Am Mittwoch, 8. März 2017 20:51:32 UTC+1 schrieb Eric MSP Veith:
> Jörg,
> 
> Am Samstag, 4. März 2017 21:52:45 UTC+1 schrieb Jörg Steffens:
> > I always used bpipe for this kind of purpose. However, I assume,
> > readfifo is working properly, as there is a working regression test for it:
> > https://github.com/bareos/bareos-regress/blob/master/tests/fifo-test
> 
> the regression test works because the whole testsuite is run on the same 
> machine. Bareos (and Bacula, too?) backs up the fifo, but never reads from 
> it. The program feeding into the FIFO waits patiently until the *verifcation 
> script* reads from it.
> 
> What happens is, in effect, the same as:
> 
> mkfifo /tmp/some.fifo
> cat /boot/vmlinuz > /tmp/some.fifo
> cmp /boot/vmlinz /tmp/some.fifo
> 
> Bareos never reads the FIFO. I fear that this might be a bug.

Sorry, this was quite short.

To verify my assumption, I created a simple FileSet that contains only one 
FIFO. I used three separate machines: One Director + SD, one FD (read, FIFO 
initially created here) and one other FD as restore target. When I backup from 
and restore to the same FD, the FIFO works. When I backup from one FD and 
restore to another, a read from the FIFO returns nothing.

Could you please verify this?

TIA!

--- Eric

-- 
You received this message because you are subscribed to the Google Groups 
"bareos-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to