glusterfs.git, branch v3.7.1

features/changelog: Remove inline keyword to avoid warnings (gcc v5.1.1)

2015-06-01T13:56:04+00:00

        Backport of http://review.gluster.org/11004

When compiled with gcc5, following warnings were displayed
and volume start failed:

changelog-helpers.h:499:1: warning: inline function 'changelog_dispatch_event'
declared but never defined
 changelog_dispatch_event (xlator_t *, changelog_priv_t *, changelog_event_t *);

gf-changelog-journal-handler.c:692:17: warning: 'list_add_tail' is static but
used in inline function 'gf_changelog_queue_journal' which is not static
                 list_add_tail (&entry->list, &jnl_proc->entries);

Fix is to remove the keyword from function prototype and
definitions.

Change-Id: I188b35b7ca087a94d7a48a052b05a6d845e3b74b
BUG: 1226853
Signed-off-by: Anoop C S 
Reviewed-on: http://review.gluster.org/11030
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Venky Shankar

spec: move file peer_add_secret_pub to server RPM

2015-06-01T12:48:21+00:00

peer_add_secret_pub was developed as part of geo-replication
feature. glusterfind has a dependency on the peer_add_secret_pub
for distributing SSH secrets. The dependency was manifest while
QA testing glusterfind. peer_add_secret_pub is now moved to the
server RPM to support glusterfind without creating a dependency
on the geo-replication RPM

Change-Id: I1345e13fd6d218eb722cd72ca1b9dd04adf4d316
BUG: 1225552
Signed-off-by: Milind Changire 
Signed-off-by: Aravinda VK 
Reviewed-On: http://review.gluster.org/10807
Reviewed-on: http://review.gluster.org/11016
Tested-by: NetBSD Build System 
Reviewed-by: Kotresh HR 
Tested-by: Gluster Build System 
Reviewed-by: Venky Shankar

snapshot/scheduler: Do not enable scheduler if another scheduler is running

2015-06-01T11:13:31+00:00

     Backport of http://review.gluster.org/#/c/10641/

Check if another snapshot scheduler is running before
enabling the scheduler

Also introducing a hidden option, disable_force

"snapshot_scheduler.py disable_force" will disable the
cli snapshot scheduler from any node, even though the node
has not been initialised for the scheduler, as long as the
shared storage is mounted

This option is hidden, because we don't want to encourage
users to use all commands from nodes that are not initialised.

Change-Id: I7ad941fbbab834225a36e740c61f8e740813e7c8
BUG: 1226120
Signed-off-by: Avra Sengupta 
Reviewed-on: http://review.gluster.org/10641
Reviewed-by: Rajesh Joseph 
Tested-by: NetBSD Build System
Reviewed-by: Kaushal M 
(cherry picked from commit d67eb34b2a5b5e3cb926ff4c86a163148743829c)
Reviewed-on: http://review.gluster.org/10993
Tested-by: Gluster Build System 
Reviewed-by: Aravinda VK 
Reviewed-by: Krishnan Parthasarathi 
Tested-by: Krishnan Parthasarathi

dht/rebalance: Change log_level to DEBUG

2015-06-01T09:27:35+00:00

BUG: 1221503
Change-Id: I4ac87bb69e05e2dd445fc0a5bcf48d5bd3d0020b
Reviewed-on: http://review.gluster.org/10756
Reviewed-by: N Balachandran 
Tested-by: Gluster Build System 
Reviewed-by: Shyamsundar Ranganathan 
Signed-off-by: Susant Palai 
Reviewed-on: http://review.gluster.org/10778

cluster/tier: load libgfdb.so properly in all cases

2015-06-01T09:23:37+00:00

We should load libgfdb.so.0, not libgfdb.so

Cherry picked from commit 628406f28364f6019261a3bb37335a494ccf8dda:
> Change-Id: I7a0d64018ccd9893b1685de391e99b5392bd1879
> BUG: 1222092
> Signed-off-by: Dan Lambright 
> Reviewed-on: http://review.gluster.org/10796
> Reviewed-by: Kaleb KEITHLEY 
> Reviewed-by: Joseph Fernandes
> Reviewed-by: Niels de Vos 
> Tested-by: Gluster Build System 

Change-Id: I7a0d64018ccd9893b1685de391e99b5392bd1879
BUG: 1221534
Signed-off-by: Niels de Vos 
Reviewed-on: http://review.gluster.org/10799
Tested-by: Gluster Build System 
Reviewed-by: Dan Lambright 
Reviewed-by: mohammed rafi  kc

glusterd : allocate peerid to store in frame->cookie

2015-06-01T09:23:26+00:00

Backport of http://review.gluster.org/10842

commit a1de3b05 was using peerid from the stack and storing it in the
frame->cookie and in the subsequent callback it was referred. The existance of
this variable is not guranteed in the cbk since its not dynamically allocated.
Fix is to dynmacially manage peerid in the frame cookie.

This patch also fixes one problem in gd_sync_task_begin () where unlock is not
triggered if the cluster is running with lesser than 3.6 op-version resulting into
commands failing with another transaction is in progress.

Change-Id: I0d22cf663df53ef3769585703944577461061312
BUG: 1223215
Signed-off-by: Atin Mukherjee 
Reviewed-on: http://review.gluster.org/10842
Tested-by: NetBSD Build System
Tested-by: Gluster Build System 
Reviewed-by: Kaushal M 
(cherry picked from commit 37f365843bed87728048da1f56de22290f5cb70f)
Reviewed-on: http://review.gluster.org/10959
Reviewed-by: Krishnan Parthasarathi 
Tested-by: Krishnan Parthasarathi

nfs: Use uuid_clear() instead of memset

2015-06-01T09:21:00+00:00

Fixes the following incorrect usage:

mount3.c: In function '__mnt3_build_mountid_from_path':
mount3.c:705:24: warning: 'sizeof' on array function parameter 'mountid' will
return size of 'unsigned char *' [-Wsizeof-array-argument]
         length = sizeof(mountid);
                        ^
mount3.c:699:58: note: declared here
 __mnt3_build_mountid_from_path (const char *path, uuid_t mountid)
                                                          ^
mount3.c: In function '__mnt3_get_mount_id':
mount3.c:732:24: warning: 'sizeof' on array function parameter 'mountid' will
return size of 'unsigned char *' [-Wsizeof-array-argument]
         length = sizeof(mountid);
                        ^
mount3.c:726:46: note: declared here
 __mnt3_get_mount_id (xlator_t *mntxl, uuid_t mountid)

Cherry picked from commit df12f00851a234b0e892af4b958e606a1fa660ed:
> Change-Id: I08f46c5994578fc99a7b61681e808d1115e41d71
> BUG: 1221095
> Signed-off-by: Vijay Bellur 
> Reviewed-on: http://review.gluster.org/10765
> Reviewed-by: Niels de Vos 
> Reviewed-by: Kaleb KEITHLEY 
> Tested-by: Gluster Build System 

Change-Id: I08f46c5994578fc99a7b61681e808d1115e41d71
BUG: 1222198
Signed-off-by: Niels de Vos 
Reviewed-on: http://review.gluster.org/10798
Tested-by: Gluster Build System 
Tested-by: NetBSD Build System 
Reviewed-by: Kaleb KEITHLEY

gfapi: zero size issue in glfs_h_acl_set()

2015-06-01T09:20:23+00:00

When setting the stringified ACLs in the xattr dict through
pub_glfs_h_setxattrs(), the size of the string is always passed as 0.
The correct way is to pass the length of the ACL in text form.

Backport of:
> Change-Id: Ia7a7fa1f3a7d615a813c703057dc97b09a0bbb34
> BUG: 789278
> Reviewd-on: http://review.gluster.org/10782
> Signed-off-by: Jiffin Tony Thottan 

Change-Id: Ia7a7fa1f3a7d615a813c703057dc97b09a0bbb34
BUG: 1224241
Signed-off-by: Niels de Vos 
Reviewed-on: http://review.gluster.org/10890
Tested-by: Gluster Build System 
Reviewed-by: Shyamsundar Ranganathan

quota: quota.conf backward compatibility fix

2015-06-01T09:13:35+00:00

This is a backport of http://review.gluster.org/#/c/10889/

> In release-3.7 the format of quota.conf is changed.
> There is a backward compatibility issues during upgrade
> 1) There can be an issue when peer sync between node-3.6 and node-3.7
> 2) If the user sets/removes limit, there is will different format of
>    file in node-3.6 and node-3.7
>
> This patch fixes the issue:
> 1) restrict the user to execute command quota enable, limit-usage, remove
> 2) write quota.conf in older format if op-version is less than 3.6
>
> Change-Id: Ib76f5a0a85394642159607a105cacda743e7d26b
> BUG: 1223739
> Signed-off-by: vmallika 
> Reviewed-on: http://review.gluster.org/10889
> Tested-by: NetBSD Build System
> Reviewed-by: Krishnan Parthasarathi 

Change-Id: I3d5075b96de1e125d30b0f6045389573aaaaecdb
BUG: 1226153
Signed-off-by: vmallika 
Reviewed-on: http://review.gluster.org/10985
Tested-by: Gluster Build System 
Tested-by: NetBSD Build System 
Reviewed-by: Krishnan Parthasarathi 
Tested-by: Krishnan Parthasarathi

Build: glusterd socket file cleanup to set SElinux context properly.

2015-06-01T08:25:38+00:00

Issue : glusterd runs as rpm_script_t when it's executed from the rpm scriptlet, socket file
created in this context is set as rpm_script_t type, glusterd unable to access socket file
when it runs in glusterd_t context (glusted not cleaning socket file while stoping due to
some cleanup issues, so cleanup required in rpm install).

Fix: In rpm post upgrade,remove the glusterd.socket file which is created by glusterd in rpm context.
so that glusterd recreates socket file when it runs in glusterd_t context.

Backport of:
>Change-Id: I57041e2e0f9d3e74e12198507a9bd47d09f4a132
>BUG: 1223185
>Signed-off-by: anand 
>Reviewed-on: http://review.gluster.org/10815
>Reviewed-by: Niels de Vos 
>Tested-by: NetBSD Build System
>Tested-by: Gluster Build System 

Change-Id: Ic97458f41e1184054cbedbeb547118b217768bbd
BUG: 1222869
Signed-off-by: anand 
Reviewed-on: http://review.gluster.org/10996
Reviewed-by: Niels de Vos 
Tested-by: Niels de Vos