glusterfs.git/xlators/cluster/afr, branch v3.3.0qa30

Logs: Improved logs in lock/unlock execution path

2012-03-18T08:53:41+00:00

Statedump will now start showing the lk-owner of the stack.

Change-Id: I9f650ce9a8b528cd626c8bb595c1bd1050462c86
BUG: 803209
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2968
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

Self-heald: Handle errors gracefully and show errors to users

2012-03-18T07:33:28+00:00

Change-Id: I5424ebfadb5b2773ee6f7370cc2867a555aa48dd
BUG: 800352
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2962
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

cluster/afr: set_read_child when xactions in progress in fresh lookup

2012-03-18T07:30:19+00:00

Change-Id: I33e0268635ae7a1f247b0052994e027f990083da
BUG: 800755
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2963
Tested-by: Gluster Build System 
Reviewed-by: Anand Avati

afr: Copy loc->gfid independent of lookup being fresh or otherwise

2012-03-18T07:01:55+00:00

This change ensures that entry self-heal following a lookup on that entry
would have loc->gfid 'filled'.

Change-Id: If723c71ca43e1f062dcb99cbe5488342514dace0
BUG: 786087
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.com/2950
Tested-by: Gluster Build System 
Reviewed-by: Jeff Darcy 
Reviewed-by: Anand Avati

cluster/afr: Enable eager-lock

2012-03-18T06:31:38+00:00

Eager-lock is disabled by default.
Use cluster.eager-lock on/off to change the config.
write-behind on and eager-lock off is not supported configuration.

In afr, when eager-lock is enabled the inode lock on fd is taken
using the fd address as the lk-owner. So the lock is
interchangableale between the inode-locks on the same fd.

Change-Id: I7eef1ecd510f8028f5395dee882782da53c0de3f
BUG: 802515
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2925
Tested-by: Gluster Build System 
Reviewed-by: Jeff Darcy 
Reviewed-by: Anand Avati

afr: Corrected getxattr 'key' matching in case of clrlk cmd

2012-03-14T17:58:35+00:00

- Added local->dict cleanup into afr_local_cleanup

Change-Id: Ie1b96615735a9d2a2be1757cd016dbe225aae31c
BUG: 800412
Signed-off-by: Krishnan Parthasarathi 
Reviewed-on: http://review.gluster.com/2922
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

cluster/afr: handle sending NULL dentry name for inode link in self-heal-daemon

2012-03-14T10:53:35+00:00

* Without the dentry name, dentry cannot be created in inode_link, which
 leads to trying to access the null dentry to check if it is cyclic and
 thus segfault. So send the parent inode also NULL, which just returns
 the proper inode after assigning the gfid and type to the inode without
 trying to create dentry.

* Handle failures such as dentry_create returning NULL, in inode_link
  properly and return NULL in such cases.

* Increase the lru limit of inode table of self-heal-daemon to 2048

Change-Id: I7ae0e0e9be279d1694b6aafb5e054585e43f03ff
BUG: 801149
Signed-off-by: Raghavendra Bhat 
Reviewed-on: http://review.gluster.com/2893
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Vijay Bellur

cluster/afr: save the xattr obtained in the {f}xattrop_cbk in local

2012-03-12T12:22:56+00:00

If the {f}xattrop operation succeeds on one of the subvolumes and fails
on another (thus the xattr dict obtained from the failed subvolume in
the callback will be NULL), then afr would be unwinding with op_ret = 0
(since the operation was successful on one subvolume), but the xattr dict
would be NULL (afr is not saving the xattr it has received in the callback
in its local structure and will send the xattr it has received in the last
callback). xlators above afr might segfault when they access the xattr since
they would have assumed that xattr would be present as op_ret is 0.

Change-Id: I50761a302150285f31dfdaa397f890c9370a989a
BUG: 797119
Signed-off-by: Raghavendra Bhat 
Reviewed-on: http://review.gluster.com/2813
Tested-by: Gluster Build System 
Reviewed-by: Amar Tumballi 
Reviewed-by: Vijay Bellur

cluster/afr: handle node failures in lookup

2012-03-06T06:55:15+00:00

When a transaction is in progress lookup depends on inode ctx
for read-child. If the lookup fails on the read-child while
another transaction is in progress, it should select the
read-child as the next success_child which is in fresh_children.

Change-Id: I33a04b102966b63a64bacf8d2e29f0d0119fdac6
BUG: 773225
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2858
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

cluster/afr: Reset re-usable sh args in sh_*_done

2012-03-05T18:36:01+00:00

The bug is observed due to stale value of active_sink count
set in metadata self-heal.

Change-Id: I41996999213c323c0f4d4db575d87b2d0b4b3fec
BUG: 798874
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.com/2849
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur