pmxbot IRC Log Viewer

[00:24:27] <ckoning> Is it possible to retrieve the replica set name from the command line?

[00:29:11] <cirwin> ckoning: mongo --quiet --eval 'print(rs.status().set)'

[00:29:42] <ckoning> cirwin: thx

[02:34:23] <mnaser> I am looking at the profiler on a mongo database, some inserts are taking up to 50s

[02:35:47] <mnaser> Any idea where I can look for that type of issue/

[02:39:34] <cirwin> mnaser: start with mongostat and iostat

[02:39:38] <cirwin> chances are it's not cpu time

[02:39:50] <mnaser> cirwin: it looks like it's the write lock (i think)

[02:39:58] <cirwin> that'll be the problem then

[02:40:05] <cirwin> mongostat will tell you that

[02:40:11] <mnaser> i have MMS setup

[02:40:19] <mnaser> i'll have a look at mongostat

[02:44:08] <mnaser> cirwin: indeed

[02:44:09] <mnaser> *0 3 *0 *0 0 11|0 1 18.1g 37.8g 9.41g 0 ceilometer:129.9% 0 0|0 2|0 1k 14k 136 02:41:46

[02:44:35] <mnaser> so what can i do regarding this..? ceilometer is an openstack system

[02:45:13] <cirwin> you can either write less, write faster, or shard your db

[02:45:19] <cirwin> the latter is the "best" idea

[02:46:15] <mnaser> cirwin: does sharding the db happen on the app layer or the db layer? sorry i dont do a lot of mongo really

[02:46:31] <mnaser> i'm just using mongo as it's the recommended backend for this app

[02:46:43] <cirwin> you don't have to make any code changes to your app

[02:46:50] <cirwin> you just point it to mongos

[02:46:54] <cirwin> and mongos deals with the sharding

[02:47:00] <cirwin> and has the same api as mongod

[02:47:05] <mnaser> i see

[02:47:13] <mnaser> this is really really stupid but

[02:47:17] <cirwin> how much data are you trying to write?

[02:47:18] <mnaser> can i run multiple mongos on the same instance?

[02:47:29] <cirwin> yes

[02:47:44] <mnaser> i dont know.. the app does the writes, it writes statistics from time to time.. i ran the profiler.. not that much queries

[02:47:47] <cirwin> but you'll probably run into issues with disk contention

[02:48:00] <mnaser> this is in array of ssds

[02:48:04] <cirwin> what's your disk utilisation?

[02:48:08] <cirwin> heh

[02:48:13] <cirwin> that should be fine

[02:48:16] <mnaser> iostat is almost 0

[02:48:24] <mnaser> *iowait

[02:48:38] <mnaser> i think the problem is that it's "big" writes

[02:48:40] <cirwin> yeah, mongo does pretty well on lots of terrible disks

[02:48:48] <cirwin> (instead of one amazing one)

[02:48:53] <cirwin> how big?

[02:49:11] <mnaser> avg obj size in mms is 2.7KB

[02:49:26] <mnaser> 4.94M objects though

[02:50:46] <mnaser> but i am seeing a lot of updates

[02:51:07] <cirwin> updates can cause problems because resized documents may have to move

[02:51:17] <cirwin> using power of 2 sizing helps with that

[02:51:23] <cirwin> if you're not using 2.6 you'll have to set that up

[02:51:32] <mnaser> i am using 2.6

[02:52:10] <cirwin> should be that by default

[02:52:32] <mnaser> cirwin: http://pastie.org/private/ykvfrkcgf6oogkrvp88xw

[04:18:39] <lethjakman> what's the difference between save and insert?

[04:19:31] <lethjakman> ohhh

[04:19:35] <lethjakman> insert or update

[06:38:51] <styles> #hadoop

[12:31:39] <kkuno> hello

[12:32:07] <kkuno> I'm trying to remove an array element (an object) if a property of that object equal something

[12:32:27] <kkuno> i.e. I have tags: [{name: 'hello'},{name: 'world'}]

[12:32:39] <kkuno> and I want to remove the object which have name = 'hello'

[13:21:52] <obiwahn> mh

[13:22:32] <obiwahn> there is some mongodb homework about profiling and i simply fail to get to the data:(

[13:22:41] <obiwahn> there is this json: https://university.mongodb.com/static/10gen_2014_M101P_June/handouts/sysprofile.acfbb9617420.json

[13:23:08] <obiwahn> i load the data with the given command: mongoimport -d m101 -c profile < sysprofile.json

[13:23:33] <obiwahn> but

[13:23:35] <obiwahn> switched to db school2

[13:23:37] <obiwahn> > db.system.profile.findOne()

[13:23:39] <obiwahn> null

[13:35:52] <obiwahn> ok is should have not just copied and pasted and trusted the text

[17:50:27] <brammator> How do I work with stored js in pymongo? Specifically, I try to insert document like that: db.pages.insert({_id:getNextSequence("pageid"), name:"new document"})

[17:50:37] <brammator> work fine from mongo shell

[17:51:22] <brammator> preferably withoou using eval()

[22:07:14] <mango_> hi, I'm using CentOS, I'm just determining where the best place to store data files are?

[22:07:20] <mango_> or is rather.

[22:07:44] <mango_> by default where are programs placed in CentOS or linux in general.

[22:08:07] <mango_> I'm referring to mongo data files here.

[23:46:46] <cathode> hi there. i've been doing research on various nosql databases and it looks like mongodb may be suitable for my use case. i'm looking to store 3d volumetric data for an open-world game (think similar to Minecraft)

[23:47:44] <cathode> so, lots and lots of transactions per second, small bits of data where the schema isn't fixed (e.g. a specific type of block that has a custom piece of data attached to it)

[23:53:05] <cathode> if anyone had some feedback on this i would appreciate it :)

Log file Viewer

Help | Karma | Search:

#mongodb logs for Saturday the 19th of July, 2014