Log file Viewer
#mongodb logs for Saturday the 25th of July, 2015
[01:01:46] <Doyle> Boomtime, To do an upgrade you just install 3.x.x an join to an RS with existing 2.6 hosts, then stepdown the old ones when the initsync is done?
[01:02:53] <Boomtime> please follow the tutorial, depending on what you are running there may be extra things to check
[01:42:32] <Doyle> Boomtime, smallfiles, what if you have a lot of DBs that hold a lot fo small files?
[01:46:56] <Boomtime> do you expect a problem with that? there is nothing wrong with running smallfiles, you just get lots of files which burns through file handles faster too
[01:59:07] <cheeser> i'm not 100% but i'm wondering about moving bits between two files rather than within just the one.
[02:00:03] <cheeser> hrm. maybe it's ok: http://docs.mongodb.org/manual/reference/configuration-options/#storage.mmapv1.smallFiles
[02:00:07] <Boomtime> no, it isn't, but i thought in large part that's because it eats into ulimits faster, which on it's own also has some heap overhead
[02:24:31] <Doyle> On centos, would you want to set fork: true? Why wouldn't you want daemon behavior by default?
[03:31:40] <jigax> hi, i was wondering if someone could give me a hand with the following question? http://stackoverflow.com/questions/31622597/please-help-get-grand-total-of-count-field
[03:32:06] <Boomtime> Doyle: i imagine fork is always required unless you want to run the server as a test in screen session or something
[03:33:52] <Boomtime> the aggregation pipeline provides a single stream of output documents that have the same form - your 'example' clearly shows two different formats
[03:36:27] <Boomtime> ok.. you can do it but i don't have the time to look more closely for your case
[03:38:19] <Boomtime> the $project can't be last, whatever it is doing, the $group that creates the values array and the total count must be last
[03:39:13] <Doyle> Thanks again Boomtime. Just wrapping up the limits and logrotate and I'll be done. Do you know offhand if centos has a virtual memory limit, or memory size that can be set? I'm not seeing it in limits.conf
[04:08:58] <jigax> is there a better way than having a bunch of nested if statements handling if the user selected "All" or a specific user from a dropdown in the front end?
[06:00:57] <Doyle> What disk setup has the best performacne in AWS? LVM 4 stripes on 4 disks x2 segments? (Raid 11?)
[10:08:01] <Doyle> how does mongoperf know which drive to test ? I'm running it, but it's only hitting one device, xvda
[10:51:24] <vramana> Hi, This is my first time using mongodb. I am working on this project https://github.com/yasp-dota/yasp. I have downloaded a mongodb dump. I want to serve it locally. I unerstood that I need to use mongorestore. I don't know what to do after that.
[10:52:04] <vramana> I have to multiples .bson files all of which have to served under a database named dota.
[10:54:06] <vramana> but I dont understand what should be the host and port in my case, how should I start mongod service Everytime I do service mongod start I get mongod: unrecognized service
[11:04:43] <Doyle> vramana, you root or using sudo? if so and you still get that message, it's time to revisit the install instructions. Something was missed.
[11:05:57] <vramana> Doyle: No. I didn't use sudo. I found some stack exchange answer specifying sudo. It worked out.
[11:11:24] <vramana> Doyle: I am using mongorestore --db dota . now. But I am getting root directory must be a dump of a single database when specifying a db name with --db what does this mean?
[11:46:57] <aps> A secondary member of my replica set is in RECOVERING mode. It seems it's unable to catch-up. What do I do now?
[16:09:24] <hypermist> Unable to connect to database: mongodb://coincoin:passwashere@localhost:27017/explorerdb
[16:13:50] <flok420> halp? do you mean http://childrensnational.org/departments/hyperactivity-attention-and-learning-problems-clinic-halp ?
[19:41:25] <blizzow> I have three shards and each member of the replica set has been allocated 48GB RAM. Each mongod instance has ~40GB of indexes. Top reports that only 7GB of RAM are being used. and 6GB of data in cache. free and htop report similar stats.
[19:41:59] <blizzow> mongotstat however reports mapped = ~600GB vsize=~1200GB and non-mapped=~600GB.
[19:42:57] <blizzow> This is using mongostat --discover -u myuser -p mypassword --all --authenticationDatabase admin
[19:44:12] <blizzow> I read up a little and the touch command is supposed to force a lot of data into memory. Is there a different/better way to get mongo to eat more of the RAM that's been allocated to it?
[19:57:23] <blizzow> Also, why the disparity between what the OS reports and what mongostat reports?