Log file Viewer

[00:15:59] <daidoji1> falieson: then denormalize away

[06:25:58] <Fr0g> sharding is the secret sauce

[06:30:13] <svm_invictvs> in Morphia, how do I create a unique compound index?

[06:30:51] <svm_invictvs> I see how to create ac ompound index

[06:30:54] <svm_invictvs> I see how to create a unique index

[06:31:03] <svm_invictvs> but not how to create an index that is both compound and unique

[06:45:21] <aps> Hi recently updated to mongo 3.0.4 and I have started getting this message -

[06:45:21] <aps> WARNING: /sys/kernel/mm/transparent_hugepage/defrag is 'always'.", We suggest setting it to 'never'",

[06:45:21] <aps> I can see in documentation what do I need change, but my question is, why is this showing now when I haven't changed and `transparent_hugepage` settings?

[06:46:09] <preaction> you just said you upgraded mongo

[06:47:38] <aps> Yes, but why is it like, it was okay with Mongo 2.6 but not with 3.0?

[06:47:55] <preaction> because mongo 3 detects it, and 2.6 didn't

[06:48:06] <preaction> it's still fine. read the message

[06:48:16] <preaction> it says "warning", and "suggest"

[06:49:07] <aps> cool, thanks. I have some other DBs still running at 2.6, so I was skeptical if I should make these changes

[08:04:27] <rusty78> anyone still up?

[09:08:04] <aps> Is it possible to set an alert in MMS for DB storage size being greater than a percentage of the respective disk size?

[09:17:18] <jamiel> Hi all, I'm having an issue where my secondary crashed over night and won't restart with the following error:

[09:17:20] <jamiel> [repl writer worker 13] _getOpenFile() invalid file index requested 687846576

[09:17:28] <jamiel> Full dump here: http://pastebin.com/64As3jKN

[09:17:34] <jamiel> Anyone seen this before?

[09:19:37] <jamiel> Seems like a strange index file number, as data files go upto <db>.54 , and largest collection is 28624374 records

[09:19:49] <jamiel> Mongod v. 3.0.4

[09:21:40] <lijeesh> Hello guys, I have a cluster of two shards (two replica sets, rs1 and rs2). I am trying to move a shard to rs1 using sh.moveChunk(). But it gives me following error:

[09:21:47] <lijeesh> "errmsg" : "ReplicaSetMonitor no master found for set: rs2",

[09:21:53] <lijeesh> what could be the problem?

[09:22:45] <lijeesh> I am using 3.0.4 and rs.status() on rs2 does not report any problem

[09:23:37] <lijeesh> correction: I am trying to move a *chunk*

[09:24:58] <jamiel> Have you tested connectivity between the shards?

[09:25:30] <lijeesh> yes, they are on a private network and can communicate without any problem

[09:31:53] <aps> lijeesh: did you check output of rs.status( ) for rs2?

[09:32:00] <aps> oops

[09:32:07] <aps> you did that

[09:33:56] <aps> lijeesh: are you getting something like "balancer move failed" as well?

[09:46:09] <lijeesh> here is the full message

[09:47:36] <lijeesh> "errmsg" : "ReplicaSetMonitor no master found for set: rs2",

[09:47:43] <lijeesh> "errmsg" : "data transfer error",

[09:47:48] <lijeesh> "errmsg" : "move failed"

[09:47:53] <lijeesh> no other error message

[09:50:08] <lijeesh> http://pastebin.com/Dz1DA4pm

[09:50:23] <lijeesh> this is the output of moveChunk()

[09:55:07] <lijeesh> aps: here is rs.status() from rs2 http://pastebin.com/yUGpfJu5

[13:32:09] <johnflux> Can I mark a collection as readonly?

[13:32:14] <johnflux> to prevent accidental changes

[13:32:20] <johnflux> or accidental drop()'s

[13:57:42] <Derick> johnflux: nope, but you can setup authentication to do this - and grant a user only read-only access

[14:00:42] <johnflux> Derick: ah good idea

[14:00:51] <johnflux> Derick: hmm, is that easy, or a lot of steps?

[14:03:45] <Derick> johnflux: a lot of steps

[14:20:08] <m3t4lukas> how do I set the current Screen in ApplicationAdapter?

[14:21:49] <johnflux> Can I do something like db.countries.find().forEach() {

[14:21:59] <johnflux> and then do something like:

[14:22:29] <johnflux> db.data.find({country: "$country"}).aggregate(....)

[14:22:52] <johnflux> basically I want to run a aggregate command on my data for each country

[14:23:00] <johnflux> does that make sense?

[14:28:22] <johnflux> ah, I could assign the cursor to a variable, then iterate over that variable

[14:30:40] <m3t4lukas> ooops, wrong channel :/

[14:32:26] <johnflux> m3t4lukas: i'm an android dev too :)

[14:44:52] <m3t4lukas> cool :) libGDX?

[15:50:30] <Lope> Should I be at all concerned about storing phone numbers (11 digit integers) as integers in javascript and MongoDB? Is there any chance of Javascript or MongoDB flipping a bit due to some kind of floating point thing? I seem to recall reading something once about Javascript not having a clear distinction between integers and floating point numbers?

[15:55:10] <kali> Lope: store them as string

[15:55:37] <kali> Lope: or you'll loose spacing and leading 0

[15:55:45] <kali> Lope: leading +

[15:57:02] <Lope> Thanks, I don't need spacing, and I don't need leading 0 or +.

[15:58:02] <Lope> My data should be a lot smaller when storing the phone numbers as integers rather than strings. As long as JS and MongoDB don't flip any bits due to treating the integer as a floating point number, but I'm not sure if there's any risk of that with an integer?

[15:58:05] <kali> you only have javascript issues if your client application uses javascript. mongodb does not use javascript in the storage engine (BSON has stronger typed than json)

[15:59:26] <kali> i honestly would not use integer, even with a sensible language