Node.js + Cluster + Mongoose

[Jorgen B.]

Hello,

I'm learning Node.js by writing a simple HTTP server application using
Cluster plus MongoDB through Mongoose for sharing state between
different Cluster workers.

Is there a way to have one MongoDB connection per Cluster worker?
Currently I'm setting up a single connection using mongoose.connect()
and then using that connection in my HTTP server callback, which then
gets cluster()ed to multiple workers... but that doesn't seem like a
good solution to me -- a bottleneck if nothing else.

I'm really new to this and I have little idea how things work under
the hood with Node.js (or Cluster and MongoDB, for that matter), so
maybe I'm thinking the wrong way about this. Any suggestions?

Jorgen

[tjholowaychuk]

You already would have one connection per worker.. you can't share the
connection, so you have already accomplished that :p

[Ricardo Tomasi]

If you allow me to hijack the thread: say I'm pop()ing from a
collection in this setup, am I likely to have any racing conditions?

[Ricardo Tomasi]

*race :)

I'm sharing a waiting list between processes and need to make sure an
item is retrieved only once. Might have more to do with how mongo
works, I'm not sure...

[Jorgen B.]

Sorry, uh, I'm not sure if I understood. If I cluster() an HTTP server
that accesses MongoDB through a single Mongoose connection in its
callback, what exactly happens?

[Marcello Bastéa-Forte]

I'm guessing each node worker gets its own mongoose connection.

Marcello

--
You received this message because you are subscribed to the Google Groups "nodejs" group.
To post to this group, send email to nod...@googlegroups.com.
To unsubscribe from this group, send email to nodejs+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/nodejs?hl=en.

[Jonas Huckestein]

Sounds like you're having only one file server.js that looks something
like this:

-- snip --
cluster(http.createServer( ... do stuff ... )).listen();
-- snip --

Cluster will run server.js once for the master and then once for every
worker, which means that every worker has their own database
connection Alternatively (making things a little clearer), you can
pass cluster the path of a module that exports an http server. This
module will only get required in the workers. Check out the API docs
at http://learnboost.github.com/cluster/docs/api.html

[Jorgen B.]

OK, yes, now I understand, thanks a lot.

Jorgen

On Apr 4, 8:47 pm, Jonas Huckestein <jonas.huckest...@gmail.com>
wrote:

> Sounds like you're having only one file server.js that looks something
> like this:
>
> -- snip --
> cluster(http.createServer( ... do stuff ... )).listen();
> -- snip --
>
> Cluster will run server.js once for the master and then once for every
> worker, which means that every worker has their own database
> connection Alternatively (making things a little clearer), you can
> pass cluster the path of a module that exports an http server. This
> module will only get required in the workers. Check out the API docs

> athttp://learnboost.github.com/cluster/docs/api.html