Kevin Conaway

@gehel I'm testing locally and I'm seeing the OutputWriters initialized twice. Below is the test config I'm using ``` { "servers": [{ "port": "11061", "host": "localhost", "numQueryThreads": 2, "outputWriters": [{...

> yes, this is most probably wrong, but also most probably of no consequence Are you sure? It will create one pool per outputwriter. If users specify a writer per...

> Actually, you're right, those unneeded pools are not shutdown and not garbage collected. Seems like a cheap fix would be to change `ConfigurationParser#parseServers` to do away with the ServerListBuilder....

@gehel you're right. I created #453 to track this issue separately so as not to derail this one where the the main focus should be stormpot IMO

Any update here? 254 is unusable due to this issue. Can we revert back to using commons-pool (or something else) until stormpot is fixed?

> The problem with this is that there's only one allocation thread, so if you make it sleep then it won't be able to do anything else. Does imply that...

> So blocking it is exactly what we want. Is there a potential to starve out other slots from being allocated though? Lets say there are two socket poolables, one...

Not sure why the AppVeyor build failed. It doesn't look related to my change.

@alistairking did you get around to testing this?

@jliunyu We tested the PR in a fork and it still does not fix the issue. I was able to reproduce the issue with the following test case on my...