sin/samba-mirror
1
0
Fork 0
mirror of https://github.com/samba-team/samba.git synced 2024-12-23 17:34:34 +03:00
Code

r23261: Merge WHATSNEW back into the main branch. Comments/omissions greatly

Browse Source 
appriciated.

Andrew Bartlett
(This used to be commit f803e563cb)
This commit is contained in:
Andrew Bartlett 2007-05-31 05:05:45 +00:00 committed by Gerald (Jerry) Carter
parent c46e5868de
commit 57be767a24
1 changed files with 37 additions and 57 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

94
WHATSNEW.txt
View File

@ -1,69 +1,49 @@
'Samba4 TP4' presents you with an opportunity to see a Technology
Preview (TP) snapshot of Samba4's development, as at January 2007.
'Samba4 TP5' presents you with a snapshot into Samba4's ongoing
development, as we move towards our first alpha releases. This Technology
Preview (TP) is snapshot of Samba4's development, as at June 2007.
In the last few months since TP3 was released in October 2006,
significant work has been done across many parts of Samba4. Since that
time, we have added the basis for some new and exciting features:
In the time since TP4 was released in January 2007, Samba has
continued to evolve, but you may particularly notice these areas:
PKINIT support to Samba4's KDC will allow, smart-card login to a
Samba4 domain. TP4 demonstrates this with static key files, but
work will continue to enable actual hardware cards.
Work has continued on SWAT, the the libnet API behind it. These we
hope will grow into a full web-based management solution for both
local and remote Samba and windows servers.
Clustering support was always a design goal of Samba4, and with TP4
we have the ctdb framework, a cluster-aware shared database. This
allows Samba4 to share a shared cluster file-system with it's clients.
Presented at this year's linux.conf.au, including a highly rigged
demo, you can expect to see this mature over the next few months.
The DRAUAPI research effort has largely concluded, and an initial
implementation of AD replication is present, included in torture
test-cases. This includes the decryption of the AD passwords, which
were specially and separately encrypted. This should be recognised
as vital milestone.
Non-blocking and Asynchronous IO support, has always been a design
goal in Samba4, and TP4 will use new Linux Kernel features to
implement event driven asynchronous IO. This makes Samba more
efficient on systems where some data may be 'further away' than a
local disk, such as HSM systems. This allows the Kernel to handle
reading the returned data from the disk, only notifying Samba when
the data is ready for dispatch to the client.
Likewise, the LDAP Backend project has moved from a research
implementation into something that can be easily deployed outside
the test infrastructure.
Our web-management console, known as SWAT, is being revamped, and in
TP4 you can find a new Web 2.0 style user interface, being used to
support a web-based ldb browser. We hope this new system will allow
things simple not possible with the form-submit style of web
management.
Testing has been an area of great work, with renewed vigour to
increase our test coverage over the past few months. In doing so,
we now demonstrate PKINIT and many other aspects of kerberos, as
well as command-line authentication handling in our testsuite.
Using LDB LDAP back-end integration has improved in this release, with an
improved mapping module allowing the start of Fedora DS back-end
support.
The testsuite infrastructure has been rewritten in perl and
extended, to setup multiple environments: allowing testing of the
domain member, as well as the domain controller, roles. Samba4's
initial implementation of winbind has been revived, to fill in these
tests.
In continuing our research effort, TP4 includes the work to better
understand and implement the DRSUAPI replication protocols. By better
understanding the needs of replication now, we can structure our
databases so that their format will have to change less in future.
In clustering, work on CTDB (an implementation of a clustered Samba)
has moved ahead very well, but the current code has not
been merged into Samba4 in time for this release.
We hope to use this replication function to replace the SamSync based
Vampire process so effectively demonstrated since TP1, and to
eventually join an Active Directory domain, as a replicating partner.
To support better management, we have investigated group policy
support, and include the infrastructure required. Unfortunately
without MMC write support, you will need to place the polices into
the directory by hand.
Behind the scenes, much of the core infrastructure of Samba4 continues
development:
In Kerberos, we have continued to track the development of the
Heimdal Kerberos implementation, and reduce the custom diff between
our branch and upstream. Heimdal now provides plug-in APIs for
almost all of the hooks we need, including management and validation
of the PAC.
In testing, our test infrastructure has undergone a quiet
revolution, as we improve our unit test framework. Likewise, the
tests themselves have continued to expand, as we follow our
test-driven development pattern.
In providing an abstraction above our raw RPC layer, the libnet
library continues to expand, becoming a C and JS management API for
Samba4 and remote servers.
To ensure that, as an administrator and developer, you can easily
read and edit our internal databases, our LDB layer has been
optimised for speed. The aim here is to avoid needing to use the faster, but
more opaque, TDB layer.
As we move forward, we have many of the features we feel are required
for a Samba4 Alpha. Similarly, we know enough about the data
formats (particularly those that are encrypted) to be confident that
we won't need to change the LDB format. Our plan is to publish a
Samba4 alpha in the next few months.
These are just some of the highlights of the work done in the past few
months. More details can be found in our SVN history.