Bug 1265300

Description of problem:
The hardcoded 5 minute timeout period was replace with Integer.MAX_VALUE for the timeout period of "remote" distributed transaction branches.

- https://issues.jboss.org/browse/WFLY-2789
- https://bugzilla.redhat.com/show_bug.cgi?id=1056585#c1

An "infinite" timeout is a problem as this value is also passed to the database to control statement execution. This means that database statements may run uncontrolled in most cases. Too, in one case we have seen that the database driver or the database is mishandling the value and aborting almost immediately (e.g. after only 2 or 3 seconds).

Version-Release number of selected component (if applicable):


How reproducible:
Consistently


Steps to Reproduce:
1. Start a transaction in server "one"
2. Using JTA/EJB remoting, propagate that transaction to a remote server
3. The timeout for the transaction on the remote can be seen to be Integer.MAX_VALUE

Actual results:
Either an "infinite" timeout period for database statement execution or else an immediate abort

Expected results:
A "reasonable" smaller value must be used.

Additional info:
We do not have a test where we can reproduce the immediate abort but we confirmed that using a small value worked with the database/driver.