Jason Dillaman [Wed, 22 Apr 2015 15:27:35 +0000 (11:27 -0400)]
librbd: updated cache max objects calculation
The previous calculation was based upon the image's object size.
Since the cache stores smaller bufferheads, the object size is not
a good indicator of cache usage and was resulting in objects being
evicted from the cache too often. Instead, base the max number of
objects on the memory load required to store the extra metadata
for the objects.
Fixes: #7385
Backport: firefly, hammer Signed-off-by: Jason Dillaman <dillaman@redhat.com>
Jason Dillaman [Tue, 21 Apr 2015 16:59:33 +0000 (12:59 -0400)]
librbd: better handling for duplicate flatten requests
A proxied flatten request could be replayed, resulting in a
-EINVAL error code being generated on the second attempt. Filter
out that error if it is known the parent did exist before the
op started.
Fixes: #11370
Backport: hammer Signed-off-by: Jason Dillaman <dillaman@redhat.com>
rgw: don't use end_marker for namespaced object listing
Fixes: #11437
Don't deal with end_marker if there's a namespace set for the object
listing. The cur_end_marker.empty() check is incorrect if no end_marker
is provided, but a namespace is set. Just avoid this issue entirely, as
it's not needed.
If ImageCtx::init() fails, perf counters are still used when closing
the image. To prevent uninitialized use, it's easiest to always
initialize them before anything that can fail in ImageCtx::init().
Jason Dillaman [Mon, 20 Apr 2015 16:40:33 +0000 (12:40 -0400)]
cls_rbd: get_features needs to support legacy negative tests
During Ceph upgrade testing, older Ceph test suites assume that
get_features will return -ENOENT if provided a missing snapshot.
Support these negative tests until the older releases are no
longer supported.
Fixes: #11380 Signed-off-by: Jason Dillaman <dillaman@redhat.com>
Add a test to exercise rbd unmap. This was inspired by and is mostly
for rbd unmap <image-name> (aka spec), but rbd unmap <device> is tested
as well. Any significant change to krbd.cc or corresponding rbd cli
should be tested in the following three configurations at least:
Make use of krbd_unmap_by_spec() and allow unmapping by <image-name>
(i.e. pool/image@snap spec) with a bunch of careful adjustments to the
fragile command args/options parsing process to make --pool, --image
and --snap options work as expected for rbd unmap <image-name> case.
Current client code relies on cap message to trim unlinked inode
from cache. This method is too sensitive to message ordering. If
client receives the cap message while directory inode still contains
dentries that are referenced by unsafe requests, the inode can't
get trimmed.
The fix is, when rmdir/rename request finishes, try trimming the
unlinked inode again.
Jason Dillaman [Fri, 17 Apr 2015 00:55:53 +0000 (20:55 -0400)]
tests: separate check_PROGRAMS from TESTS
Not all programs that should be built during 'make check' should
be executed as tests. A new variable, check_TESTPROGRAMS, was
introduced to specify which programs should be executed as a test
during a 'make check'.
Signed-off-by: Jason Dillaman <dillaman@redhat.com>
Danny Al-Gaaf [Fri, 10 Apr 2015 09:25:12 +0000 (11:25 +0200)]
test_async_driver.cc: fix UNINIT
Fix for:
CID 1254380 (#1 of 1): Uninitialized scalar variable (UNINIT)
uninit_use_in_call: Using uninitialized value sa. Field
sa.sin_zero is uninitialized when calling connect
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
Danny Al-Gaaf [Wed, 18 Mar 2015 14:03:28 +0000 (15:03 +0100)]
rgw_rest_user.cc: UserQuotas init vars in ctor init list
Fix for:
[src/rgw/rgw_rest_user.cc:631]: (performance) Variable 'bucket_quota' is
assigned in constructor body. Consider performing initialization in
initialization list.
[src/rgw/rgw_rest_user.cc:632]: (performance) Variable 'user_quota' is
assigned in constructor body. Consider performing initialization in
initialization list.
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
Danny Al-Gaaf [Tue, 17 Mar 2015 21:58:19 +0000 (22:58 +0100)]
libcephfs.h: fix unnecessary forward declarations
This is an issue in case non-cpp code. Move the declaration of
inodeno_t and vinodeno_t structs into __cplusplus block where
they are needed.
Fix for cppcheck issue:
[src/include/cephfs/libcephfs.h:89] -> [src/include/cephfs/libcephfs.h:74]:
The struct 'inodeno_t' forward declaration is unnecessary. Type struct is
already declared earlier.
[src/include/cephfs/libcephfs.h:93] -> [src/include/cephfs/libcephfs.h:83]:
The struct 'vinodeno_t' forward declaration is unnecessary. Type struct is
already declared earlier.
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
Danny Al-Gaaf [Tue, 17 Mar 2015 16:12:04 +0000 (17:12 +0100)]
rgw_rest_log.h: remove http_ret from some RGWOp_*_list classes
Remove http_ret from child class since there is already an integer
with the same name in the parent/base class.
Fix for:
[src/rgw/rgw_rest_log.h:79] -> [src/rgw/rgw_rest.h:275]: (warning) The
class 'RGWOp_MDLog_List' defines member variable with name 'http_ret'
also defined in its parent class 'RGWRESTOp'.
[src/rgw/rgw_rest_log.h:99] -> [src/rgw/rgw_rest.h:275]: (warning) The
class 'RGWOp_MDLog_Info' defines member variable with name 'http_ret'
also defined in its parent class 'RGWRESTOp'.
[src/rgw/rgw_rest_log.h:182] -> [src/rgw/rgw_rest.h:275]: (warning) The
class 'RGWOp_DATALog_List' defines member variable with name 'http_ret'
also defined in its parent class 'RGWRESTOp'.
[src/rgw/rgw_rest_log.h:202] -> [src/rgw/rgw_rest.h:275]: (warning) The
class 'RGWOp_DATALog_Info' defines member variable with name 'http_ret'
also defined in its parent class 'RGWRESTOp'.
Signed-off-by: Danny Al-Gaaf <danny.al-gaaf@bisect.de>
ceph-disk: set '--cluster=ceph' in 'prepare' subcommand
otherwise the 'ceph' default value won't be catched by "main_prepare()",
this breaks TEST_corrupt_and_repair_replicated. seems the behaviour
was changed in python 2.7.9, so we move the default value to where
the '--cluster' argument is added instead of setting it in the
main argument parser.
Ken Dreyer [Tue, 14 Apr 2015 16:22:37 +0000 (10:22 -0600)]
Merge pull request #4359 from ceph/wip-fix-ceph-dencoder-build-master
Move ceph-dencoder build to client
Rework mds/Makefile.am to support a dencoder client build
rgw/Makefile.am: Populate DENCODER_SOURCES properly
Dencoder should never be built with tcmalloc
Reviewed-by: Sage Weil <sage@redhat.com> Reviewed-by: Ken Dreyer <kdreyer@redhat.com>
Boris Ranto [Mon, 13 Apr 2015 13:07:03 +0000 (15:07 +0200)]
Rework mds/Makefile.am to support a dencoder client build
The patch adds all the mds sources to DENCODER_SOURCES to allow a
dencoder client build. The patch also splits the Makefile.am file to
better accomodate the change.
projects / ceph.git / log