[PATCH] smbd: Fix snapshot query on shares with DFS enabled

Jeremy Allison jra at samba.org
Tue Aug 16 15:40:15 UTC 2016


On Tue, Aug 16, 2016 at 08:05:55AM +0300, Uri Simchoni wrote:
> On 08/16/2016 02:44 AM, Jeremy Allison wrote:
> >>
> >> I will try to take a closer look tomorrow, maybe that can be easily
> >> handled in the shadowcopy2 module. The backport should probably wait
> >> until this is resolved
> > 
> > We should probably add a smbclient-based torture test
> > for both file and directory inside source3/script/tests/test_shadow_copy.sh
> > once this gets fixed to make sure we don't regress.
> > 
> Looking again at the shadow_copy2 test, it's a great step forward vs no
> test, but it is lacking in some areas.
> 
> The current test list previous versions of a file, which queries
> snapshot file information.
> a. Not sure it's complete enough - maybe successfully reading the file
> will be safer
> b. It does not handle directories at all
> c. It only runs on SMB1, and with respect to paths, SMB2 may behave
> differently. Strangely, the test seems to pass on SMB3 (adding -m SMB3),
> although AFAICT smbclient uses SMB1 calls to list snapshots (so it
> should have yielded 0 snapshots). I'll have to look into it.
> 
> The SMB3 test must, IMHO, access snapshot files using TWrp create
> context (I don't mind doing it once I have time and it shouldn't stop
> any bugfixes if it's been reasonably verified that they don't break
> things, so it's a note to self).

For the current change, it would be enough to test over SMB2 using
path with /@GMT-YYYY... appended to the end, as that's exactly
what the server now does.

Further tests can add access to the Twrp context, but I don't
think they're a blocker for adding tests (one for file, one
for directory) for this specific case.



More information about the samba-technical mailing list