Bacula fails on FreeBSD 10.x / "mt fsf" infinitely proceeds
Kenneth D. Merry
ken at FreeBSD.ORG
Thu Jul 31 21:17:04 UTC 2014
On Thu, Jul 31, 2014 at 22:17:42 +0200, Joerg Wunsch wrote:
> As Joerg Wunsch wrote:
>
> > Next attempt: in saerror(), rather than basing the decision whether
> > this status report belongs to a read/write or control operation on
> > some CCB flag that needs to be set explicitly, base it on the actual
> > command that led to the status report (SA_READ or SA_WRITE
> > vs. anything else). A similar decision is already done in saerror().
>
> Now, tested for real, too.
>
> I am content with that patch, and would be willing to commit it.
Okay, it looks good to me; go ahead and commit it.
> Note that much of the changes there are whitespace changes only since
> I removed a switch/case statement where Alexander Motin only left a
> single case in, so a large block of code is unindented by one TAB now.
>
> Here's a typescript of handling a real Bacula tape with the suggested
> patch.
Great, thanks for tracking this down!
Ken
--
Kenneth Merry
ken at FreeBSD.ORG
More information about the freebsd-scsi
mailing list