glusterfs.git/xlators/features/gfid-access/src, branch release-3.6

feature/gfid-access: Always send setattr down in overloaded setxattr.

2014-11-14T09:53:14+00:00

Problem: File ownership is not being preserved for root in geo-rep
         mountbroker setup.

Analysis and Cause:
         Entry creations for geo-rep is overloaded in ga_setxattr.
         It happens in two phase, entry creation followed by setattr
         to preserve ownership as in master.

         If uid and gid of file being synced is root, setattr was
         not being sent down. Since, the file creation happens with
         non-root user in mountborker geo-rep setup, if setattr is
         not done explicitly, file ownership is not preserved for root.

Solution:
         Always pass setattr down in overloaded ga_setxattr.

BUG: 1159213
Change-Id: I0a6ef16333190b069e2ae326721d2b983f6a1a44
Reviewed-on: http://review.gluster.org/9051
Signed-off-by: Kotresh HR 
Reviewed-on: http://review.gluster.org/9083
Tested-by: Gluster Build System 
Reviewed-by: Venky Shankar 
Tested-by: Venky Shankar

features/gfid-access: error handling for entry creation

2014-07-14T13:26:55+00:00

Proceed with setattr() only on a successfull entry creation.
Winding a setattr() using a freshlyOC initiated inode would
most likely fail in one translator or the other (e.g. DHT
expecting the layout information to be set in the inode
context), which is the case if the inode was not looked up.

Therefore, gfid-access handles failure entry creations and
passes the _correct_ errno back to the client instead of
continuing with setattr() call and probably returning back
incorrect errno. Also, filling up inode->gfid is required
as the new inode is not looked up and ->gfid would be
certainely required for inode operations.

Change-Id: Ie92f5647a89bf558c07710ab0400bce69d59fc31
BUG: 1111490
Signed-off-by: Venky Shankar 
Reviewed-on: http://review.gluster.org/8260
Tested-by: Gluster Build System 
Reviewed-by: Kotresh HR 
Reviewed-by: Vijay Bellur

features/gfid-access: Fix entry operations

2014-07-07T04:01:20+00:00

Problem:
When more than one aux-mounts are performing rmdir .gfid//dir
simultaneously, then sometimes a hang is observed.  In gfid-access xlator When
virtual parent/inode are replaced with real parent/inode in loc, virtual
pargfid/gfid are not replaced with real pargfid/gfid respectively. Afr is using
parent_loc->gfids to order the entry locks. But parent_loc->gfid contains
random/virtual gfid generated by gfid-access xlator. Entrylk in client xlator
is using loc->inod->gfid for sending entrylk which has 'real' gfid. Because the
ordering is happening based on random gfids, One mount orders the locks as (L1,
L2) where as the other orders them as (L2, L1) leading to a dead-lock thus
a hang.

Fix:
Replace virtual pargfid/gfid with real pargfid/gfid when virtual-inodes are
replaced with real-inodes in loc.

BUG: 1114501
Change-Id: Ie94e816122ef9e7aad51605adbf49291de60827e
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/8204
Reviewed-by: Kotresh HR 
Reviewed-by: Vijay Bellur 
Tested-by: Vijay Bellur

features/gfid-access: Handle loc modification correctly for virtual dirs

2014-06-30T09:55:45+00:00

Change-Id: I6e3321534dc2f711519b18e8bffb691ab952a8ba
BUG: 1112659
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/8163
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra G 
Reviewed-by: Vijay Bellur

feautre/gfid-access: Fix EINVAL when stat on .gfid

2014-06-20T09:21:58+00:00

Problem: Some of the inode operations on '.gfid'
virtual directory was resulting in the error
EINVAL from dht after failing to find the layout.

Solution: Inode operations on '.gfid' virtual directory
should not wind further down and should be handled
accordingly in the gfid-access translator itself.

Change-Id: I156cb10ffea0c46b0d747e26f74538d7fb01a1dd
BUG: 1105891
Signed-off-by: Kotresh H R 
Reviewed-on: http://review.gluster.org/8011
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra G 
Reviewed-by: Venky Shankar 
Reviewed-by: Vijay Bellur

features/gfid-access: Fix inode leaks and loc path corruption

2014-06-10T15:01:13+00:00

Change-Id: Ib506db28415d57b344aa90e07e74666ad2063c64
BUG: 1104919
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/8009
Tested-by: Gluster Build System 
Reviewed-by: Kotresh HR 
Reviewed-by: Vijay Bellur

features/gfid-access: Fix memory leaks.

2014-06-10T09:06:41+00:00

Change-Id: I90f6cdb1c8c4face1bb72a9cc77818d308389e45
BUG: 1104919
Signed-off-by: Raghavendra G 
Reviewed-on: http://review.gluster.org/7982
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Gluster Build System 
Reviewed-by: Venky Shankar 
Reviewed-by: Vijay Bellur

features/gfid-access: calloc gfid and set in xdata

2014-06-08T14:53:09+00:00

Problem: the gfid passed to ga_fill_tmp_loc() was a stack variable which
the function set in the xdata dictionary. Accessing it in a later point
in time gave unexpected values.

This was easy to hit when AFR was involved like so:
ga_mknod()--->xxx-->afr_mknod():
In afr_mknod transaction, once the stack-winds for the lock-phase are sent,
the gfid in xdata becomes out of scope. When we send the actual op i.e.
afr_mknod_wind(), the gfid in xdata is stale, causing posix to set junk
gfids on the files.

Fix:
calloc the gfid and set it in the dict.

Thanks to Pranith for the RCA!

Change-Id: I0291fce34745268dc63095fc41fe31c7213dd5e1
BUG: 1104707
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/7978
Reviewed-by: Raghavendra G 
Reviewed-by: Pranith Kumar Karampuri 
Reviewed-by: Venky Shankar 
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

features/gfid-access: Fix possible inode memory corruption.

2014-04-03T16:36:42+00:00

        During lookup, the inode is not ref'd. Added code
        to ref the inode in call path and unref in cbk path.
        Also fixed a case where we should always be putting
        linked inode into context as it is not guaranteed
        that we get same inode that we passed in a call to
        inode_link.

Change-Id: Iaec083a9258658bef3047e83956729d3dbcd9a59
BUG: 1080295
Signed-off-by: Kotresh H R 
Reviewed-on: http://review.gluster.org/7329
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra G 
Reviewed-by: Venky Shankar

geo-rep/gfid-access: Fix errno for non-existent GFID.

2014-02-28T05:45:31+00:00

Because of http://review.gluster.org/#/c/6318/ patch,
ESTALE is returned for a lookukp on non-existent GFID.
But ENOENT is more appropriate when lookup happens
through virtual .gfid directory on aux-gfid-mount
point. This is avoids confusion for the consumers
of gfid-access-translator like geo-rep which expects
ENOENT.

Change-Id: I4add2edf5958bb59ce55d02726e6b3e801b101bb
BUG: 1069191
Signed-off-by: Kotresh H R 
Reviewed-on: http://review.gluster.org/7154
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra G 
Reviewed-by: Venky Shankar 
Reviewed-by: Anand Avati