This reverts commit 243ed53.
…tmpdirs to be fully cleaned up on worker machines.
…d 'server to node ' Server to node communication is supposed to be fast, always. But node/worker to server can be slow, very slow. So, if you have this issue with timeouts, you can configure how you want this.
…break loop and add filter conditions.
… so it can handle another action.
…eturn as soon as possible to client. By this way, when distribution fail or take too long, client request won't get affected.
…r and node has failed.
…wnload. Slower, but more portable.