I'm surprised that they still think it is Bacula
after my last email. In my tests, I used Bacula,
btape simulating Bacula calls, and a loop in
btape doing nothing but write(), and in all
cases, I got data loss.
I suppose that if they don't want to deal with
the problem, I can write a simple stand alone
program that shows the problem. That should
get their attention, but I don't understand why
that should be necessary.
I spent a lot of time devising/running/analyzing
the tests, and I would like to know what
concrete steps they have taken to show that
my results were not valid.
In addition, Bacula runs fine on Linux and Sun=20
Can they explain why the data loss at EOM occurs=20
only on FreeBSD?
Let me know if you think it is necessary
to write a simple program.
On Mon, 2003-07-21 at 17:18, Dan Langille wrote:
> What needs to be done to get the FreeBSD SCSI issues resolved? It=20
> sounds like both Bacual and FreeBSD think the issue lies with the=20
From: Dan Langille <dan@la...> - 2003-07-21 16:09:56
On 21 Jul 2003 at 17:50, Kern Sibbald wrote:
> Let me know if you think it is necessary
> to write a simple program.
I think a simple example program which they can run themselves and
which demonstrates the problem in a very easy to understand situation
would be useful. That would give everyone a very clear and concise
understanding of the problem and removes any possibility that someone
could say "it's Bacula".
Dan Langille : http://www.langille.org/