When default transaction timeout is redefined during the server lifetime

/subsystem=transactions:write-attribute(name=default-timeout, value=60)
{"outcome" => "success"}

such change is not propagated to WFTC and for example in case of
MDB the default transaction timeout is being left as it was at time of
container startup. The transaction timeout is considered after server is
reloaded. But it's different behavior than it was and in general for
transaction timeout redefinition should not be needed container reload.