glusterfs.git/xlators/mgmt/glusterd, branch v3.7.3

glusterd: initialize the daemon services on demand

2015-07-27T14:31:30+00:00

Backport of http://review.gluster.org/#/c/11488/

As of now all the daemon services are initialized at glusterD init path. Since
socket file path of per node daemon demands the uuid of the node, MY_UUID macro
is invoked as part of the initialization.

The above flow breaks the usecases where a gluster image is built following a
template could be Dockerfile, Vagrantfile or any kind of virtualization
environment. This means bringing instances of this image would have same UUIDs
for the node resulting in peer probe failure.

Solution is to lazily initialize the services on demand.

Change-Id: If7caa533026c83e98c7c7678bded67085d0bbc1e
BUG: 1247012
Signed-off-by: Atin Mukherjee 
Reviewed-on: http://review.gluster.org/11488
Tested-by: Gluster Build System 
Tested-by: NetBSD Build System 
Reviewed-by: Gaurav Kumar Garg 
Reviewed-by: Kaushal M 
Reviewed-on: http://review.gluster.org/11766

rpc,server,glusterd: Init transport list for accepted transport

2015-07-26T17:36:39+00:00

GlusterD or a brick would crash when encrypted transport was enabled and
an unencrypted client tried to connect to them. The crash occured when
GlusterD/server tried to remove the transport from their xprt_list due
to a DISCONNECT event. But as the client transport's list head wasn't
inited, the process would crash when list_del was performed.

Initing the client transports list head during acceptence, prevents this
crash.

Also, an extra check has been added to the GlusterD and Server
notification handlers for client DISCONNECT events. The handlers will
now first check if the client transport is a member of any list.
GlusterD and Server DISCONNECT event handlers could be called without
the ACCEPT handler, which adds the transport to the list, being called.
This situation also occurs when an unencrypted client tries to establish
a connection with an encrypted server.

Change-Id: Icc24a08d60e978aaa1d3322e0cbed680dcbda2b4
BUG: 1246809
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.org/11692
Tested-by: Gluster Build System 
Tested-by: NetBSD Build System 
Reviewed-by: Raghavendra G 
(cherry picked from commit a909ccfa1b4cbf656c4608ef2124347851c492cb)
Reviewed-on: http://review.gluster.org/11762

glusterd: Pass NULL in glusterd_svc_manager in glusterd_restart_bricks

2015-07-24T03:51:38+00:00

On restarting glusterd quota daemon is not started when  more than one
volumes are configured and quota is enabled only on 2nd volume.
This is because of while restarting glusterd it will restart all the bricks.
During brick restart it will start respective daemon by passing volinfo of
first volume. Passing volinfo to glusterd_svc_manager will imply daemon
managers will take action based on the same volume's configuration which
is incorrect for per node daemons.

Fix is to pass volinfo NULL while restarting bricks.

BUG: 1242882
Change-Id: Ie53fc452dc79811068a9397abca13c65de4a8359
Signed-off-by: Gaurav Kumar Garg 
Reviewed-on: http://review.gluster.org/11660
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Atin Mukherjee

cluster/ec: Make background healing optional behavior

2015-07-21T21:37:38+00:00

Provide options to control number of active background heal count and qlen.

 >Change-Id: Idc2419219d881f47e7d2e9bbc1dcdd999b372033
 >BUG: 1237381
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/11473
 >Reviewed-by: Xavier Hernandez 
 >Tested-by: Gluster Build System 

BUG: 1238476
Change-Id: I22ba902d9911195656db9e458c01b54cf0afcd7a
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11680
Tested-by: Gluster Build System 
Reviewed-by: Xavier Hernandez

glusterd: Fix failure in replace-brick when src-brick is offline

2015-07-14T18:39:39+00:00

Change-Id: I0fdb58e15da15c40c3fc9767f2fe4df0ea9d2350
BUG: 1242728
Reviewed-on: http://review.gluster.org/11651
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Krishnan Parthasarathi 
Signed-off-by: Anuradha Talur 
Reviewed-on: http://review.gluster.org/11656

quota/marker: use smaller stacksize in synctask for marker updation

2015-07-14T11:27:24+00:00

This is a backport of http://review.gluster.org/#/c/11499/

Default stacksize that synctask uses is 2M.
For marker we set it to 16k

Also move market xlator close to io-threads
to have smaller stack

> Change-Id: I8730132a6365cc9e242a3564a1e615d94ef2c651
> BUG: 1207735
> Signed-off-by: vmallika 

Change-Id: Id1cb6288a38d370956cc47aed5253ff95f04c966
BUG: 1229282
Signed-off-by: vmallika 
Reviewed-on: http://review.gluster.org/11595
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra G 
Tested-by: Raghavendra G

glusterd: use a real host name (instead of numeric) when we have one

2015-07-13T17:52:54+00:00

Change-Id: Ie9cc201204d3d613e3e585cab066a07283db902c
BUG: 1241963
Signed-off-by: Jeff Darcy 
Reviewed-on: http://review.gluster.org/11587
Reviewed-by: Shyamsundar Ranganathan 
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.org/11624

glusterd: Get the local txn_info based on trans_id in op_sm call backs.

2015-07-13T11:33:31+00:00

Issue: when two or more transactions are running concurrently in op_sm,
global op_info might get corrupted.

Fix: Get local txn_info based on trans_id instead of using global txn_info for
commands (re-balance, profile ) which are using op_sm in originator.

TODO: Handle errors properly in call backs and completely remove the global op_info from
op_sm.
Backport of:
>Change-Id: I9d61388acc125841ddc77e2bd560cb7f17ae0a5a
>BUG: 1229139
>Signed-off-by: anand 
>Reviewed-on: http://review.gluster.org/11120
>Tested-by: Gluster Build System 
>Tested-by: NetBSD Build System 
>Reviewed-by: Krishnan Parthasarathi 
(cherry picked from commit c9765bcb1557ab1e921080e7de4f3ebac1e424d5)

Change-Id: I4405a4a1cf4c853b5d2ce478fd65043e45a79826
BUG: 1230523
Signed-off-by: anand 
Reviewed-on: http://review.gluster.org/11557
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Atin Mukherjee

glusterd: Fix management encryption issues with GlusterD

2015-07-13T10:12:18+00:00

  Backport for 01b82c6 from master

Management encryption was enabled incorrectly in GlusterD leading to
issues of cluster deadlocks. This has been fixed with this commit. The
fix is in two parts,

1. Correctly enable encrytion for the TCP listener in GlusterD and
re-enable own-threads for encrypted connections.
  Without this, GlusterD could try to esatblish the blocking SSL
  connects in the epoll thread, for eg. when handling friend updates,
  which could lead to cluster deadlocks.

2. Explicitly enable encryption for outgoing peer connections.
  Without enabling encryption explicitly for outgoing connections was
  causing SSL socket events to be handled in the epoll thread. Some
  events, like disconnects during peer detach, could lead to connection
  attempts to happen in the epoll thread, leading to deadlocks again.

Change-Id: I438c2b43f7b1965c0e04d95c000144118d36272c
BUG: 1241784
Signed-off-by: Kaushal M 
Reviewed-on: http://review.gluster.org/11610
Tested-by: Gluster Build System 
Tested-by: NetBSD Build System 
Reviewed-by: Atin Mukherjee

glusterd/geo-rep: Fix failure of geo-rep pause

2015-07-11T02:29:07+00:00

Geo-replication pause fails if one or more of the nodes
in the master cluster is not part of master volume.
If the master volume bricks are not part of the node,
it should be ignored. The check is added to fix the issue.

BUG: 1240616
Change-Id: If10da6921d0c87d6e1c1c4ceba975c4b1e3bf302
Reviewed-on: http://review.gluster.org/11549
Tested-by: NetBSD Build System 
Reviewed-by: Avra Sengupta 
Reviewed-by: Krishnan Parthasarathi 
Signed-off-by: Kotresh HR 
Reviewed-on: http://review.gluster.org/11564
Tested-by: Gluster Build System 
Reviewed-by: Milind Changire 
Reviewed-by: Saravanakumar Arumugam