glusterfs.git/tests/bugs, branch v3.8.0

glusterd: fail volume delete if one of the node is down

2016-06-13T14:39:46+00:00

Backport of http://review.gluster.org/14681

Deleting a volume on a cluster where one of the node in the cluster is down is
buggy since once that node comes back the resync of the same volume will happen.
Till we bring in the soft delete feature tracked in
http://review.gluster.org/12963 this is a safe guard to block the volume
deletion.

Please note the test file which is backported from this commit has an issue
where we start the volume and then try to delete it which is anyway going to
fail. So the test actually doesn't validate the fix.
http://review.gluster.org/#/c/14693/ in master fixed the problem and the same is
ported as part of this commit as well.

Cherry picked from commit 5016cc548d4368b1c180459d6fa8ae012bb21d6e:
> Change-Id: I9c13869c4a7e7a947f88842c6dc6f231c0eeda6c
> BUG: 1344407
> Signed-off-by: Atin Mukherjee 
> Reviewed-on: http://review.gluster.org/14681
> Smoke: Gluster Build System 
> CentOS-regression: Gluster Build System 
> Reviewed-by: Kaushal M 
> NetBSD-regression: NetBSD Build System 

Change-Id: I9c13869c4a7e7a947f88842c6dc6f231c0eeda6c
BUG: 1344631
Signed-off-by: Atin Mukherjee 
Reviewed-on: http://review.gluster.org/14691
Reviewed-by: Niels de Vos 
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

afr: Consider ENOSPC and EDQUOT as symmetric errors

2016-06-13T10:14:49+00:00

Backport of http://review.gluster.org/#/c/14604/

Problem:
Since commit 8eaa3506ead4f11b81b146a9e56575c79f3aad7b, in replica 3, if a
brick is down and a create fails on the other 2 brick with EDQUOT, we consider
it an unsymmetric error and hence do not do post-op. So the dirty xattr
remains set on the parent dir, leading to conservative merges during heal when
all bricks are up. i.e. a file deleted on the source might re-appear after heal.

Fix:
Consider ENOSPC and  EDQUOT as symmetric errors since there is no
possibility of partial inode or entry modification operations possible when
quota is enabled. IOW, if quota reports EDQUOT, the no. of bytes written
(or not written) will be the same on all bricks of the replica.
Likewise, the entry operation (create, mkdir...) will either succeed or
not succeed on all bricks.

Change-Id: Iacb1108e9ef4a918e36242fb4a957455133744e9
BUG: 1344559
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/14687
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
Reviewed-by: Niels de Vos

nfs: build exportlist with multiple groupnodes

2016-06-13T10:11:00+00:00

The EXPORT procedure of the MOUNT protocol does not correctly create
structures for the 'groupnodes' in the reply. Each 'groupnode' should be
a single entry in the 'nfs.rpc-auth-allow' volume option. Because the
value is handled as a single string, the encoding of the
groupnode->gr_name fails when the value of the volume option is longer
than 255 characters.

In the error case, encoding the EXPORTS reply fails, and the waiting
'showmount' command will not receive a reply and times out.

Splitting the allowed entries and creating a groupnode for each one
prevents the too long ->gr_name. This is following the structures for
the EXPORTS reply in the MOUNT protocol more correctly as well. Note
that the contents of ->gr_name is expected to be server dependent.

This is backport of below mainline fix -
	http://review.gluster.org/#/c/14667/

Change-Id: Ibbabad581cc9aa00feb80fbbc851a1b10b28383d
BUG: 1343287
Signed-off-by: Niels de Vos 
Signed-off-by: Bipin Kunal 
Reviewed-on: http://review.gluster.org/14700
Smoke: Gluster Build System 
Reviewed-by: bipin kunal 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System

glusterfsd/main: Add ability to set oom_score_adj

2016-06-01T18:14:43+00:00

Give the administrator a possibility to set oom_score_adj for glusterfs
process. Applies to Linux only.

This is a backport of cb8f5e01f639cb6e8715b33bb725210cb0493887.

Change-Id: Iff13c2f4cb28457871c6ebeff6130bce4a8bf543
BUG: 1341697
Signed-off-by: Oleksandr Natalenko 
Reviewed-on: http://review.gluster.org/14399
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Jeff Darcy 
Reviewed-on: http://review.gluster.org/14605
Reviewed-by: Kaleb KEITHLEY 
Reviewed-by: Vijay Bellur

cluster/afr: Do not inode_link in afr

2016-05-25T10:31:27+00:00

Race is explained at
https://bugzilla.redhat.com/show_bug.cgi?id=1337405#c0

This patch also handles performing of self-heal with shd-pid.
Also performs the healing with this->itable's inode rather than
main itable.

 >BUG: 1337405
 >Change-Id: Id657a6623b71998b027b1dff6af5bbdf8cab09c9
 >Signed-off-by: Pranith Kumar K 
 >Reviewed-on: http://review.gluster.org/14422
 >Smoke: Gluster Build System 
 >NetBSD-regression: NetBSD Build System 
 >CentOS-regression: Gluster Build System 
 >Reviewed-by: Krutika Dhananjay 

BUG: 1337870
Change-Id: Ifb476eeed2ff73a44e481d64074599ab0707c725
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/14455
Smoke: Gluster Build System 
Reviewed-by: Krutika Dhananjay 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Niels de Vos

tests/write-behind: move 1279730.t to BAD tests

2016-05-24T13:26:16+00:00

There is a race condition which is causing the test to fail. For lack
of bandwidth I am moving this test to BAD, though clearly there is
some issue with codebase.

Backport of:
> BUG: 1337777
> Change-Id: If4f3eff8a5985f37a4dee65d2df29fa7b6bda7ae
> Reviewed-on: http://review.gluster.org/14443
> Signed-off-by: Raghavendra G 

BUG: 1337780
Change-Id: If4f3eff8a5985f37a4dee65d2df29fa7b6bda7ae
Signed-off-by: Raghavendra G 
Reviewed-on: http://review.gluster.org/14445
CentOS-regression: Gluster Build System 
NetBSD-regression: NetBSD Build System 
Smoke: Gluster Build System 
Reviewed-by: Vijay Bellur

features/shard: Get hard-link-count in {unlink,rename}_cbk before deleting shards

2016-05-24T08:57:46+00:00

        Backport of http://review.gluster.org/#/c/14334/

Change-Id: Iff0e90bee22e20c309eaea6c6a19e4fa6e101ed7
BUG: 1337839
Signed-off-by: Krutika Dhananjay 
Reviewed-on: http://review.gluster.org/14451
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Niels de Vos

cluster/afr : Do post-op in case of symmetric errors

2016-05-24T08:37:10+00:00

        Backport of: http://review.gluster.org/#/c/14310/

In afr_changelog_post_op_now(), if there was any error,
meaning op_ret < 0, post-op was not being done even when
the errors were symmetric and there were no "failed
subvols".

Fix:
When the errors are symmetric, perform post-op.

How was the bug found :
In a 1 X 3 volume with shard and write behind on
when writes were done into a file with one brick down,
the trusted.afr.dirty xattr's value for .shard directory
would keep increasing as post op was not done but pre-op was.
This incorrectly showed .shard to be in split-brain.

RCA:
When WB is on, due to multiple writes being sent on
offset lying in the same shard, chances are that
same shard file will be created more than once
with the second one failing with op_ret < 0
and op_errno = EEXIST.

As op_ret was negative, afr wouldn't do post-op,
leading to no decrement of trusted.afr.dirty xattr.
Thus showing .shard directory to be in split-brain.

        >Change-Id: I711bdeaa1397244e6a7790e96f0c84501798fc59
        >BUG: 1335652
        >Signed-off-by: Anuradha Talur 

Change-Id: I711bdeaa1397244e6a7790e96f0c84501798fc59
BUG: 1335829
Signed-off-by: Anuradha Talur 
Reviewed-on: http://review.gluster.org/14331
NetBSD-regression: NetBSD Build System 
Smoke: Gluster Build System 
Reviewed-by: Ravishankar N 
CentOS-regression: Gluster Build System 
Reviewed-by: Niels de Vos

nfs: strip trailing / when clients do subdir mounts

2016-05-24T08:35:09+00:00

Mounting a volume over NFS with a subdir followed by a / does not work:

  # mount -t nfs -o vers=3 storage.example.com:/media/installation/ /mnt
  mount.nfs: an incorrect mount option was specified

In the nfs.log:

  [client-rpc-fops.c:2930:client3_3_lookup_cbk] 0-media-client-0: remote operation failed. Path: /installation/ (00000000-0000-0000-0000-000000000000) [Invalid argument]
  [client-rpc-fops.c:2930:client3_3_lookup_cbk] 0-media-client-1: remote operation failed. Path: /installation/ (00000000-0000-0000-0000-000000000000) [Invalid argument]
  [mount3.c:1134:mnt3_resolve_subdir_cbk] 0-nfs: path=/installation/ (Invalid argument) [Invalid argument]

It is not possible to resolve paths with a trailing /. Stripping
trailing /'s from the subdir to mount is sufficient to make it work
again.

Backport:
> Change-Id: I4075d4cd351438de58e1ff81f0fb65a1ff076da4
> BUG: 1337597
> Reviewed-on: http://review.gluster.org/14421
> Signed-off-by: Niels de Vos 

Change-Id: I4075d4cd351438de58e1ff81f0fb65a1ff076da4
BUG: 1337596
Signed-off-by: Niels de Vos 
Reviewed-on: http://review.gluster.org/14424
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Smoke: Gluster Build System 
Reviewed-by: jiffin tony Thottan

Revert glusterd/afr: store afr pending xattrs as a volume option

2016-05-19T04:25:32+00:00

This patch reverts changes introduced by commit
6e635284a4411b816d4d860a28262c9e6dc4bd6a in the release-3.8 branch. The
commit itself was inherited from http://review.gluster.org/#/c/12738/ in
master when the branch was created.

Reverting it for the same reason it was reverted in 3.7 branch as well:
It breaks the rolling upgrade scenario and these changes will be
required only when server side AFR materializes, possibly for gluster
4.0.

Change-Id: Ib3bb78994d7375f7c34df9897dfaf653ea909924
BUG: 1337130
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/14414
Smoke: Gluster Build System 
NetBSD-regression: NetBSD Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Vijay Bellur