Uri Simchoni | 28 May 19:44 2016
Picon

[PATCHES] more quota cleanups - remove EDQUOT checks

Hi,

The attached patch set removes checks for EDQUOT errno in linux-specific
code, in the platform-independent sysquota code (which is the default
VFS implementation), and in disk-free code, which relies on VFS quota
interface.

On some Unices (e.g. HP-UX), there's a convention that the quotactl
system call for getting/setting user/group quota may return an EDQUOT
errno to signal that the request has been fulfilled, but the user/group
is over quota. This convention has trickled into platform-independent code.

The attached patch set removes this convention from platform-independent
code. The interface is thus simplified to the regular unix interface - a
system call returns 0 if and only if it has succeeded. Specific drivers
(sysquotas_4A.c, which supports HP-UX) have the responsibility to return
0 if their underlying OS uses the above mentioned convention and has
returned -1 with EDQUOT as errno. On Linux, such a convention does not
exist and therefore the patch removes the extra checks for EDQUOT on the
sysquota_linux.c driver.

Review appreciated.

Thanks,
Uri.
From 21eab0eb19f4da642ec279285bad5e8450084083 Mon Sep 17 00:00:00 2001
From: Uri Simchoni <uri <at> samba.org>
Date: Thu, 26 May 2016 10:24:58 +0300
(Continue reading)

Kenny Dinh | 27 May 21:59 2016
Picon

TDB lock contention during "startup" event caused winbind crash

Hello,

I ran into a situation where contention for tdb lock caused "winbind" to
crash. Below is the scenario

-- ctdb process                       smbd process           winbindd
process

-- "service start winbind"------------------------------         winbindd
 started (pid 14461)
-- "service start smb" ------ smbd started ( pid 14602)
-------------------------------------- smbd acquires lock g_lock.tdb - hung
-- invoke hung script ------  smbd child pid (14602) is still hung
-- ----------------------------------- smbd still lock g_lock.tdb
-- CTDB restart all services
-- kill existing winbindd ------------------------------------------
winbindd (pid 14461 term)
-- service start winbind ---------------------------------------- winbindd
started (pid 14733)
-- 
 -----------------------------------------------------------------------try
to lock g_lock.tdb but failed
--

-- 
 ------------------------------------------------------------------------
PANIC
--  kill existing smbd -------- Kill pid 14602
--  service start smb ------- smbd started

Attached are log file from ctdb, winbind, and smb, and winbind core
backtrace.
(Continue reading)

Jeremy Allison | 27 May 20:48 2016
Picon

[PATCH] Small cleanups struct sockaddr -> struct sockaddr_storage

So after noticing the cleanup in Volker's patch,
I did:

git grep 'struct sockaddr ' | grep -v 'struct sockaddr \*'

and here are the 2 cleanups I found that were needed.

Please review and push if you agree. The smbldap one,
is that worth a bug report ?

Comments welcome,

Jeremy.
From 8421d481afcaddb5b74c6af0f0a94eeff8787579 Mon Sep 17 00:00:00 2001
From: Jeremy Allison <jra <at> samba.org>
Date: Fri, 27 May 2016 11:42:58 -0700
Subject: [PATCH 1/2] s3: lib: ldap: Use struct sockaddr_storage to cope with
 IPv6.

Signed-off-by: Jeremy Allison <jra <at> samba.org>
---
 source3/lib/smbldap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/source3/lib/smbldap.c b/source3/lib/smbldap.c
index 5a876e0..e9cc48b 100644
--- a/source3/lib/smbldap.c
+++ b/source3/lib/smbldap.c
(Continue reading)

Volker Lendecke | 27 May 17:01 2016
Picon

[PATCH] prepare streams based messaging

Hi!

Attached find two patches: The first is a set of patches preparing the
second one. "look" cleans up some dependencies, making some
infrastructure more generally useful. I'd like to put that in now.
Review appreciated!

The second one, "tfd", is the start of what I've talked about  <at> 
sambaxp. It's the stream socket introducer. Our plan is to also use it
for ctdb communication, I've already done some changes after talking
to Amitay about it. So, not really used yet, but if you're curious,
feel free to take a look :-)

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
http://www.sernet.de, mailto:kontakt <at> sernet.de

SerNet & BSI laden ein: 29. Juni 2016,
2. IT-Grundschutztag 2016, BPA Berlin.
Anmeldung: https://www.sernet.de/gstag
From 51c3ec668e10bdb6a6e49c999c1422b4adfb3c73 Mon Sep 17 00:00:00 2001
From: Amitay Isaacs <amitay <at> gmail.com>
Date: Sun, 15 May 2016 21:43:26 +1000
Subject: [PATCH 1/7] lib/async_req: Reduce the dependency to tevent_unix_util
(Continue reading)

Abhidnya Joshi | 27 May 11:26 2016

[PATCH] Efficient xattr handling for VxFS

Hi Jeremy, list,

Please find attached patch which handles storing xattrs in efficient way for Veritas File System (VxFS)
using VxFS library.

This patch handles:
1.            Storing xattrs using VxFS library calls
2.            Handling backward compatibility where if xattrs are stored with old way then store them with new way .
3.            This makes use of vxfsmisc.so in user space which is provided by VxFS and is open sourced. It can be found
at: http://www.veritas.com/community/downloads/vxfsmisc-library

This patch applies to Samba master. Please let me know your comments.

Thanks and Regards
Abhidnya Joshi

Andrew Bartlett | 27 May 08:00 2016
Picon

[PATCH] Fix replication, re-sort replPropertyMetaData and add many tests

This patch series, is at:

git://git.catalyst.net.nz/samba.git repl_move-replPropertyMetaData

http://git.catalyst.net.nz/gitweb?p=samba.git;a=shortlog;h=refs/heads/r
epl_move-replPropertyMetaData

This patch series addresses a large number of issues in our replication
code, particularly that we did not correctly honour the parentGUID when
there was no apparent change in the DN, and that we would sometimes
have a 'name' attribute that did not match the RDN.

Finally, we find that our replPropertyMetaData was incorrectly sorted,
and did not match Windows behaviour for our RDN so we fix that.

The code has comprehensive tests, and earlier versions have passed many
autobuilds without flapping in the replication code.  The latest code
is again under test in the Catalyst Cloud. 

Please look at this carefully, if you have time, as this changes some
base assumptions that we had, which I show to be unfounded.  I plan to
merge it late next week after peer-review.

This should address a number of our flapping tests and non-
deterministic behaviours, and make Samba replication much more reliable
for Samba 4.5.

The patch is, at this stage, ready to go, except that I need to decide
on if this patch will have the HACK tag removed, or a better solution
implemented:
(Continue reading)

Jeremy Allison | 27 May 01:38 2016
Picon

[PATCH] Fix out-of-scope variable reference.

This *may* be the root cause of:

https://bugzilla.samba.org/show_bug.cgi?id=10618

(I'm going to get the users to test).
But it's certainly an obvious flaw.

Review appreciated !

Cheers,

Jeremy.
Uri Simchoni | 26 May 23:16 2016
Picon

[PATCHES] s3-smbd: do not clamp free disk reporting with quota if quota is not enforced

Hi,

The attached patch set fixes free-space-reporting on an XFS file system
where quota accounting is enabled but quota enforcement is disabled.

Out of the supported file systems and OS's, it seems like only XFS has a
separate notion of quota enforcement and accounting. An alternative fix
could have been to just fix the XFS driver. However, this would break
getting the user's quota, which should not be affected by whether the
quota is actually being enforced (I verified this on a Windows 2008R2
server).

Because code common to all file systems is being changed, I verified
that it doesn't break ext4 on Linux, ufs on FreeBSD (the 4B driver), and
NFS. Haven't tested the 4A driver though.

Review appreciated.
Thanks,
Uri.
From 65ee7348c7b47b9effdc9af09f0fd475666836ac Mon Sep 17 00:00:00 2001
From: Uri Simchoni <uri <at> samba.org>
Date: Thu, 26 May 2016 21:59:38 +0300
Subject: [PATCH 1/3] vfs_fake_dfq: add more mocking options

Add support for mocking FS user/group quotas (default quota and
quota flags).

Make the default block size 4096 instead of 0. This
(Continue reading)

Volker Lendecke | 26 May 16:56 2016
Picon

Re: Patch for a memory leak issue in share mode locking.

On Thu, May 26, 2016 at 12:31:49AM +0000, Saji VR wrote:
> Jeremy,
>  That exactly was my analysis too…
> 
>  Here is what I sent to Volker today morning trying to explain it…
> 
> ————————
> 
> Volker,
> 
> Thanks for the response, Hemanth posted this on my behalf as I am not active on samba (Yet!)

I wonder what will happen when you get active on Samba given the depth
of this analysis while not being active .... :-)

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
http://www.sernet.de, mailto:kontakt <at> sernet.de

SerNet & BSI laden ein: 29. Juni 2016,
2. IT-Grundschutztag 2016, BPA Berlin.
Anmeldung: https://www.sernet.de/gstag

Noel Power | 26 May 12:45 2016

RFC rawpipe

Hi,
I'm trying once again to get my head around the WSP stuff, one of the
prerequisites for that is working with rawpipes, some time ago I posted
some patches around dealing with such rawpipes that was too heavy weight
(see 
https://lists.samba.org/archive/samba-technical/2014-July/100884.html)
See attached patches that attempt to rework this,

The patches provide a simple framework to support an internal/external
service that wants to use such rawpipes. The patches also provide a very
simple reference implementation (for testing), note: this test
implementation is only built with developer builds, also there are also
some simple torture test suite added (but not so simple to figure out
how to write ;-))

I confess that I didn't understand all of what was said in the thread
above but I think I have captured the spirit of if (and am happy to work
to do whatever is needed to clean it further).

Note: Also included in there is a patch for adding the ability to change
the default 'Max Ioctl' for pipe trans, that patch is not strictly a
perquisite for the WSP server but is necessary for any test clients that
will use the WSP service.

thanks
Noel
From 0b677b91b1ce196b21788093e384e1a4e2b7f20b Mon Sep 17 00:00:00 2001
From: Noel Power <noel.power <at> suse.com>
(Continue reading)

Karolin Seeger | 26 May 10:35 2016
Picon

[Release Planning 4.3] Samba 4.3.10 on June 15

Hi,

Samba 4.3.10 is scheduled for Wednesday, June 15 2016.

https://wiki.samba.org/index.php/Release_Planning_for_Samba_4.3
has been updated accordingly.

Cheers,
Karolin

--

-- 
Karolin Seeger			https://samba.org/~kseeger/
Release Manager Samba Team	https://samba.org
Senior Consultant SerNet	https://sernet.de


Gmane