On Thu, Mar 17, 2005 at 12:05:21PM -0500, David Magda wrote:
> We have multiple jobs that use the virtual hostname: they all run
> fine to the virtual hostname, but then try to run to the real
> hostname and return the 239 error code.

Another datapoint from an observation I just made because of an off
list thread about this.

We backup the following directories:

/u02/oradata/DBNAME
/u03/oradata/DBNAME
...
/u13/oradata/DBNAME

Some databases have all the /u?? direcotories defined in one policy
(they're smaller databases), while other databases have the /u??
directories broken up into three policies (somedb_blib, somedb_blib1,
somedb_blib2). In the past it was found that the backups went faster
if we ran the streams in parallel.

What I just noticed is that even though we may have three policies to
get all the files for one database (blib, blib1, blib2), only one of
the policies (blib, with no number) gets the 239 error. The other
policies (blib1, blib2) do not show up with the 239 error.

I'm not sure whether this is significant but I though I'd mention
it.

An individual mentioned offlist that the data stream may be binding
to a particular IP to send data back to the media server. I'll try to
see if this is the case.