destroy about node-pool HOT 7 CLOSED

Hi there,

Keeping track of objects in the pool seems reasonable -- I'm not sure there's a huge downside. Can you think of one?

Seems like that would be fairly easy to implement. Do you have a fork of the code? If so, please take a pass at implementing this and issue a pull request to me.

thank you!

-- James

from node-pool.

The only downside is, if the application using the pool doesn't clean up after itselft, the objects in the pool may not get garbage collected, as there is still reference to them. So solving it is not quite trivial, as it can introduce memory leak as well as the pool filling in without "knowing" why.

from node-pool.

We could introduce a max checkout timeout to workaround this. It would work similarly to the idle timeout. if an element in the pool is checked out for more than the checkout timeout, we destroy it and remove it from the pool.

would that work?

from node-pool.

well, this problem is as old as ehmm.. as something really-really old.
don't you remember about
var conn = undefined;

try{
conn = pool.getConnection();
}
finally{
conn.release();
}

this max checkout is hacky solution which can posibly bring more pain then profit when some long running jobs will eventually get dropped because of it.

But maybe will work :)

from node-pool.

As I checked through the Object pool pattern recently as I'm writing a Node.js article, I realized that the destroy shouldn't even be a public property. Only the borrow / returnToPool should be public. Maybe it should even be renamed to those that are in the pattern description. Anyways I've already forked the module, and will commit my changes in the next couple of days.

from node-pool.

cool. looking forward to your commit.

from node-pool.

The destroy is now moved out of public scope.

from node-pool.