glusterfs.git, branch v3.7.8

cluster/dht: Cleanup dict in dht_do_rename()

2016-02-08T13:26:19+00:00

Change-Id: Ib4b3a843e78eccf5b8e0e7776cd0128013a59a3e
BUG: 1304889
Signed-off-by: Vijay Bellur 
Reviewed-on: http://review.gluster.org/13322
CentOS-regression: Gluster Build System 
Smoke: Gluster Build System 
Reviewed-by: Kaleb KEITHLEY 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Raghavendra G 
Reviewed-on: http://review.gluster.org/13358

Revert "glusterd/afr: store afr pending xattrs as a volume option"

2016-02-08T08:18:23+00:00

This reverts commit 486b07dfc33782d27e3458659cdd6090f496ad35.

Change-Id: I2da12ef77fb01c52cae3c3b5547398b6a4d7d17e
BUG: 1293536
Signed-off-by: Pranith Kumar Karampuri 
Reviewed-on: http://review.gluster.org/13339
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System

Revert "afr: warn if pending xattrs missing during init()"

2016-02-05T07:53:28+00:00

This reverts commit 101243860bca1bbda8a03ab76f6753ecc10799ed.

Change-Id: I042b251250515d703b5bc8bac55e0fc49f2dfc69
BUG: 1293536
Signed-off-by: Pranith Kumar Karampuri 
Reviewed-on: http://review.gluster.org/13338
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System

cluster/afr: Fix heal-info slow response while IO is in progress

2016-02-05T04:31:54+00:00

        Backport of: http://review.gluster.org/#/c/13326/

Now heal-info does an open() on the file being examined so that
the client at some point sees open-fd count being > 1 and releases
the eager-lock so that heal-info doesn't remain blocked forever
until IO completes.

Change-Id: I7d4a8aa4de459216408b666894ee7bb42e406547
BUG: 1303899
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/13348
Smoke: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

fuse: fix inode and dentry leaks

2016-02-04T07:55:34+00:00

When a readdirp was executed, the nlookup count for each inode of the
returned entries was incremented. However the kernel does not increment
the counter for '.' and '..' entries.

This caused that kernel sent forgets with a counter smaller than the
inode's current value. This prevented these inodes to be retired when
ref count was 0.

Change-Id: I36f3736c86d1cc20adabbbc133ce4edeff62db78
Signed-off-by: Xavier Hernandez 
Reviewed-on: http://review.gluster.org/13324
Smoke: Gluster Build System 
Reviewed-by: Raghavendra Bhat 
Tested-by: Kaleb KEITHLEY 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Kaleb KEITHLEY

build: use 'make install' to install the hook scripts

2016-02-04T06:11:08+00:00

This is a backport of http://review.gluster.org/#/c/13072

The installation should be the same on all distributions, and doing
manual installation of files in the .spec is very ugly. This change adds
the rules so that 'make install' places the hook scripts in the right
location.

Also, the hook script(s) for NFS-Ganesha should be part of the
glusterfs-ganesha sub-package and got moved there.

> BUG: 1174765
> Change-Id: Iba25a7a5112c7d40db4c10ff4a5ac7a5fb4f7c4e
> Signed-off-by: Niels de Vos 
> Reviewed-on: http://review.gluster.org/13072
> Tested-by: Gluster Build System 
> Reviewed-by: Kaleb KEITHLEY 

Change-Id: I8fade7e9e540eedc1e619f116d356a2b26f599a9
BUG: 1302955
Signed-off-by: vmallika 
Reviewed-on: http://review.gluster.org/13313
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System 
Reviewed-by: Kaleb KEITHLEY

marker : handle variable `ret` properly in marker_unlink_cbk

2016-02-03T05:20:04+00:00

Backport of http://review.gluster.org/#/c/12584/

While compiling new source, the following issue got hit :
marker.c: In function 'marker_unlink_cbk':
marker.c:994:29: warning: variable 'ret' set but not used
         int32_t             ret     = 0;
This patch will fix the same.

Upstream reference :
cherry picked from commit eada31c88f929768df5fa337137b2ffcd57f2e73
>Change-Id: I0de60bed3351b3aa1cc80f52d178e447826210e7
>BUG: 1257694
>Signed-off-by: Jiffin Tony Thottan 
>Reviewed-on: http://review.gluster.org/12584
>Reviewed-by: Vijaikumar Mallikarjuna 
>Tested-by: NetBSD Build System 
>Tested-by: Gluster Build System 
>Reviewed-by: Raghavendra G 

Change-Id: Ia8f87131e90599cad018cd31269de68049a2ade9
BUG: 1279331
Signed-off-by: Jiffin Tony Thottan 
Reviewed-on: http://review.gluster.org/12596
Reviewed-by: Manikandan Selvaganesh 
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Raghavendra G

fuse: use-after-free fix in fuse-bridge, revisited

2016-02-03T05:16:04+00:00

Prompted by the email exchange in gluster-devel between Oleksandr
Natalenko, xavi, and soumyak, I looked at this because the fuse client
on the longevity cluster has also been suffering from a serious memory
leak for some time. (longevity cluster is currently running 3.7.6)

The longevity cluster manifests the same kernel notifier loop terminated
log message the Oleksandr sees, and some sample runs suggest that the
length passed to the (sys_)write call is unexpectedly and abnormally large.

Basically this fix
 a) uses correct types for len and rv,
 b) copies the len from potentially incorrectly aligned memory (in a
    way that should minimize potential performance issues related to
    accessing unaligned memory.)
 c) changes log level of the kernel notifier loop terminated message
 d) fixes a potential mutex lock/unlock issue

See
    Change-Id: Icedb3525706f59803878bb37ef6b4ffe4a986880
    BUG: 1288857
    http://review.gluster.org/13274

Change-Id: I1a2f957a95978468baa33c2b1f3667934d88939c
BUG: 1288922
Signed-off-by: Kaleb S KEITHLEY 
Reviewed-on: http://review.gluster.org/13275
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Raghavendra G

tests: Fix spurious failure in bug-1221481-allow-fops-on-dir-split-brain.t

2016-02-02T10:47:28+00:00

        Backport of: http://review.gluster.org/#/c/13172/

Occasionally, when ls is executed, prior to READDIRP, a STAT is wound
on the operand directory. And AFR fails STAT with EIO if it is in
metadata split-brain which "dir" is in the test case in question.
As a result, ls also fails with EIO, causing test 20 to return negative
exit status.
The fix is in the test script where the parts that cause the dir to go
into metadata split-brain have been removed. Now "dir" will only have
entry split-brain.

Change-Id: Icf3998ad6f8735c283171e22445406a2eaaaa23f
BUG: 1296400
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/13190
Smoke: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

glusterd: improve error logs for unsupported clients

2016-02-01T11:03:12+00:00

Backport of http://review.gluster.org/11831

Change-Id: I90be46a7b25cc3efd0e61c9a145852e9027f5f64
BUG: 1302521
Signed-off-by: Atin Mukherjee 
Reviewed-on: http://review.gluster.org/11831
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri 
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Prashanth Pai 
Reviewed-by: Kaushal M 
NetBSD-regression: NetBSD Build System 
Reviewed-on: http://review.gluster.org/13305