glusterfs.git/tests/basic, branch v3.7.4

ec : trusted.ec.version xattr of all root directories of all bricks should be same.

2015-08-30T05:54:44+00:00

Problem:
After replacing the brick using "replace-brick" command and running "heal
full", the version of the root directory of the newly added brick is not
getting healed. heal starts running on the dentries of the root but does not
run on root directory.

Solution:
Run heal on root directory.

> Change-Id: Ifd42a3fb341b049c895817e892e5b484a5aa6f80
> BUG: 1243382
> Signed-off-by: Ashish Pandey 
> Reviewed-on: http://review.gluster.org/11676
> Reviewed-by: Pranith Kumar Karampuri 
> Tested-by: NetBSD Build System 

Change-Id: Ifd42a3fb341b049c895817e892e5b484a5aa6f80
BUG: 1243384
Signed-off-by: Ashish Pandey 
Reviewed-on: http://review.gluster.org/11755
Tested-by: Gluster Build System 
Tested-by: NetBSD Build System 
Reviewed-by: Pranith Kumar Karampuri

tests: remove unwanted tests from volume-snapshot.t

2015-08-26T08:53:28+00:00

    Backport of http://review.gluster.org/#/c/11972/

volume-snapshot.t failspuriously because of having additional test cases
which restarts glusterd and they are really not needed as far as the test
coverage is concerned. Currently glusterd doesn't have a mechanism to indicate
that volumes handshaking has been completed or not, due to this even if the peer
handshaking finishes and all the peers are back to the cluster there could be a
case where any command which accesses the volume structure might end up in
corruption as volume handshaking is still in progress. This is because of volume
list is still not been made URCU protected.

Change-Id: Id8669c22584384f988be5e0a5a0deca7708a277d
BUG: 1255636
Signed-off-by: Atin Mukherjee 
Reviewed-on: http://review.gluster.org/11975
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Avra Sengupta 
Reviewed-by: Raghavendra Talur

libgfapi: adding 'follow' flag to glfs_h_lookupat()

2015-08-25T12:31:35+00:00

This patch is backport of http://review.gluster.org/#/c/11883/

Previously glfs_h_lookupat won't follow the symlink, this patch
introduces new flag `follow` which will resolve the same. Applications
linking against the new library will need to use the new glfs_h_lookupat
API call.

In order to stay compatible with existing binaries that use the previous
glfs_h_lookupat() function, the old symbol needs to stay available.

Verification that there are two versions of glfs_h_lookupat:

  $ objdump -T /usr/lib64/libgfapi.so.0 | grep -w glfs_h_lookupat
  0000000000015070 g    DF .text  000000000000021e  GFAPI_3.7.4 glfs_h_lookupat
  0000000000015290 g    DF .text  0000000000000008 (GFAPI_3.4.2) glfs_h_lookupat

Testing with a binary (based on anonymous_fd_read_write.c from ./tests/)
that was linked against the old library:

  $ objdump -T ./lookupat | grep -w glfs_h_lookupat
  0000000000000000      DF *UND*  0000000000000000  GFAPI_3.4.2 glfs_h_lookupat

Enable debugging for 'ld.so' so that we can check that the GFAPI_3.4.2
version of the symbol gets loaded:

  $ export LD_DEBUG_OUTPUT=lookupat.ld.log LD_DEBUG=all
  $ ./lookupat
  $ grep -w glfs_h_lookupat lookupat.ld.log.2543
  2543:     symbol=glfs_h_lookupat;  lookup in file=./lookupat [0]
  2543:     symbol=glfs_h_lookupat;  lookup in file=/lib64/libgfapi.so.0 [0]
  2543:     binding file ./lookupat [0] to /lib64/libgfapi.so.0 [0]: normal symbol `glfs_h_lookupat' [GFAPI_3.4.2]

This change has been successfully cherry-picked as 1ead86a8bcbfe4045729466e4b98f765f3c13c8d
in master

Upstream reference
>Change-Id: I8bf9b1c19a0585f681bc1a7f84aad1ccd0f75f6a
>BUG: 1252410
>Signed-off-by: Jiffin Tony Thottan 
>Signed-off-by: Niels de Vos 
>Reviewed-on: http://review.gluster.org/11883
>Reviewed-by: soumya k 
>Reviewed-by: Kaleb KEITHLEY 
>Tested-by: NetBSD Build System 
Signed-off-by: Jiffin Tony Thottan 

Change-Id: Idbbf0cd6802f86c53b16377d90d08ff6d99e7b08
BUG: 1256616
Reviewed-on: http://review.gluster.org/12009
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Kaleb KEITHLEY

glusterd/tier: configure tier daemon during volume restart

2015-08-19T08:11:03+00:00

rebalance daemon will be running on every tier volume for
promoting/demoting the files. When volume/glusterd is restarted,
then we need to configure the daemon.

Back port of >
>Change-Id: Ib565240a70edea2ec8bc1601c52b40c0783491d3
>BUG: 1225330
>Signed-off-by: Mohammed Rafi KC 
>Signed-off-by: Joseph Fernandes 
>Reviewed-on: http://review.gluster.org/10933
>Reviewed-by: Dan Lambright 
>Tested-by: Dan Lambright 

(cherry picked from commit 7aaa414c67be460aca3c1c5eed047e3db17bd435)

Change-Id: Id15f5a18fd2ebaf8642b02561ae3666e619e3f83
BUG: 1235202
Signed-off-by: Mohammed Rafi KC 
Reviewed-on: http://review.gluster.org/11376
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Dan Lambright 
Tested-by: Dan Lambright

tests: add basic/afr/arbiter-statfs.t

2015-08-14T04:00:02+00:00

Backport of http://review.gluster.org/#/c/11864/

This is a test-case for BZ 1251380

Change-Id: Icbde8d17c823a3f2c98056c14a75f0ef5227b848
BUG: 1251380
Signed-off-by: Ravishankar N 
(cherry picked from commit f550efa2879c02dec4c8f6876a94fcfd95b5bf91)
Reviewed-on: http://review.gluster.org/11900
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Raghavendra Talur

glusterd: glusterd crash due to race between handshake and snapshot remove threads

2015-07-30T03:50:37+00:00

Issue : glusterd was crashing due to race between handshake thread and snapshot
remove
RCA : Snapshot  thread referring  voinfo and same time volinfo is modified during handshake,
glusterd was crashing  due to this inconsistent data of volinfo .

Note: Sending commands without checking cluster status may lead to crash

Fix:.Wait for handshake complete/cluster ready before proceeding commands.
Back port of :
>Change-Id: Iefd986664bd9dd225f0abf8f85476d6afd206914
>BUG: 1246432
>Signed-off-by: anand 
>Reviewed-on: http://review.gluster.org/11757
>Tested-by: Gluster Build System 
>Reviewed-by: Atin Mukherjee 
>Tested-by: NetBSD Build System 
>Reviewed-by: Raghavendra Talur 
>(cherry picked from commit 51f48bc9a41a5e2004d9051ff90517b01626b08f)

Change-Id: I4ca707f84544ba9e4e3ea77752c6576af8572f0c
BUG: 1247917
Reviewed-on: http://review.gluster.org/11787
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

features/quota : Fix spurious failure

2015-07-14T04:02:21+00:00

This is a backport of http://review.gluster.org/#/c/11125/

Problem : Basically, in this test case a file is created
which exceeds the quota limit. Once the limit is reached
that file will be deleted. At the same moment we are
testing inode-quota. It can so happen that before the
marker updates the information related to deletion of
file, a new file creation operation comes and sees that
quota limit is still exceeded.

Solution : Inducing a check to see if marker updation
completed successfully.

Updated all the test case which has the similar
machanism and also moved the "usage" function
to a common place "volume.rc"

> Change-Id: I36ddbc5ebbf1b74c9d326a0d1d5f3b32f20a906a
> BUG: 1229297
> Signed-off-by: Sachin Pandit 
> Signed-off-by: vmallika 
> Reviewed-on: http://review.gluster.org/11125
> Tested-by: NetBSD Build System 
> Tested-by: Gluster Build System 
> Reviewed-by: Krishnan Parthasarathi 

Change-Id: Iccc36de2b3a1e1a068d1a8d5e98d413c3afa1bc7
BUG: 1242329
Signed-off-by: vmallika 
Reviewed-on: http://review.gluster.org/11642
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Krishnan Parthasarathi

cluster/ec: wind fops on good subvols for access/readdir[p]

2015-07-02T06:42:17+00:00

        Backport of http://review.gluster.com/11246

BUG: 1234679
Change-Id: I2d774f62740c82e922efab50fc78fa74050ede93
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/11357
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System

cluster/afr : set pending xattrs for replaced brick

2015-06-27T13:08:54+00:00

       Backport of: http://review.gluster.org/10448/
                  & http://review.gluster.org/11416

This patch is part two change to prevent data loss
in a replicate volume on doing a replace-brick commit
force operation.

Problem: After doing replace-brick commit force, there is a
chance that self heal might happen from the replaced (sink) brick
rather than the source brick leading to data loss.

Solution: Mark pending changelogs on afr children for
the replaced afr-child so that heal is performed in the
correct direction.

Credits to Ravishankar N for patch 11416.

Change-Id: Icb9807e49b4c1c4f1dcab115318d9a58ccf95675
BUG: 1232173
Reviewed-on: http://review.gluster.org/10448
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: NetBSD Build System 
Reviewed-by: Krutika Dhananjay 
Signed-off-by: Anuradha Talur 
Reviewed-on: http://review.gluster.org/11254
Tested-by: Gluster Build System