blame someone else because you cannot be bothered to compile coffee script into a language that many more people in this channel understand. I have been in this channel for the better part of 3 years and I have NEVER seen coffee script mentioned once by any of the regulars so I asked you to compile it down to HELP YOU OUT.
[10:11:04] <Nodex> Your chioldish attitude immediately saw it as an attack on your beloved language and you went on a nerd rage. Perhaps if you had thought before you typed you would have an answer by now
[10:15:21] <Industrial> I actually spent time thinking about converting it to js first before I asked. I don't blame anyone for my inability to convert CS to JS. I am blaming you for the type of response you gave me that is by now so typical of JS users that dislike coffeescript. It's not the first nor last time this will happpen but I opted to go for the CS gist
[10:16:06] <Industrial> I would have been fine with receiving no help because maybe you don't have experience with the python API
[10:16:40] <Industrial> the same goes for coffeescript, but the usual response you get from people is 'hey its somewhat similat to what i.. oh no.. coffeescript..."
[10:17:31] <Industrial> It's also not my beloved language, because I do this at work. I prefer LiveScript if you have to know
[10:19:54] <Nodex> I personaly don;t know python but many people in here do. I am not debating languages with you, I have better things to do. You don't seem to want to do what's needed to get help so either wait for someone who knows CS or compile it / write it in a more sparse language that more people know. Else you probably won't get help
[10:27:11] <lxsameer> hey guys, is it possible to search for all Site documents which have an embedded Domain document with value of "example.com"
[11:08:09] <mischat> yay to RC1 ... there is a bug fix in there which is stopping us upgrading to 2.6.0 ... I wonder how many more release candidates there will be
[11:45:17] <Derick> sleepr: with a difference is that not every row (document) needs to have the same fields
[12:02:50] <Gr1> I could see that sometimes, some slave is giving slow queries, when compared to other slaves
[13:26:20] <michaelchum> Hi, I made a new server with mongod --syncdelay 0, I am inserting 30 million documents with pymongo but it going exponentially slow, the insertion pauses (no CPU usage) every 10 minutes before inserting for 30 second, the pauses grow longer and longer
[13:26:45] <michaelchum> My insertion script has been runnin 4 days now, it's just going slower and slower any ideas?
[13:31:00] <michaelchum> rspijker: I was looking into flushing and fsync() but I don't really understand how it works and during the pauses, there is no CPU usage nor Hard disk writing
[13:34:09] <michaelchum> cheeser: You mean using the terminal command mongoimport to import CSV/JSON files? I need to do some computations in Python before insertion
[13:35:26] <michaelchum> rspijker: How do I check the memory footprint? htop tells me 80.2GB of virtual memory and 74.4 MEM% not sure what this means
[13:35:48] <rspijker> well… if your memory footprint approaches RAM, mongo will have to do something
[13:37:14] <cheeser> "Do not set this value on production systems. In almost every situation, you should use the default setting."
[13:38:01] <rspijker> what your mongod is doing is storing as much as possible on RAM. When that’s full it needs to sync to disk and will do so quite aggresively. This is most likely causing the delays
[13:38:16] <rspijker> although it is weird that you;re not seeing any disk activity during that time.. how are you checking that exactly?
[13:40:02] <michaelchum> Thanks guys! I'll try taking out --syncdelay 0 although I had the same issue last week, and read somewhere on StackOverflow that --syncdelay 0 would fix it but it seems to be getting worse -.-'
[13:40:22] <rspijker> there’s really no reason to use syncdelay unless there are some very specific circumstances
[13:41:00] <michaelchum> Because the pauses were about 60seconds at the beginning so it made sense, but after, it's weird
[14:08:02] <rspijker> in linux that’s fairly easy. You can just use directoryperdb and mount different drives to the dirs in question
[14:08:34] <rspijker> if you have only 1 DB, the only option is LVM, but then it’s not really partitions anymore at all
[14:09:02] <Nodex> you can still mount a partition onto a directory INSIDE your /path/to/data directory
[14:11:13] <rspijker> and the layout of the partitions. If they are all on the same drive then it might be a little crappier… But nothing horrible, I’d imagine
[14:19:48] <user55> Working on a tagging system where a tag-item relation is managed with arrays embedded in documents.
[14:19:53] <user55> As an implementation tag may hold a list of items OR an item may hold a list of it's tags.
[14:21:07] <rspijker> it can, as long as the document size (including embeded arrays) is not larger than 16MB
[14:25:39] <ELFrederich> hey guys... curious what kind of API mongodb has underneith. Is it as simple to use as Redis where I can use /usr/bin/nc to communicate with it?
[14:26:39] <JT-EC> Should mongo 2.6.x install headers and lib files? I seem to be missing a huge number of files trying to build it the same way that worked for 2.4.
[14:27:59] <og01_> ELFrederich: perhaps your looking for the wire protocol? http://docs.mongodb.org/meta-driver/latest/legacy/mongodb-wire-protocol/
[14:29:15] <ELFrederich> og01_, Redis has a TCL client...not sure if MongoDB has one, but if it is simple enough to implement it doesn't matter
[14:30:44] <rspijker> mongo objectIDs are 12 bytes. So 1.5 M of them will be more than will fit in a single document
[14:34:13] <rspijker> well… you could do chunking. As in have a tags collections where the actual unique tags live. Then have a tagChunks collection which refer to the tags by _id and each of them can have a part of the itemList. Then you need to ensure in your application layer that each chunk will stay below the 16MB limit
[14:43:42] <rspijker_> but instead of having a single embeded array on the Tag document (item would work exactly the same way), the array is split over several documents each referring to the tag document
[14:47:47] <Darkwater> I want to store quests in a collection; each quest has multiple objectives. Should I just put the objectives in an array in the quest document or is there something better I can do?
[14:55:06] <Nodex> Darkwater : it's hard to know what your usecase is, so people saying it's fine really won't help you
[14:55:51] <rspijker> if you want to use the objectives outside of a quest scenario, it might be usefull to have them stored separately from that quest. Bit realistically, will you ever do that? I think not
[14:56:18] <Nodex> you might not think that rspijker but not everyone thinks the same and everyone has different usecases
[14:57:16] <Nodex> if you think quests will take up more than 16mb then you will have to re-think your strategy
[14:58:01] <rspijker> true Nodex, but I think for 99% of all people this will work. The 1% can come back if they actually run into a problem. I prefer that over having to explain every little detail or ask about every little detail before giving an answer. It’s a choice :)
[14:58:33] <Nodex> I find it better to educate people the reasons then perhaps next time they can answer themselves of help others :)
[15:01:31] <rspijker> certainly something to consider. And I do tend to be a bit more ‘explaining’ on questions that I find interesting (no offense Darkwater, just because I don’t find it interesting doesn’t mean it’s not a valid question)
[15:28:16] <cozby> Hi, I've setup a replica set - works fine, but I would like to make some config changes in mongod.conf and have them take affect
[15:39:56] <inad922> is there a way to make a query in mongodb where I have say 5000 entries and I want 100 entries from 4100 to 4200?
[16:13:32] <shesek> Is there any cost incurred of storing large subdocuments, versus storing them as completely separate documents on another collection?
[16:15:18] <saml> Tue Jun 10 11:17:54 [conn1891318] command articles.$cmd command: { getlasterror: 1, j: true } ntoreturn:1 keyUpdates:0 reslen:101 1206ms
[16:39:48] <shesek> Nodex, if I load the parent document without the subdocument (by specifying a list of fields), will there by any performance penalty from the existence of a large subdocument?
[16:53:20] <shoshy> question , the involves mongodb... i hope someone can help. I'm using the bitnami AMI with mongo on, tried to do "mongo admin -u admin -p bitnami" (which is the default according to http://wiki.bitnami.com/Components/mongoDB#How_to_change_the_MongoDB_root_password.3f) and i get auth faild. I also tried resetting the password by modifing mongodb.conf and restarting the server and following instructions. Still can't login.
[16:53:39] <shoshy> It's a cryout for help basically as this isn't bitnami.. .but maybe someone has been in my spot before.
[17:00:01] <inad922> How can I indicate the total number of entries in datatables if I get pages via ajax calls?
[17:06:50] <Lawhater> Hello, if I have a DocumentField.. but I want multiple DOcuments.. how do I type that in the terminal? http://pastebin.com/dG8rxiL3
[17:07:10] <Lawhater> I have options = db.DocumentField(Option) but there will be multiple options.. :S
[17:19:02] <shesek> When an upsert results in an update to an existing document, is there any way to get the ID of that document?
[18:00:27] <insanidade> quick question: I have a simple flask-mongoengine app with two models (A and B, where A has many B's and B has a reference to A through ReferenceField). I created a sample .js file so that I could insert some testing data for both collections in mongodb. The problem is that it looks like that relationship between both models (collections A and B in mongodb) do not exist. Is that the expected behavior?
[18:26:12] <linojon> hi, using mongo_mapper, if i subsclass a model, is there a way to store the subclass items in separate collection from the parent ones?
[18:31:07] <joshua> Forgot about the logs, but I don't think that would have a lot of impact on query performance.
[18:38:25] <linojon> wrt my question, i tried set_collection_name on the subclass but gets an unexpected error
[18:38:48] <saml> pymongo.MongoClient(uri, w=1, j=True) I have this. and writes ton of data every 10 minutes. during writes, some queries return empty. why?
[18:52:43] <arghav> So, I got this error "Btree::insert: key too large to index" and the entire document disappeared. Is that weird?
[19:00:04] <arghav> But I wasn't expecting mongo would just give up the data just because indexing failed. :(
[19:01:04] <saml> another thing i notice is whenever that script runs (every 10 minutes), so many connections are open from other members of replicaset
[19:09:29] <Darkwater> I want to change a string in an object in an array in a document, what's the best way to do that?
[20:26:07] <Derick> if this is a production environment, you should step the primary down first before you modify the config, but only after you've udpated for all the secondaries first
[20:30:06] <cozby> ugh, the strange thing is whenever I do sudo service mongod restart, it just hangs at stopping mongod:
[20:57:05] <saml> hey, I have {id1: {count: count1}, id2: {count: count2}, ..} how can I update multiple documents (id1, id2, ..) ?
[22:36:30] <ac360> Can anyone recommend the best way to save long text (i.e., product decsriptions) in a mongoDB database? I'm using the MEAN stack with Mongoose, and I would like to preserve line breaks if possible...