Thanks for your contribution. I have two questions:

1. Why don't set timeout 0 to avoid the timout issue?
2. how to test the default check for connection?

1. Why don't set timeout 0 to avoid the timout issue?
   Sorry didn t understand what you mean, here is temporary set, if set to 0 it should be the same as Integer.MAX_VALUE ,should be ok ,just I set a condition here

if (timeout > 0) {
    transport.setTimeout(timeout);
}

2. how to test the default check for connection?

This problem we have already validated in the production environment and will automatically disappears when we catch this excepiton and release the connection

connection is managed by ConnObjectPool,conncetion·s created、borrow、return will checks if is available,

    @Override
    public boolean validateObject(PooledObject<SyncConnection> p) {
        if (p.getObject() == null) {
            return false;
        }
        if (!p.getObject().ping()) {
            p.getObject().close();
            return false;
        }
        return true;
    }

I mean since the 0 timeout can avoid the timeout issue, why need a extra temporary timeout for each execution?

I mean since the 0 timeout can avoid the timeout issue, why need a extra temporary timeout for each execution?

Time out time Some scenes have to be set, or it may drag down the called application, causing an avalanche
Not set every time because the session can be reused and temporary settings are only required when the session is created

I mean since the 0 timeout can avoid the timeout issue, why need a extra temporary timeout for each execution?

As a simple example, the data in the nebula is weakly dependent on the nebula, but the nebula response time in some case, the response time slows down
If the timeout is not set, the response time of the application will become longer, so that the throughput of the application will reduce, resulting in a large number of timeout upstream, resulting in an avalanche

I mean since the 0 timeout can avoid the timeout issue, why need a extra temporary timeout for each execution?

I wonder if I express it clear

I mean since the 0 timeout can avoid the timeout issue, why need a extra temporary timeout for each execution?

I wonder if I express it clear

Yeah, I got you. But why don't set a monitor to make it timeout in the business layer.

I mean since the 0 timeout can avoid the timeout issue, why need a extra temporary timeout for each execution?

I wonder if I express it clear

Yeah, I got you. But why don't set a monitor to make it timeout in the business layer.

you kan see this issues