VigneshDhanraj G | 1 Dec 08:18 2015

NFSV4 Client setup problem


I tried to bring up nfsv4 client setup, but when i joining AD server from
my LINUX machine i always get below error

"kinit succeeded but ads_sasl_spnego_krb5_bind failed: Server not found in
Kerberos database
Failed to join domain: failed to connect to AD: Server not found in
Kerberos database"

wbinfo -u command gives the user list
net ads info gives the details of the AD

when i tried to login from AD administrator user i am not able to login
using ssh.

i am using debian wheezy as client and windows 2003 Server as AD.

my samba conf
        security = ADS
        realm = INDIA.LOCAL
# If the system doesn't find the domain controller automatically, you may
need the following line
        password server = INDIA.LOCAL
# note that workgroup is the 'short' domain name
        workgroup = INDIA
#       winbind separator = +
        winbind refresh tickets = yes
        winbind enum users = yes
(Continue reading)

Richard Sharpe | 30 Nov 23:59 2015


Hi folks,

my copy of [MS-SMB].pdf contains these words WRT DFS referrals:

----------------------------------------------- Application Requests DFS Referral Information

The client MUST search for an existing Session and TreeConnect to any
share on the server identified by ServerName for the user identified
by UserCredentials. If no Session and TreeConnect are found, the
client MUST establish a new Session and TreeConnect to IPC$ on the
target server as described in section using the supplied
ServerName and UserCredentials.

However, Windows clients seem not to do that and always seem to use
IPC$ for DFS referrals even when they have connections open on the
target server.

Has anyone seen any cases where Windows clients use shares other than IPC$?


Richard Sharpe

Jacob Holtom | 30 Nov 19:42 2015

[PATCH] smbd: add reparse points at symlinks


This patch adds a reparse point at symlinks, with the correct 
accompanying tag as defined by microsoft.
this may be better implemented with a compile time option of smb.conf 
flag, review appreciated

Martin Schwenke | 30 Nov 10:30 2015

[PATCH] CTDB IP allocation separation - nearly done

This is the last of the patches before CTDB's IP allocation code can be
separated into a separate build subsystem.

Still to come or to-do:

* The patch that does the separation is not included here because it is
  quite large (just lots of code movement).  I'll post it separately
  once this batch goes in.

* Then there will be some cleanup to try to make the "ipalloc" subsystem
  no longer depend on the old ctdb_protocol.h.

* Write the a standalone helper, using only new protocol/client code, to
  do the actual IP allocation/takeover run.

* Then back to the election code...  or perhaps this and the item
  above in parallel... :-)

Please review and push...

peace & happiness,
Attachment (ctdb.patches): application/octet-stream, 59 KiB
Martin Schwenke | 30 Nov 06:36 2015

[PATCH] CTDB connection tracking (tickle) improvements

* ctdb-daemon: Do not process tickle updates for hosted IP addresses

  Fix a potential race in the tickle transfer code.  I can't believe
  I've hung onto this for so long...  :-(

* ctdb-daemon: Move VNN lookup out of ctdb_remove_tcp_connection()
  ctdb-daemon: Don't delete connection information for released IP

  Future-proof connection tracking code

Please review and push...

peace & happiness,
Attachment (ctdb.patches): application/octet-stream, 7944 bytes
Martin Schwenke | 30 Nov 02:41 2015

[PATCH] CTDB updates

4 different patches, for variety...  :-)

* ctdb-daemon: Drop handling of ban control sent to unexpected node

  Drop unused code

* ctdb-recoverd: Drop function unban_all_nodes()

  Drop no-op code

* ctdb-scripts: Add exportfs cache to NFS Linux kernel callout

  Add caching to avoid a performance problem

* ctdb-docs: Rewrite event script documentation

  Update old documentation

Please review and push...

peace & happiness,
Attachment (ctdb.patches): application/octet-stream, 31 KiB
Noel Power | 27 Nov 16:38 2015

small valgrind fix


valgrind was complaining [*] on winbindd startup (sometimes) maybe there
is a better solution that avoids this kindof ugly patch (see attached)



==7913== Invalid read of size 1
==7913==    at 0xC4F23EE: smb2cli_ioctl_done (smb2cli_ioctl.c:245)
==7913==    by 0x747A744: _tevent_req_notify_callback (tevent_req.c:112)
==7913==    by 0x747A817: tevent_req_finish (tevent_req.c:149)
==7913==    by 0x747A93C: tevent_req_trigger (tevent_req.c:206)
==7913==    by 0x7479B2B: tevent_common_loop_immediate
==7913==    by 0xA9CB4BE: run_events_poll (events.c:192)
==7913==    by 0xA9CBB32: s3_event_loop_once (events.c:303)
==7913==    by 0x7478C72: _tevent_loop_once (tevent.c:533)
==7913==    by 0x747AACD: tevent_req_poll (tevent_req.c:256)
==7913==    by 0x505315D: tevent_req_poll_ntstatus (tevent_ntstatus.c:109)
==7913==    by 0xA7201F2: cli_tree_connect (cliconnect.c:2764)
==7913==    by 0x165FF7: cm_prepare_connection (winbindd_cm.c:1276)
==7913==  Address 0x16ce24ec is 764 bytes inside a block of size 813 alloc'd
==7913==    at 0x4C29110: malloc (in
==7913==    by 0x768A0C1: __talloc_with_prefix (talloc.c:668)
==7913==    by 0x768A27E: _talloc_pool (talloc.c:721)
==7913==    by 0x768A41E: _talloc_pooled_object (talloc.c:790)
(Continue reading)

Volker Lendecke | 26 Nov 21:09 2015

[PATCH] loadparm: Do not talloc empty params


Review appreciated!

Thanks, Volker

SerNet GmbH, Bahnhofsallee 1b, 37081 Göttingen
phone: +49-551-370000-0, fax: +49-551-370000-9
AG Göttingen, HRB 2816, GF: Dr. Johannes Loxen, mailto:kontakt <at>
From d79436f338f1693b0fc83210fc66c9672301bbcb Mon Sep 17 00:00:00 2001
From: Volker Lendecke <vl <at>>
Date: Thu, 26 Nov 2015 13:23:41 +0100
Subject: [PATCH] loadparm: Do not talloc empty params

Reduce memory footprint for large smb.conf files. With 1400 shares this
gets smbd -i from 7259kb to 3087kb according to 64-bit x86 massif. If
you want to serve 1000 clients, this saves you 4GB.

Signed-off-by: Volker Lendecke <vl <at>>
 lib/param/loadparm.c     | 15 ++++++++++-----
 source3/param/loadparm.c |  2 +-
 2 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/lib/param/loadparm.c b/lib/param/loadparm.c
index 640c602..13ca2b5 100644
(Continue reading)

Karolin Seeger | 26 Nov 11:35 2015

[PATCH] Fix typos in the idmap backend section


please find attached a patch to fix some typos in man smb.conf.5.
Please feel free to review/push... ;)



Karolin Seeger
Release Manager Samba Team
Senior Consultant SerNet

Jiri Sasek - Solaris Prague | 26 Nov 06:20 2015

Do anyone here have the experience with 64bits compiled Samba on SPARC?

Hi geeks,

I have built 64bits Samba and winbindd is receiving the SIGBUS(*)

...on:`talloc_dict_traverse_fn+0xf0:   ldx [%l2], %l3

where l2 is equal to 2b50c7275c as can be seen from stackregs trace:
fffffdfa7f401190, fffffdfa7f400bf8, fffffdfa7f400bd8, 1, fffffdfa7f400c20)
   %l0-%l3:    7ffc619014800               44 2b50c7275c               44
   %l4-%l7:       2b50c72718               44 2b50c72718               44`db_rbt_traverse_internal+0x198:jmpl %l3, %o7

> 2b50c7275c,10::dump
               0 1 2 3  4 5 6 7  8 9 a b \/ d e f 0123456789abvdef
2b50c72750:  00000000 00000000 00000000 0000002b ...............+
2b50c72760:  50c73170 000a8750 0000002b 50c72500 P.1p...P...+P.%.

So I suppose the problem is the 0x2b50c7275c address is not the 64bits 

I have also localized the problem is (with some [RISC optimized code] 
uncertainty :-) ) in samba-4.1.17/source3/lib/talloc_dict.c 
(talloc_dict_traverse_fn) line:
(Continue reading)

Andreas Schneider | 25 Nov 12:24 2015

[PATCH] Array out of bound access


gcc 5.1.1 is complaining:

lib/util/idtree.c:184:15: error: array subscript is above array bounds

attached is a patch addressing the issue.

Review and puth much appreciated!

	-- andreas


Andreas Schneider                   GPG-ID: CC014E3D
Samba Team                             asn <at>
Attachment (guxtdu.patch): text/x-patch, 855 bytes