glusterfs.git, branch v3.5.4beta1

cluster/afr: Make read child match check in afr optional

2015-03-18T11:21:28+00:00

        Backport of: http://review.gluster.org/#/c/9917

Having this particular check which was introduced by
commit bb2df4e63fa8a5d65f18b4a5efc757e8d475fbff causes a drop in
performance in readdirp. So the behavior is made configurable with this
patch.

Change-Id: I9012a6bb955229a0cbb48f06e4e2edc0782dfead
BUG: 1202675
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/9924
Reviewed-by: Pranith Kumar Karampuri 
Reviewed-by: Atin Mukherjee 
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos

afr: exit out of stack winds in for loops if call_count is zero

2015-03-12T04:14:38+00:00

....in order to avoid a race where the fop cbk frees the frame's local
variables and the fop tries to access it at a later point in time.

Change-Id: I91d2696e5e183c61ea1368b3a538f9ed7f3851de
BUG: 1200764
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/9856
Tested-by: Gluster Build System 
Reviewed-by: pranith karampuri 
Reviewed-by: Niels de Vos

configure: argp-standalone is always built, even when not needed

2015-02-26T21:17:12+00:00

This patch still configures argp, i.e. creates Makefile in
/contrib/argp-standalone, for the `make dist`, but there's no reason
to compile it when it's not going to be linked into anything.

Backport of commit 09d2dcac3a238cfe6bde1a7aed06facfa16caf20:
> Change-Id: Id0fcb717b5821d8a4c176e6274339f46b3a4a249
> BUG: 1086773
> Signed-off-by: Kaleb S. KEITHLEY 
> Reviewed-on: http://review.gluster.org/7440
> Tested-by: Gluster Build System 
> Reviewed-by: Harshavardhana 
> Reviewed-by: Niels de Vos 
> Reviewed-by: Anand Avati 

Change-Id: Id0fcb717b5821d8a4c176e6274339f46b3a4a249
BUG: 1191006
Signed-off-by: Niels de Vos 
Reviewed-on: http://review.gluster.org/9761
Tested-by: Gluster Build System 
Reviewed-by: Kaleb KEITHLEY

protocol: Log ENODATA & ENOATTR logs at DEBUG loglevel in removexattr_cbk.

2015-02-25T08:55:24+00:00

Prevents messages of the following type from being seen by default in the log files:

[2014-09-19 07:57:39.877806] W
[client-rpc-fops.c:1232:client3_3_removexattr_cbk] 0--client-0:
remote operation failed: No data available
[2014-09-19 07:57:39.877963] W
[client-rpc-fops.c:1232:client3_3_removexattr_cbk] 0--client-1:
remote operation failed: No data available

Change-Id: I3b1a121b0fc272eb772547275bb8085ed19db5a1
BUG: 1192832
Signed-off-by: Vijay Bellur 
Reviewed-on: http://review.gluster.org/8781
Reviewed-by: Niels de Vos 
Tested-by: Gluster Build System 
Reviewed-by: Jeff Darcy 
Reviewed-on: http://review.gluster.org/9655

features/quota: Send the immediate parent with limit in quota statfs

2015-02-23T12:56:47+00:00

adjustment

This is a backport of http://review.gluster.org/#/c/7330

> Problem:
> Assume the directory structure /quota_limit_dir/subdir and
> quota_limit_dir is
> set with some limit. When quota-deem-statfs is enabled the output of
> 'df /quota_limit_dir' would display quota modified values wrt to
> quota_limit_dir where as 'df /quota_limit_subdir/subdir' would display
> the quota modified values wrt volume root (/).
>
> This behaviour is not expected since, when mounted with subdirectory
> admin doesn't want users to know information above the subdirectory mounted.
>
> Solution:
> Any subdirectory within a quota_limit_dir would show the modified values
> as in the /quota_limit_dir. It searches for the nearest parent that has quota
> limit set and modifies the statvfs wrt that.
>
> Change-Id: Ie10fae8999bddbb766b1dbeb881723ed80dce688
> BUG: 1080296
> Signed-off-by: Varun Shastry 
> Reviewed-on: http://review.gluster.org/7330
> Reviewed-by: Raghavendra G 
> Tested-by: Raghavendra G 

Change-Id: I5232a82da20d3dfd48303ba2748b86cfc96ff864
BUG: 1115197
Signed-off-by: vmallika 
Reviewed-on: http://review.gluster.org/9510
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos

afr: Don't write to sparse regions of sink.

2015-02-12T12:09:13+00:00

Corresponding afr-v2 fix: http://review.gluster.org/#/c/9480/

Problem:
When data-self-heal-algorithm is set to 'full', shd just reads from
source and writes to sink. If source file happened to be sparse (VM
workloads), we end up actually writing 0s to the corresponding regions
of the sink causing it to lose its sparseness.

Fix:
If the source file is sparse, and the data read from source and sink are
both zeros for that range, skip writing that range to the sink.

Change-Id: Iade957e4173c87e45a2881df501ba2ad3eb1a172
BUG: 1190633
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/9611
Reviewed-by: Pranith Kumar Karampuri 
Reviewed-by: Krutika Dhananjay 
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos

features/marker: Filter internal xattrs in lookup

2015-02-12T12:08:12+00:00

        Backport of http://review.gluster.com/9061

Afr should ignore quota-size-key as part of self-heal
but should heal quota-limit key.

BUG: 1162230
Change-Id: I639cfabbc44468da29914096afc7e2eca1ff1292
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/9091
Tested-by: Gluster Build System 
Reviewed-by: Ravishankar N 
Reviewed-by: Raghavendra G 
Reviewed-by: Krutika Dhananjay 
Reviewed-by: Niels de Vos

cluster/afr: serialize inode locks

2015-02-11T10:15:30+00:00

      Backport of http://review.gluster.com/9372

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: I939a1d101e313a9f0abf212b94cdce1392611a5e
BUG: 1177928
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/9374
Reviewed-by: Krutika Dhananjay 
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos

features/locks: Add lk-owner checks in entrylk

2015-02-11T10:14:21+00:00

        Backport of http://review.gluster.com/9125

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 3.6.2 and above:
3.6.2 and above gets 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. Make 3.5 locks accept nested locks from same lk-owner,
client.

BUG: 1177339
Change-Id: I65d4c3732d1e074720422533f85f6e13f933ecad
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/9352
Reviewed-by: Krutika Dhananjay 
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos

cluster/afr: Preserve errno in case of failures on all subvols

2015-02-11T10:13:46+00:00

        Partly backported from http://review.gluster.org/8984

Problem:
When quorum is enabled and the fop fails on all the subvolumes,
op_errno is set to EROFS which overrides the actual errno returned
from bricks.

Fix:
Don't override the errno when fop fails on all subvols.

PS: Afr-v2 code differs from afr-v1 so that pre-op part of code doesn't apply

Change-Id: I61e57bbf1a69407230ec172a983de18d1c624fd2
BUG: 1162150
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/9088
Tested-by: Gluster Build System 
Reviewed-by: Niels de Vos