me.destroy() not destroying about node-pool HOT 3 CLOSED

I think this may just be a documentation issue. The intent is that destroy() is an alternative to release(). obj won't be in availableObjects after it is dispensed via acquire, so destroy simply needs to decrement the count and run the destroy() function registered with the pool. For example:

pool.acquire(function(err, client) {
    client.query("select * from foo", [], function() {
        // instead of: pool.release(client);
        // we destroy the client
        pool.destroy(client);
    });
});

Does that make sense? An example use case might be that you want to close the connection to the db after 100 acquires, so you stash a counter on the client obj, incrementing on acquire(). if n < 100 you release(), else you destroy()..

from node-pool.

ok. now I get it, so if say you detect your conn is corrupted somehow, you need to destroy it and not release it. and yes, it makes sense you only "touch" your connections when you're "inside" the acquire callback. The use case I proposed before implies you can touch the conn once it's been released, and , well, you shouldn't worry about conns in the "available" pool, that is supposed to be a task for the pool manager.

anyway,regarding the use case you said, if you set the pool maximun to 100 you shouldn't get more than 100 conns at anytime. should you?

Do you think a README update is worth it?

from node-pool.

Yep, I'll update the jsdoc comment to clarify the intent of the function.

Re: max pool connections, yes that's right. If you set max to 100, the pool will stop creating objects when 100 is reached, and acquire() calls will block.

This raises an issue though -- currently destroy() doesn't call dispense(). I think it should. I'll open a separate issue for that.

thanks

-- James

from node-pool.