FileChangelogDB.initializeDB() swallows ChangelogException: a replication server with an unreadable changelog starts anyway

[vharseko]

Spotted while reviewing #795, one line above the swallow that PR removes.

FileChangelogDB.initializeDB()
(opendj-server-legacy/src/main/java/org/opends/server/replication/server/changelog/file/FileChangelogDB.java:282-300)
catches ChangelogException and only logs ERR_COULD_NOT_READ_DB:

public void initializeDB()
{
  try
  {
    replicationEnv = new ReplicationEnvironment(dbDirectory.getAbsolutePath(), replicationServer, TimeService.SYSTEM);
    ...
  }
  catch (ChangelogException e)
  {
    logger.traceException(e);
    logger.error(ERR_COULD_NOT_READ_DB, this.dbDirectory.getAbsolutePath(), e.getLocalizedMessage());
  }
}

ReplicationServer.initialize() calls it and then goes on to bind the listen port and start its
threads, so a replication server whose changelog cannot be read comes up and accepts
connections: the failure surfaces later, somewhere else — the exact shape #792 / #795 removed
for the listen port.

Why it was not fixed in #795

Propagating it means adding throws ChangelogException (or wrapping it in a ConfigException)
to the ChangelogDB interface and to every implementation and caller, which is a change of its
own.

What to do

Decide whether an unreadable changelog must prevent the replication server from starting — as
an unbindable listen port now does — and, if so, propagate the exception through ChangelogDB.