Log file Viewer
#mongodb logs for Friday the 5th of June, 2015
[01:16:23] <Thinh> anyone here have success with mongo-connector and hosted solutions like compose.io?
[03:49:39] <acidjazz> totally strange.. for some reason in my setup all dates that I create in mongodb are 18 hours old
[04:12:26] <n0t1me> i recently tried building the cxx driver but failed due to it requiring the c driver 1.0 although the quickstart guide says it requires 1.2
[04:17:11] <acidjazz> joannac: tahts what i assume.. unforunately this setup i'm using mongolab.com so i cant really tell
[06:33:37] <hypermist> i can't add an account with db.adduser. it says password cannot be blank but its not blank
[07:01:48] <hypermist> http://docs.mongodb.org/master/tutorial/install-mongodb-on-debian/ i followed that tutorial
[07:07:29] <joannac> hypermist: okay, since you don't seem to be getting it: there are 4 steps on that page. I want to see the output of all 4 steps
[07:12:13] <hypermist> Linux hypermistwp2 2.6.32-37-pve #1 SMP Wed Mar 18 08:19:56 CET 2015 i686 GNU/Linux
[07:15:43] <joannac> but if you're using it for something serious, you should jsut get a 64-bit machine or VM
[07:27:25] <joannac> is this your machine? if you delete all that data, is someone else going to be mad?
[07:30:21] <joannac> hypermist: it would be really helpful if you would just pastebin, rather than making me ask every time
[07:37:30] <hypermist> lol it says ** WARNING: You are running in OpenVZ. This is known to be broken!!!
[07:49:33] <joannac> when I type "use mydb" in the shell, I get "switched to db mydb". what do you get?
[07:51:57] <joannac> hypermist: okay, the the terminal that you typed "use mydb" on, what's the full output?
[08:04:06] <joannac> I don't think you actually understand how a database works. so follow the tutorials and it should at least get you up and running
[08:06:56] <Dev0n> hey, anyone know if there is anything that would allow you to use mongodb query on a standalone object that's not currently in mongo?
[08:10:08] <Dev0n> one hacky way to do it is insert it then do find on the id that you get back for that document
[10:11:33] <hypermist> and it says waiting for connectings but when i goto connect my program it goes unable to connect
[11:36:10] <milestone> hi folks, can someone tell me what happened to the find_one method of the ruby mongo driver for 2.6?
[11:36:56] <hypermist> cause if i dont upgrade to 64bit. since its not actually installed as a process or maybe i need to reboot
[12:22:40] <flok420> I have a collection with millions of documents. each has a ts and a value. now I would like to know the maximum and minimum ts (not the value) for all documents where ts older than x. how can i do that quickly? I tried using collection.group(...) but that takes ages
[12:35:08] <flok420> cheeser: would that be something like the following because that still takes minutes on a very small dataset: db.data2.aggregate([ { $match : { 'ts' : { '$lt' : 1433402642 } } }, { $group : { _id : "$ts", max_ts : { $max : "$ts" } } } ])
[12:38:13] <flok420> cheeser: I'm also not sure about that $group : { _id : "$ts", why does it need that _id? I don't care for the_id, I only want to know that max_ts
[12:39:33] <cheeser> you can't telnet to it because you get unknown host. so it sound like you need to fix dns or your connection string.
[12:45:36] <hypermist> it keeps saying authentication failed cheeser but i know the password is correct
[12:46:32] <cheeser> ah, i've never worked with auth really. you might need to specify the authDB but you'll need to check the docs about that one.
[15:39:30] <FanBoy> Has anyone got Edda running in the last year? Tried it in Windows, colleague tried it in MacOS
[15:43:44] <StephenLynx> who in their right mind would take so much time in visual design for a tool like this?
[15:49:34] <FanBoy> OS fanboyism is ironcially lost on me, I honestly don't care what OS you use if it works for you :)
[15:53:14] <FanBoy> as mentioned, it would nice as an educational tool or for demos. Perhaps you've never had to try and sell something to a client, but sadly it is often style over substance
[16:57:12] <ams__> I'm trying to speed up my mongodb upserts. Are upserts likely to be slower than straight replaceOnes? Are there any metrics anywhere I can take a look at?
[17:02:46] <ams__> I've not doing comparisons myself, I'm being lazy and wondering if anyone else has already done them for me
[17:31:13] <greyTEO> GothAlice, I haven't forgotten about releasing the php library for marrow.task. I have been jumping around projects lately
[17:31:53] <GothAlice> Large deployment at work redirected some effort, but it's coming along quite nicely, actually.
[17:32:23] <GothAlice> Most of the work is happening in a fork. There'll be a pull request at the end of that.
[17:34:49] <GothAlice> It's a little http://s.webcore.io/image/310Y3X3b1k1Q at the moment, unfortunately. Fun to debug, though. ^_^
[17:48:31] <StephenLynx> if I am not mistaken, that is when it has to look for something on the disk.
[18:08:12] <GothAlice> ams__, StephenLynx: Correct, a page fault is MongoDB reading data from disk that was loaded before, but had been "paged out" (removed from the in-memory cache).
[18:08:30] <GothAlice> Well, regardless of having been loaded before. Cold caches will generate a large number of page faults initially.
[18:31:05] <brotatochip> cheeser and deathanchor so I've decided to move forward with the 3 DC option with an arbiter and it seems to work well with my forced failover/failback tests. What I'm doing is setting the prod primary priority to 10, DR site to 1, so the failover is warm in the event of DC outage
[18:31:52] <brotatochip> then what I'm doing is going to the DR site (now primary) reconfig its priority to 10 and the prod priority to 1 so that when it comes back up it isn't elected as primary, and this works
[18:32:36] <brotatochip> Then I'll bring the app down in DR, wait for the sync to complete from DR to PROD secondary, then reconfiguring back to original - DR site priority 1 PRD 10, then prod gets elected as primary again
[18:33:18] <brotatochip> oh, right, I could just set DR to 100, and then drop it back to 1 after it completes the sync
[18:34:35] <brotatochip> I don't think they should, because I have one set of app servers in Prod and one set in DR, and they only need to reach the DB that is local to them when that is a primary
[18:35:39] <brotatochip> in the event of a DC outage in prod, I'll either set up hot failover for route 53 to point to DR instead, or manually do it
[18:36:59] <deathanchor> just saying if you want automatic app switching: http://docs.mongodb.org/meta-driver/latest/legacy/connect-driver-to-replica-set/
[18:38:16] <deathanchor> sorry, this is better: http://docs.mongodb.org/manual/reference/connection-string/
[18:38:17] <brotatochip> basically if there's a DC outage, it doesn't matter what the app is pointing to because the app is down in prod, so DR will be spun up to point to the DR mongodb instance, I don't think currently my use case needs automatic app switching
[18:39:06] <brotatochip> thanks for that though deathanchor that will definitely prove useful if and when we need to scale mongodb horizontally
[18:41:00] <deathanchor> so I still need help with profiler.. how do I drop the db.system.profile on a secondary where I ran the profiler?
[18:41:04] <brotatochip> also deathanchor I've decided to implement a hidden slave for backups thanks to your advice
[18:41:28] <brotatochip> don't want to impact prod and I want to be able to take regular backups - both mongodumps and snapshots
[18:43:06] <brotatochip> oh btw does anybody know if there is a way to have secure encrypted communication for mongodb 2.6 without compiling the binary myself?
[18:43:34] <StephenLynx> I remember an issue with SSL in old versions, but it was a bug, if I am not mistaken.
[18:43:44] <brotatochip> SSL requires a custom compile based on what I've read deathanchor and StephenLynx
[18:44:43] <brotatochip> also don't want to run custom compiled binaries in prod as they are such a pita to maintain without something like FPM
[18:45:28] <brotatochip> although I do have a very healthy aversion to not running code that hasn't been out in the wild for at least 6 months to a year
[18:46:44] <brotatochip> the CTO and the director of ops (both of my bosses) share that same aversion, and we don't exactly have a timetable for being able to test our social platform with mongodb 3.0
[18:46:46] <deathanchor> So no answers for how to drop the db.system.profile on a secondary where I ran and stopped the db profiler?
[18:58:07] <brotatochip> and if so, are those credentials sent in plaintext? (this would defeat the purpose for me)
[19:37:10] <crised_> When you use java with mongodb, Do you normally use GSON to convert POJO to JSON?
[19:51:35] <crised_> if mongodb uses bson internally, as a app developer, why would I care about this?
[19:54:24] <Zelest> Derick, i wrote a little thingie to handle translations.. so i basically have an object with the fields en, sv, de, etc for each language.. The page writes the text on english and i wrap the text around this function.. if the text is found in "en", replace it with whatever language the user has picked.. if available..
[19:54:44] <Zelest> Derick, however, i've noticed that texts that contains a lot of spaces isn't found. :o
[20:03:45] <Derick> Zelest: that sounds odd... you sure they're not non breaking spaces or similar types of line endings?
[20:12:17] <crised_> Derick: how do I extract the 4 byte vlue representing the seconds since Unix epoch?
[20:17:12] <diegoaguilar> Hello, I'm trying to use mongoose to perform queries and db logic ... however I got an issue as any model seems to get find, findById, etc methods available ... this is what my code looks like
[20:17:27] <diegoaguilar> Hello, I'm trying to use mongoose to perform queries and db logic ... however I got an issue as any model seems to get find, findById, etc methods available ... this is what my code looks like
[20:26:14] <crised_> Derick: How do I define a coordinate in mongodb? Do I need to define it as a column name?
[20:30:33] <Derick> db.colname.insert( { _id: 'yourid', 'location' : { 'type': 'Point', 'coordinates' : [ long, lat ] } } );
[20:30:52] <Derick> db.colname.insert( { _id: 'yourid', 'location' : { 'type': 'Point', 'coordinates' : [ -0.5, 59.5 ] } } );
[20:31:57] <crised_> Derick: so the coordinate is not either Number or Array, it's a special data type called 'Point'?
[20:34:03] <crised_> Derick: so I don't need to createIndex, I will de using default _id and coordinates
[20:35:37] <crised_> oh I get it, createIndex is done here: db.colname.createIndex( { location: '2dsphere' } );
[20:47:11] <gagalicious> i'm about to convert to cassandra, not trying to start a flame war here... but can anyone advise what's mongo doing better than cassandra? and how does the storage efficiency of mongo compares with tokudb?
[20:48:39] <StephenLynx> absolutely no idea about anything about cassandra and never heard about tokudb
[20:49:39] <StephenLynx> http://www.infoworld.com/article/2848722/nosql/mongodb-cassandra-hbase-three-nosql-databases-to-watch.html
[21:14:30] <diegoaguilar> Can anyone help me with a mongoose issue? https://stackoverflow.com/questions/30675729
[21:56:38] <crised_> When working in Java, How do you validate the structure of the json document inserted?
[22:17:32] <MacWinner> if I have a document like invoice = { notes: 'Test Invoice', products: [ {name: 'product1'}, {name:'product2'}] } is there an atomic way to add another product to the array but only if there is not existing element in the products array with the same name? like if I try adding {name: 'product3'} to products, it should work.. but {name: 'product2'} should not since it's already there.. I want this to be atomic in case 2 different processes
[23:35:55] <MeStesso> hello! do you know if there’s a way to have MongoDB using SSL and requiring authentication only on a certain network interface or only when the request comes from IPs that are not in the LAN?
[23:38:51] <MeStesso> cheeser: i know nginx is a proxy,but i don’t think i can use it to proxy mongodb… what kind of proxy would work then?
[23:40:46] <MeStesso> i can only think of SSH but that would require creating an entire new user with certain privileges i don’t want them to have
[23:41:08] <StephenLynx> but if the peer behind the proxy authenticated, wouldn't the proxy just have to send the server EXACTLY what its receiving from the peer?
[23:41:11] <MeStesso> i can also think of VPN but that requires extra hardware and extra monthly AWS fees. my boss says no
[23:44:30] <MeStesso> ChALkeR: the applications don’t have support for auth in MongoDB and that would require a lot of extra work
[23:46:00] <MeStesso> ChALkeR: look, the developers taking care of mongodb said no auth , they can’t do that
[23:49:26] <ChALkeR> Also, I recommend configuring your firewall to block everything, except for the desired subnetworks. As as _additional_ measure.