glusterfs.git/tests/bugs/glusterfs, branch v3.11dev

afr: afr-pending-xattr fallback check

2016-06-09T08:20:43+00:00

Commit 6e635284a4411b816d4d860a28262c9e6dc4bd6a introduced a comma
separated list of values to be used as AFR's pending changelogs. If
this xlator option is missing in the volfile, fall back to using client
xlator names for constructing the pending changelog names.

Also, since the aforementioned commit was reverted from 3.7 and 3.8
branches, introduce GD_OP_VERSION_3_9_0 and change the op-version for this
feature to GD_OP_VERSION_3_9_0.

Change-Id: I3639b9ab475bd8d9929cc7527d9f4584dee1ad1b
BUG: 1285152
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/14642
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Pranith Kumar Karampuri 
NetBSD-regression: NetBSD Build System

glusterd: default value of nfs.disable, change from false to true

2016-04-27T07:59:30+00:00

Next step in eventual deprecation of glusterfs nfs server in favor
of ganesha.nfsd.

Also replace several open-coded strings with constant.

Change-Id: If52f5e880191a14fd38e69b70a32b0300dd93a50
BUG: 1092414
Signed-off-by: Kaleb S KEITHLEY 
Reviewed-on: http://review.gluster.org/13738
NetBSD-regression: NetBSD Build System 
Smoke: Gluster Build System 
CentOS-regression: Gluster Build System 
Reviewed-by: Atin Mukherjee 
Tested-by: Atin Mukherjee 
Reviewed-by: Niels de Vos

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

2015-12-16T06:17:14+00:00

Problem:
When AFR xlator initialises, it uses the name of the client xlators
below it for storing the pending changelogs (xattrs). This can be
problem when some other xlator is loaded in between AFR and the client.
Though that is a trivial 'traverse-graph-till-the-client-and-use-the-name'
fix in AFR's init(), there are other issues like when there's no client
xlator at all when, say, AFR is moved to the server side.

Fix:
The client xlator names are currenly unique and stored as
brickinfo->brick_ids. So persist these ids as comma separated values in
AFR's volume_options and use them as xattr values during init().

Change-Id: Ie761ffeb3373a4c4d85ad05c84a768c4188aa90d
BUG: 1285152
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/12738
Tested-by: NetBSD Build System 
Tested-by: Gluster Build System 
Reviewed-by: Atin Mukherjee

glusterd: Return better error messages for probe and detach failures

2015-09-01T07:01:04+00:00

We handle some specific errors and return good error messages for those, but for
the default case where the error code is not recognised we just report "unknown
errno". This patch attempts to at least return the output of strerror to provide
more informative errors.

BUG: 1257149
Change-Id: I0027e74e41adac4ab0c0a929c6fff56878bf39c8
Signed-off-by: Brad Hubbard 
Reviewed-on: http://review.gluster.org/12021
Reviewed-by: Niels de Vos 
Tested-by: NetBSD Build System 
Reviewed-by: Atin Mukherjee

cluster/ec: Make background healing optional behavior

2015-07-06T10:31:27+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

cluster/afr : truncate all sinks files

2015-06-27T02:33:08+00:00

Problem : During data self-heal of sparse files,
sparseness of files is lost.

Cause : Earlier, only files with larger ia_size in sinks
were being truncated to ia_size of source. This caused
checksum mismatch of sparse blocks when ia_size of files
in sinks were lesser than ia_size of source file.
Leading to unnecessary healing of sparse blocks.
As a result of which sparseness of files was lost.

Solution : truncate files in all the sinks irrespective of
their size with respect to the source file. After this change,
checksum won't mismatch for sparse blocks and heal won't
be triggered. As a result, sparseness of the files will
be preserved.

Other fixes in this patch :
1) in afr_does_size_mismatch(), check for mismatch only
in sources. Previously, the check was being done for all
children in a replica.

2) in __afr_selfheal_data_checksums_match(), check checksum
mismatch only for children with valid responses.

Change-Id: Ifcdb1cdc9b16c4a8a7867aecf9fa94b66e5301c2
BUG: 1232238
Signed-off-by: Anuradha Talur 
Reviewed-on: http://review.gluster.org/11252
Reviewed-by: Prasanna Kumar Kalever
Reviewed-by: Ravishankar N 
Reviewed-by: Pranith Kumar Karampuri 
Reviewed-by: Krutika Dhananjay 
Tested-by: NetBSD Build System

afr: Block fops when file is in split-brain

2015-06-27T02:30:33+00:00

For directories, block metadata FOPS.
For non-directories, block data and metadata FOPS.
Do not block entry FOPS.

Change-Id: Id7f656f4a513b9d33c457dd7f2d58028dbef8e61
BUG: 1235007
Signed-off-by: Ravishankar N 
Reviewed-on: http://review.gluster.org/11371
Reviewed-by: Pranith Kumar Karampuri 
Tested-by: NetBSD Build System

Tests portability: umount(8)

2015-06-09T13:24:47+00:00

1) Avoid hangs on unmounting NFS on NetBSD

NetBSD umount(8) on a NFS mount whose server is gone will wait forever
because umount(8) calls realpath(3) and tries to access the mount before
it calls unmount(2). The non-portable, NetBSD-specific umount -R flag
prevent that behavior.

We therefore introduce UMOUNT_F, defined as "umount -f" on Linux and
"umount -f -R" on NetBSD to take care of forced unmounts, especially
in the NFS case.

2) Enforce usage of force_umount wrapper with timeout

Whenever umount is used it should be wrapped in force_umount with
tiemout handling. That saves us timing issues, and it handles the
NetBSD NFS case.

3) Cleanup kernel cache flush.

We used (cd $M0 && umount $M0 )	as a portable kernel cache flush
trick, but it does not flush everything we need on Linux. Introduce
a drop_cache() shell function that reverts to previously used
echo 3 > /proc/sys/vm/drop_caches on Linux, and keeps
(cd $M0 && umount $M0 ) on other systems.

BUG: 1129939
Change-Id: Iab1f5a023405f1f7270c42b595573702ca1eb6f3
Signed-off-by: Emmanuel Dreyfus 
Reviewed-on: http://review.gluster.org/11114
Tested-by: NetBSD Build System 
Reviewed-by: Kaleb KEITHLEY 
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

Tests: use a portable way to flush kernel cache

2015-05-07T11:37:56+00:00

On Linux, kernel cache can be flushed using
echo 3 > /proc/sys/vm/drop_caches

This non-portable approach can be replaced by an on-purpose
failed attempt to unmount: if the mount point is the current
directory and umount is called, the kernel will flush inodes
until it realize it cannot complete the operation because
root of filesystem is busy:
( cd $M0 ; umount $M0 )

Unfortunately this does not flush everything. Entries may
still be present in the kenrel FUSE cache. Using $GFS to
mount the filesystem ensure --entry-timeout=0 and clears
this problem.

Some stall information may also remain in glusterfs caches,
and that may have to be adressed by appropriate volume option.
For instance tests/bugs/rpc/bug-954057.t needs to disable
performance.stat-prefetch. Qtherwise, root's new credentials
are not evaluated after root-quash is enabled. The test could
also be done with performance.stat-prefetch enabled using
various tricks: copying the file to read, creating a hard link
on it, or just waiting long enough for metadata cache to expire.

BUG: 1129939
Change-Id: I54929e899d55c04dcd9d947809133549f01fd0e1
Signed-off-by: Emmanuel Dreyfus 
Reviewed-on: http://review.gluster.org/10411
Reviewed-by: Dan Lambright 
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur

tests: Check aux umount is unmounted for quota tests

2015-05-04T08:23:28+00:00

Change-Id: If57d08f3446755ea41f66ca258efcc8ea5a89063
BUG: 1217701
Signed-off-by: Pranith Kumar K 
Reviewed-on: http://review.gluster.org/10480
Tested-by: NetBSD Build System
Tested-by: Gluster Build System 
Reviewed-by: Vijay Bellur