Log file Viewer
#mongodb logs for Monday the 18th of March, 2013
[01:45:34] <oinkon> how did one do simple summing operations before the aggregation framework? Map reduce?
[03:35:33] <michaeltwofish> I could have sworn that they did last time I added users, but I've just added a user that hasn't replicated. I can't find anything in the documentation.
[05:12:26] <michaeltwofish> It helps when you create the user on the right database. Yes, users do replicate.
[08:06:41] <maasan> I have first name, middle name and last name. I want to search for names. I used 'or' operator among each names. But it does not work when user gave both first name and lastname in the string.
[10:29:27] <phretor> Hi. Is there any way to improve the speed of aggregated queries with a $match clause on a grouped field? E.g., db.tags.aggregate([{"$group": {"_id": "$subject", "count": {"$sum": 1}}}, {'$match': {'count': {'$gt': 2}}}])
[10:49:39] <darnte> hi guys, my app have a filter for specific fields in my document ( name, provider, email, company ). What is the best approach, create a compound index for each field combination or use map reduce ?
[10:52:11] <darnte> nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn
[10:53:32] <darnte> Nodex: compound index , the user could search for company name is "foo" and provider startwith "go"
[10:55:05] <darnte> for instance: company is "foo" and provider startwith "go" and company is "whatever"
[10:56:50] <Nodex> but if you must then I would have an "idx" array with name,provider,email,company and regex an $and
[10:56:59] <DinMamma> Hello, im doing a mongorestore into a sharded cluster, unfortunately one shard have most of the shards(the key is a hash). I saw this on a small collection(4gb) were most of the shards ended up on one shard and only when the restore of the collection was finished the chunks distributed evenly.
[10:57:39] <DinMamma> The collection im restoring now is 500gb big and yeah, one shards holds 1600 chunks while the two other shards only holds 900 chunks.
[10:58:06] <DinMamma> Can I safely ctrl-z my mongorestore and wait for the cluster to rebalance itself and the foreground the process again?
[10:58:14] <darnte> Nodex: interesting, create a new field "idx" as array and combine all fields, right ?
[11:01:26] <Nodex> the former will select anything with a AND b in any of the array fields - not really ideal when you need to serach exact field names
[11:02:32] <darnte> Nodex: yeah, as u told, I better create idx as an object to search for field name
[11:04:34] <Nodex> map/reduce cannot run in parralell and will lock your DB when you're doing a query
[11:04:51] <Nodex> so you can only ever really have one client with a garunteed query at any one time
[11:38:04] <master_op> hi, how can i convert bson.objectid.ObjectId to dict after insert with pymongo ? (python)
[11:56:29] <maasan> I have first name, middle name and last name. I want to search for names. I used 'or' operator among each names. But it does not work when user gave both first name and lastname in the string.
[12:48:45] <sivang> My guess is the first is an upstrea config file, so one cannot specify mongodb server config params into it
[12:58:58] <GeertJohan> For development (watching what happens on the server, if documents are stored the way you're expecting them to be)
[13:07:48] <DinMamma> Hiya, I did a mongoimport to a sharded cluster, I had to ctrl-c it since one of the shards had twice as many shards as the other two. The cluster works for inserts and reads but its not redistributing the shards :o
[13:41:17] <Darxval> I am getting a compiler error when trying to use Scons and the C++ Driver (Windows) any ideas?
[13:41:22] <Darxval> cl /Fobuild\mongo\bson\oid.obj /c src\mongo\bson\oid.cpp /TP /nologo /EHsc /O2 /
[13:41:22] <Darxval> D_SCONS /DMONGO_EXPOSE_MACROS /Ibuild /Isrc /Ibuild\mongo /Isrc\mongo /IC:\boost
[14:04:22] <Puma1337> Hi -- I have to apologize in advanced because I know very little about mongo or how it works or what it does, but we are using it and it crashed over the weekend. The log file says: [journal] MapViewOfFileEx for C:/path/to/mongodb/data/db.1 failed with error errno:487 Attemt to access invalid address. (file size is 134217728) in MemoryMappedFile::remapPrivateView. Any help would be much appreciated. Sorry again in adv
[14:14:14] <Darxval> @Puma1337 you may want to check that status of your hdd, as i see this issue has occured before with the hdd being the cause.
[14:20:07] <Darxval> is anyone aware of issues compiling with VS2012? via Scons? I just attempted the git repo instructions for the C++ driver and that failed as well, noticed it was referencing Visual studio 11.0\ for "found Visual Studio"
[14:50:14] <MANCHUCK> in order for mongod to ping replica sets for the heartbeat, do I need to allow ICMP?
[15:09:41] <grahamhar> there is a built in { ping: 1 } to check latency and response times but is nothing to do with ICMP
[15:13:42] <dezel> Anyone have any idea as to why this is returning a result, searching a collection for a users email and an active flag set to true
[16:20:12] <Darxval> @Puma1337 looking at your log that you outputted, does this "C:/path/to/mongodb/data/" directory even exist? or did you rename that path for putting in chat?
[16:50:41] <joshua> Has anyone seen any pre-written scripts for doing stuff like gathering all the indexes for all databases?
[16:52:06] <Darxval> woot fixed my issue, i had the 6.0a windows sdk installed and it kept messing up the c++ driver compilation.
[16:58:22] <joshua> That looks pretty handy. Its fun being the guy managing most of the DB servers but not having direct access to the developer
[17:45:29] <joshua> 10gen has some kind of thing they run on the logs to generate info about things like slow queries and make pretty output. Is that publicly available, does anyone know?
[19:23:48] <wakawaka> Hello. I have a collection that is 500gb free, and my server only have 8gb free. So im thinking of deleteing 100gb worth of old documents.
[19:24:11] <wakawaka> I know that mongo wont actually free up the space on disk for those 100gb that is deleted from the database.
[19:57:30] <Darxval> ugh, has anyone successfully compiled the test sample using VS2012? I built the mongoclient.lib using the git repo and scons --mongoclient. and have boost libs built, but appear to be getting linker errors for boost... even though the project is configured
[21:33:39] <danielbeilinson> Hi! Could answer a little question about dynamic variables in MapReduce?
[22:49:28] <pilgo> Hi all. I have an array with objects and I want to change it to an array of strings. How do I do that?
[22:50:48] <pilgo> So, right now it's like this: {name: "Test 1", users: [{_id: 'blah blah', name: 'username'}]}
[23:26:07] <Bluefoxicy> let's say you have something like a problem solving system, and you list a problem document as something that describes a deviation among other things
[23:27:55] <Bluefoxicy> would it be appropriate to have a document with an embedded document for the deviation { "Name": "Web software corrupts session info", "Description": "The Web application stores incorrect session info, so people can log in and then the second they take action they are treated as not logged in." }
[23:28:18] <Bluefoxicy> and then treat the embedded document essentially as another document, analagous to another table in SQL
[23:28:44] <Bluefoxicy> but it's HIGHLY relevant to re-use the same problem if something similar/identical happens in the future
[23:29:17] <Bluefoxicy> so i.e. I may want to search for something like "corrupt session info" and find deviations that were like that, and say, "Yes, that's what's happening", and apply it to the current problem
[23:29:46] <Bluefoxicy> in an SQL database I figured on having deviations be their own table... in MongoDB I'm thinking instead of referencing, I'd make it an embedded document, which would make more sense i think...