The BackupKey Saga (blog post)

Andrew Bartlett abartlet at
Sat Sep 5 06:31:10 UTC 2015

On Fri, 2015-09-04 at 21:34 -0500, Scott Lovenberg wrote:
> On Thu, Sep 3, 2015 at 8:59 PM, Andrew Bartlett <abartlet at>
> wrote:
> > For those interested in how we fixed bug
> > which was raised
> > with
> > credentials manager started to fail after KB2992611, you might be
> > interested in another blog post I wrote up for work:
> > 
> >
> > velopment-deadline
> > 
> > Thanks,
> > 
> > Andrew Bartlett
> > 
> Great post!  I correct a project manager just today in the context of
> a 'black box' service owned by another team that, for one reason or
> another, won't provide us with documentation on the system; he stated
> that any documentation is better than none.  I agreed with the caveat
> that the documentation is correct, otherwise I'd rather have none at
> all!

I certainly appreciate the sentiment.  In this case, and with much of
Samba, I would generally prefer the reverse:  That is, the partial and
often incorrect documentation can be turned into a hypothesis to be
tested.  Because we take an approach akin to the famous 'trust but
verify', we have a method for sifting out the truth.  Even the
apparently false statements give a great structure for making
enquiries, such as by packet analysis, more network based testing or
mailing dochelp.


Andrew Bartlett

Andrew Bartlett             
Authentication Developer, Samba Team
Samba Developer, Catalyst IT

More information about the samba-technical mailing list