glusterfs.git/xlators/cluster/afr/src, branch bug-deprecated

libglusterfs: change signature of syncop_(f)getxattr

2015-01-06T06:00:09+00:00

Pass xdata dict to syncop_(f)getxattr calls.

This patch [1/3] is required as a part of afr automated split-brain resolution
implementation.

Change-Id: I3970b3dd6daf64681a031e37f8e9afb14fb3d668
BUG: 1136769
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/9375
Reviewed-by: Pranith Kumar Karampuri 
Reviewed-by: Niels de Vos 
Tested-by: Gluster Build System 
Reviewed-by: Shyamsundar Ranganathan 
Reviewed-by: Vijay Bellur

cluster/afr: serialize inode locks

2015-01-05T03:51:43+00:00

Problem:
Afr winds inodelk calls without any order, so blocking inodelks
from two different mounts can lead to dead lock when mount1 gets
the lock on brick-1 and blocked on brick-2 where as mount2 gets
lock on brick-2 and blocked on brick-1

Fix:
Serialize the inodelks whether they are blocking inodelks or
non-blocking inodelks.

        Non-blocking locks also need to be serialized.
Otherwise there is a chance that both the mounts which issued same
non-blocking inodelk may endup not acquiring the lock on any-brick.
Ex:
Mount1 and Mount2 request for full length lock on file f1.  Mount1 afr may
acquire the partial lock on brick-1 and may not acquire the lock on brick-2
because Mount2 already got the lock on brick-2, vice versa. Since both the
mounts only got partial locks, afr treats them as failure in gaining the locks
and unwinds with EAGAIN errno.

Change-Id: Ie6cc3d564638ab3aad586f9a4064d81e42d52aef
BUG: 1176008
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/9372
Reviewed-by: Krutika Dhananjay 
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra G

afr: Fixes to commit 85427a23c238499137cbfaafdb7b6ad27f67506a

2015-01-02T05:55:04+00:00

        * Fixed a dict leak
        * Re-added 'return on failure' check

Change-Id: I07edd03e4608fd2b7c4a91019a0e43033e6e78b2
BUG: 1163804
Signed-off-by: Anuradha 
Reviewed-on: http://review.gluster.org/9368
Tested-by: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri

cluster/afr: Make entry-self-heal in afr-v2 compatible with afr-v1

2014-12-27T07:20:54+00:00

Problem:
entry self-heal in 3.6 and above, takes full lock on the directory only for the
duration of figuring out the xattrs of the directories where as 3.5 takes locks
through out the entry-self-heal. If the cluster is heterogeneous then there is
a chance that 3.6 self-heal is triggered and then 3.5 self-heal will also
triggered and both the self-heal daemons of 3.5 and 3.6 do self-heal.

Fix:
In 3.6.x and above get an entry lock on a very long name before entry self-heal
begins so that 3.5 entry self-heal will not get locks until 3.6.x entry
self-heal completes.

Change-Id: I71b6958dfe33056ed0a5a237e64e8506c3b0fccc
BUG: 1168189
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/9227
Reviewed-by: Krutika Dhananjay 
Tested-by: Gluster Build System

afr: stop encoding subvolume id in readdir d_off

2014-12-26T14:21:44+00:00

The purpose of encoding d_off in AFR is to indicate the
selected subvolume for the first readdir, and continue all
further readdirs of the session on the same subvolume. This is
required because, unlike files, dir d_offs are specific to the
backend and cannot be re-used on another subvolume. The d_off
transformation encodes the subvolume id and prevents such
invalid use of d_offs on other servers.

However, this approach could be quite wasteful of precious d_off
bit-space. Unlike DHT, where server id can change from entry to
entry and thus encoding the server id in the transformed d_off
is necessary, we could take a slightly relaxed approach in AFR.
The approach is to save the subvolume where the last readdir
request was sent in the fd_ctx. This consumes constant space (i.e
no per-entry cache), and serves the purpose of avoiding d_off
"misuse" (i.e using d_off from one server on another).

The compromise here is NFS resuming readdir from a non-0 cookie
after an extended delay (either anonymous FD has been reclaimed,
or server has restarted). In such cases a subvolume is picked
freshly. To make this fresh picking more deterministic (i.e, to
pick the same subvolume whenever possible, even after reboots),
the function afr_hash_child (used by afr_read_subvol_select_by_policy)
is modified to skip all dynamic inputs (i.e PID) for the case
of directories.

Change-Id: I46ad95feaeb21fb811b7e8d772866a646330c9d8
BUG: 1163161
Signed-off-by: Anand Avati 
Reviewed-on: http://review.gluster.org/9332
Tested-by: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri

afr: coverity fixes

2014-12-23T15:22:38+00:00

Some fixes for the 17th Dec 2014 run.
https://scan6.coverity.com:8443/reports.htm#v31028/p10714/g31029

Change-Id: Ia4410ef87a56fffb61803d0a4e62369b058e1cfb
BUG: 1176089
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/9314
Tested-by: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri

cluster/afr : Change in volume heal info command

2014-12-23T15:09:23+00:00

	gluster volume heal  info command
will now also display if the files listed (in the output
of the command) are in split-brain or possibly being
healed.

This patch also fixes build warning that occurs.
Change-Id: I1fc92e62137f23b2b9ddf6e05819cee6230741d1
BUG: 1163804
Signed-off-by: Anuradha 
Reviewed-on: http://review.gluster.org/9119
Tested-by: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri

Avoid spurious directory metedata split brain

2014-12-23T07:29:54+00:00

When directory content is modified, [mc]time is updated. On
Linux, the filesystem does it, while at least on NetBSD, the
kernel file-system independant code does it. This means that
when entries are added while bricks are down, the kernel sends
a SETATTR [mc]time which will cause metadata split brain for
the directory. In this case, clear the split brain by finding
the source with the most recent modification date.

BUG: 1129939
Change-Id: Ic0177e0df753a4748624d0b906834ed54593adb9
Signed-off-by: Emmanuel Dreyfus 
Reviewed-on: http://review.gluster.org/9291
Reviewed-by: Ravishankar N 
Reviewed-by: Krutika Dhananjay 
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri

telldir()/seekdir() portability fixes

2014-12-17T11:48:22+00:00

POSIX says that an offset obtained from telldir() can only be used
on the same DIR *. Linux is abls to reuse the offset accross
closedir()/opendir() for a given directory, but this is not portable
and such a behavior should be fixed.

An incomplete fix for the posix xlator was merged in
http://review.gluster.com/8926
This change set completes it.

- Perform the same fix index xlator.
- Use appropriate casts and variable types so that 32 bit signed
  offsets obtained by telldir() do not get clobbered when copied into
  64 bit signed types.
- modify glfs-heal.c and afr-self-heald.c so that they do not use
  anonymous fd, since this will cause closedir()/opendir() between
  each syncop_readdir(). On failure we fallback to anonymous fs
  only for Linux so that we can cope with updated client vs not
  updated brick.
- Avoid sending an EINVAL when the client request for the EOF offset.
  Here we fix an error in previous fix for posix xlator: since we
  fill each directory entry with the offset of the next entry, we
  must consider as EOF the offset of the last entry, and not the
  value of telldir() after we read it.
- Add checks in regression tests that we do not hit cases where
  offsets fed to seekdir() are wrong. Introduce log_newer() shell
  function to check for messages produced by the current script.

This fix gather changes from http://review.gluster.org/9047
and http://review.gluster.org/8936 making them obsolete.

BUG: 1129939
Change-Id: I59fb7f06a872c4f98987105792d648141c258c6a
Signed-off-by: Emmanuel Dreyfus 
Reviewed-on: http://review.gluster.org/9071
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra Bhat 
Tested-by: Raghavendra Bhat

cluster/afr: Eliminate locking in sh domain in metadata self-heal

2014-12-05T09:51:14+00:00

Change-Id: I9ef25a17c9a43ba06fac2ad3f7c18cb47de91537
BUG: 1170913
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/9240
Tested-by: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: Pranith Kumar Karampuri