find_package(lttng-ust REQUIRED) fails to find the lttng library without
this change. because find_path(LTTNG_LIBRARY_DIR ...) does not search
in the default library paths. and the second mode of
FIND_PACKAGE_HANDLE_STANDARD_ARGS() does not stop the cmake with a fatal
message, even some of the required vars are missing. so use the
implemetantion from cmake upstream.
Signed-off-by: Kefu Chai <kchai@redhat.com>
New interfaces for fetching extended (and selective) stat information.
Additionally, applications can specify AT_NO_ATTR_SYNC in the flags to
indicate that they want to do a "lazy" statx that just hands out the
inode info from the cache, or AT_SYMLINK_NOFOLLOW to avoid following
symlinks when walking the path.
Signed-off-by: Jeff Layton <jlayton@redhat.com>
DSA keys are being deprecated: http://www.openssh.com/legacy.html
drop.ceph.com will continue to allow the old DSA key but eventually,
users submitting logs using ceph-post-file will run into issues when
OpenSSH completely drops support for the algorithm.
Fixes: http://tracker.ceph.com/issues/14267
Signed-off-by: David Galloway <dgallowa@redhat.com>
and bump up the dependency of python version to 2.7.
py2.6 support was only necessary for older (now unsupported) distro
versions.
Signed-off-by: Kefu Chai <kchai@redhat.com>
since we are not using subvar of ${shlibs:Depends} in python3-* packages,
just exclude them in dh_shlibdeps.
this silences warnings like
```
warning: dpkg-gencontrol: package python3-cephfs: unused substitution
variable ${shlibs:Depends}
```
Signed-off-by: Kefu Chai <kchai@redhat.com>
so we can use subvars like ${python3:Depends} in debian/control.
this silences the warnings like:
```
warning: dpkg-gencontrol: Depends field of package python3-cephfs:
unknown substitution variable ${python3:Depends}
```
Signed-off-by: Kefu Chai <kchai@redhat.com>
Stop shipping /etc/init.d/rbdmap in the Debian packages. Ship the
rbdmap.service systemd unit file instead.
The corresponding change has already been made for RPMs, in
9224ac2ad2.
For Upstart-based systems (eg Ubuntu Trusty), the Debian packages
already contain rbdmap.conf.
(This gets us a tiny bit closer to being able to remove the rbdmap SysV
script from our tree entirely.)
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
* ceph-base: use ${python:Depends} instead of listing the python
dependencies manually, dh_python2 will scan the requirements
of ceph-detect-init. and fill the subst var for us.
* ceph-common: add ${python:Depends}, as it packages ceph,
and ceph-brag client.
* ceph-osd: it packages ceph-disk, so should add ${python:Depends}
as its dependencies.
dh_python2 will figure them out.
Signed-off-by: Kefu Chai <kchai@redhat.com>
* debian/control:
as we have listed the linked libraries in Depends section, for example,
python-rados depends on librados. and we don't need `dpkg-shlibdeps` to
help figure out shared library substvar dependencies for us. by removing
them, we can silence the warnings of
```
warning: dpkg-shlibdeps: package could avoid a useless dependency if
debian/python-rados/usr/lib/python2.7/dist-packages/rados.x86_64-linux-gnu.so
was not linked against libpthread.so.0 (it uses none of the library's
symbols)
```
-lpthread is introduced by `python-config --ldflags` but it turns out we
are not using any symbols from pthread in the extension directly. and
pthread is included in glibc. so this does not added any extra
dependency to python-* pacakges. but it's desirable to have less
warnings.
* debian/rules: exclude python-* packages from dh_shlibdeps, as we will
not use it to prepare the shlib deps substvars for these packages any
more.
Signed-off-by: Kefu Chai <kchai@redhat.com>
some packages do not package python modules or scripts. so override
dh_python2 to exclude them.
this change silences warnings like:
```
warning: dpkg-gencontrol: package ceph-mon: unused substitution
variable ${python:Provides}
```
Signed-off-by: Kefu Chai <kchai@redhat.com>
dh_auto_configure set -DCMAKE_BUILD_TYPE=None so the default does not
take effect at all. this is on purpose, see [1].
and dpkg-buildflags is able to produce the suggested compiling flags by
debian policy. dh_* can pass these flags to cmake only if it works at
compat>=9, see [2].
and we are safe to move to compat 9, as jewel's supported debian based
distro is jessie (with debhelper 9.20150101), ubuntu trusty (with
debhelper 9.20131227ubuntu1)
---
[1] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=701233#35
[2] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=653916
Fixes: http://tracker.ceph.com/issues/16744
Signed-off-by: Kefu Chai <kchai@redhat.com>
This has been deprecated for quite some time, in favour
of using proper vxattr and libcephfs interfaces.
Fixes: http://tracker.ceph.com/issues/16035
Signed-off-by: John Spray <john.spray@redhat.com>
this helps to keep the build rule simpler, and easier to customize
* -j$(NUMJOBS) is taken care of by 'dh --parallel'
* use 'autoreconf' dh add-on to autoconf cleanup
* add dh-autoreconf to Build-Depends
* bump debhelper compatibility level to 8, as jewel's supported
debian based distro is jessie (with debhelper 9.20150101),
ubuntu trusty (with debhelper 9.20131227ubuntu1)
Signed-off-by: Kefu Chai <kchai@redhat.com>
ceph-disk activate-all walks /dev/disk/by-parttypeuuid at boot time. It
is not necessary when udev fires ADD event for each partition and
95-ceph-osd.rules gets a chance to activate a ceph disk or journal.
There are various reasons why udev ADD events may not be fired at
boot (for instance Debian Jessi 8.5 never does it and CentOS 7.2 seems
to be racy in that regard when a LVM root is being used).
Populating /dev/disk/by-parttypeuuid fixes ceph-disk activate-all that
would not work without it. And it guarantees disks are activated at boot
time regardless of wether udev fires ADD events at the right time (or at
all).
The new udev file is a partial resurection of the
60-ceph-partuuid-workaround-rules that was removed by
9f77244b8e. It is given a name that
reflects its new purpose.
Fixes http://tracker.ceph.com/issues/16351
Signed-off-by: Loic Dachary <loic@dachary.org>
plugins do have unresolved symbols. so exclude them from
dpkg-shlibdeps's scan list. this silences warnings like:
```
dpkg-shlibdeps: warning:
debian/ceph-base/usr/lib/ceph/erasure-code/libec_shec_generic.so.1.0.0
contains an unresolvable reference to symbol
_ZN4ceph6buffer3ptrC1EPNS0_3rawE: it's probably a plugin
```
Signed-off-by: Kefu Chai <kchai@redhat.com>
Prior to this change, ceph-base required the "parted" package, which
meant that any installation of ceph-osd, ceph-mon, or ceph-mds would
pull in the parted package.
Move the parted dependency to ceph-osd, since ceph-disk is the only
thing that uses parted.
The advantage of this change is that Mon and MDS systems will not need
to install the parted package.
Fixes: http://tracker.ceph.com/issues/16095
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
The PowerDNS integration code itself is not packaged at this point, so
remove the python-flask dependency from the radosgw package.
Maybe the PowerDNS integration bits could live in a separate
sub-package, eventually.
Fixes: http://tracker.ceph.com/issues/16032
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
* enable it using dh_systemd_enable
* start the target using dh_systemd_start
* move the dh_installinit, dh_systemd_enable, dh_systemd_start calls
down, so they can identify the service files if they care about them.
Fixes: http://tracker.ceph.com/issues/15573
Signed-off-by: Kefu Chai <kchai@redhat.com>
These were added to get /dev/disk/by-partuuid/ symlinks to work on
wheezy. They are no longer needed for the supported distros (el7+,
jessie+, trusty+), and they apparently break dm by opening devices they
should not.
Fixes: http://tracker.ceph.com/issues/15516
Signed-off-by: Sage Weil <sage@redhat.com>
When the package name changed from ceph to ceph-base, dh_installinit
started installing the init script into /etc/init.d/ceph-base. Fix
this by using --name ceph with dh_installinit, which requires
1) naming the .init file ceph-base.ceph.init, and
2) calling dh_installinit separately for each package
Fixes: http://tracker.ceph.com/issues/15329
Signed-off-by: Dan Mick <dan.mick@redhat.com>
Debian Jessie (at least) now has the virtualenv command in
package 'virtualenv', which depends on python3-virtualenv, and
that's how you get the virtualenv command there. Earlier
Debian-based releases still use python-virtualenv. Jessie's
virtualenv command still creates a python2 venv by default.
This will require a followup fix for install-deps.sh
to handle the | correctly, because currently it sends
the string to apt-get, which treats it as a regular expression
alternation of the two package names, so will try to install both.
The problem occurs when packages are installed without Recommended
packages (because jessie's virtualenv package currently Recommends
python-virtualenv); this is the case under pbuilder, and also using
'mk-build-deps --install', and I suspect under other automated
package-building tools.
Note that Build-Requires processing is not specified to
perform the searches in left-to-right order, so even this is
a tenuous workaround. We probably need distro-specific
debian/ directories.
Signed-off-by: Dan Mick <dan.mick@redhat.com>
The python scripts are not yet compatible with python3, yet it is the
default on jessie. Force the creation of the virtualenv to use python2.7
instead. The wheelhouse is already explicitly populated for both python3
and python2.7 by install-deps.sh, regardless of the default interpreter.
Signed-off-by: Loic Dachary <loic@dachary.org>
First, make the Debian package description mention that RGW aims to
implement the Swift API.
Second, replace the RPM package description with the Debian one, both for
consistency and because the Debian one is better.
Signed-off-by: Nathan Cutler <ncutler@suse.com>
The CephFS repair tools are generally useful to run on any node, not
just the MDS nodes themselves.
Move the utilities from the ceph-mds package to the ceph-common package.
Fixes: #15145
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Attempts to install jewel ceph-common, ceph-mon, ceph-osd, and ceph-base
package over infernalis ceph package fail due to files existing in both.
See comment #4 in the tracker issue for a deeper analysis.
http://tracker.ceph.com/issues/15047Fixes: #15047
Signed-off-by: Nathan Cutler <ncutler@suse.com>
Based on feedback from upstream RGW, introduce a new json-format
token structure representing credentials for pass-through authentication
(e.g., LDAP without Keystone's digest authentication).
Signed-off-by: Matt Benjamin <mbenjamin@redhat.com>
When doing a 'make check', 'qa/workunits/ceph-helpers.sh' does expect the 'btrfs'
command to be available. If not, that make many tests failing.
Signed-off-by: Erwan Velu <erwan@redhat.com>
Prior to this change, users could possibly install some ceph
sub-packages with differing versions. Make each sub-package explicitly
depend on a particular version of its dependencies.
I've purposly excluded ceph-test's dependencies for now (see
http://tracker.ceph.com/issues/10989 for background.)
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
The lsb_release executable is being run in multiple places, not least in
src/common/util.cc, which calls it via shell in the collect_sys_info() code
path.
This patch addresses this issue on SUSE- and Debian-derivatives, as well
as reinstating the dependency for RHEL/Fedora after it was dropped in
1560057226.
http://tracker.ceph.com/issues/14906Fixes: #14906
Signed-off-by: Nathan Cutler <ncutler@suse.com>
Drop duplicate rados-classes libraries from ceph-osd and align the RPM
packaging with Debian by using a wildcard to package everything that
autotools puts in the rados-classes directory.
Signed-off-by: Nathan Cutler <ncutler@suse.com>
Flask is used by MON (in the Ceph REST API) and by RGW (powerdns).
Therefore, it only needs to be in the ceph-mon and ceph-radosgw packages.
Also, this commit encloses the specification of the python-flask runtime
dependency in distro-conditional blocks to account for a minor difference
in the package name between RHEL and SUSE.
Signed-off-by: Nathan Cutler <ncutler@suse.com>
The ceph-base package contains files shared between the -mon, -osd, and
-mds. In other words, the MON, OSD, and MDS servers depend on ceph-base,
but ceph-base should not depend on the leaf packages.
Reported-by: Kefu Chai <kchai@redhat.com>
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
SUSE does not allow the use of /usr/libexec, so commit
69291f872e changed the RPM
packaging to put the libexec files into /usr/lib instead. Do the same
thing on Debian.
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Split up the "ceph" package into four new packages:
1. ceph-mon
2. ceph-osd
3. ceph-mds (already done on Debian)
4. ceph-base (files shared among multiple servers)
and then:
5. Make "ceph" into a metapackage that depends on both -mon and -osd
(and -mds, for RPMs).
To describe the outcome of this change another way:
For RPMs:
- "ceph-{mon,osd,mds}" Require: ceph-base.
- "ceph" will become a metapackage that Requires: ceph-{mon,osd,mds}.
For DEBs:
- "ceph-{mon,osd,mds}" will Depends: ceph-base.
- "ceph" will become a metapackage that Depends: ceph-{mon,osd}.
- "ceph" will continue to Recommends: ceph-mds
New users should "yum install ceph-mon" or "yum install ceph-osd" (or
"apt-get install ceph-mon", etc) in order to install the exact daemons
that they need.
http://tracker.ceph.com/issues/10587Fixes: #10587
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
First, it makes sense for both ceph_common.sh and ceph-osd-prestart.sh to
reside in the same directory: make it so.
Second, /usr/lib exists on both RHEL/Fedora and SLE/openSUSE, whereas
the later lacks /usr/libexec. To make this less painful, package
ceph_common.sh and ceph-osd-prestart.sh in /usr/lib/ceph.
Third, allow e.g. FreeBSD to do its own thing by using the $(libexecdir)
Autoconf variable (but set it to /usr/lib in the spec file).
http://tracker.ceph.com/issues/14687Fixes: #14687
Signed-off-by: Nathan Cutler <ncutler@suse.com>
This is required for building python modules with cython, but for some
reason is not a direct dependency (on debian only a recommends) for
the cython package itself.
Refs: #14059
Signed-off-by: Josh Durgin <jdurgin@redhat.com>
The only thing that uses bzip2-devel is RocksDB, and it's optional, not
requirement. Drop the bzip2-devel/libbz2-dev dependency entirely, and
let RocksDB use it only if it is already present.
Fixes: #13981
Signed-off-by: Piotr Dałek <piotr.dalek@ts.fujitsu.com>
ctypes is fundamentally incapable of providing safe callback support
(exceptions during callbacks are not handled properly). To fix this, and also
gain more flexibility, rewrite/port the librbd bindings to Cython instead.
Other Python bindings are expected to migrate to Cython in the future.
Reference: http://tracker.ceph.com/issues/13115
To support this, this change also:
- Introduces a Makefile for pybind which calls setup.py
- Updates the installation files for packages
- Adds a hack to admin/build-doc to build a dummy librbd, avoiding the need to
build the entire client library just to generate documentation (Sphinx needs
to be able to import rbd.so, which depends on librbd.so symbols).
With contributions by Josh Durgin <jdurgin@redhat.com>
Signed-off-by: Hector Martin <marcan@marcan.st>
As the Cython bindings build to a binary module, backwards compatibility with
older librbd ABIs at runtime is not possible, so depend on at least the same
version (or newer).
Signed-off-by: Hector Martin <marcan@marcan.st>
To prepare for the migration of rbd.py to Cython, add Cython as a package
dependency and have the configure script look for it.
Also adds Cython as a dependency for admin/build-doc, and changes the
virtualenv to allow usage of system packages (to avoid having to compile
Cython inside the virtualenv).
With contributions by Josh Durgin <jdurgin@redhat.com>
Signed-off-by: Hector Martin <marcan@marcan.st>
This patch creates rbdmap shell script that is called from init-rbdmap
init script. The patch also renames src/rbdmap configuration file to
src/etc-rbdmap so that rbdmap shell script can be installed via build
system directly. Finally, the patch accomodates these changes in spec
file and build system.
Fixes: #13374
Signed-off-by: Boris Ranto <branto@redhat.com>
cd ./ceph-detect-init ; python setup.py build
Traceback (most recent call last):
File "setup.py", line 23, in <module>
from setuptools import setup
ImportError: No module named setuptools
Signed-off-by: Sage Weil <sage@redhat.com>
The rbd-replay* utilities are useful for Ceph users with RBD clients.
Currently the rbd-replay* utilities ship in the "ceph-test" package, and
we intend this ceph-test package for Ceph developers and contributors,
not normal users.
Move the rbd-replay* utilities to "ceph-common".
http://tracker.ceph.com/issues/12994Fixes: #12994
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Specify the nofile ulimit in one standard place, where everyone expects it
to be. Drop it from the ceph-osd unit file.
Leave upstart and sysvinit untouched for the time being to avoid compat
issues.
Signed-off-by: Sage Weil <sage@redhat.com>
The number of log files is generally bounded; safe to chown these.
Allow ceph group members to write to this dir.
Signed-off-by: Sage Weil <sage@redhat.com>
The boost mt code uses uninitialized memory for extra randomness,
which is a bad idea in general but more importantly makes valgrind
unhappy. Use /dev/urandom instead.
Unfortunately this introduces a link time dependency.. meh!
Fixes: #12736
Signed-off-by: Sage Weil <sage@redhat.com>
If the mimecap RPM or mime-support DEB is not installed, then the
/etc/mime.types file is not present on the system. RGW attempts to read
this file during startup, and if the file is not present, RGW logs an
error:
ext_mime_map_init(): failed to open file=/etc/mime.types ret=-2
Make the radosgw package depend on the mailcap/mime-support packages so
that /etc/mime.types is always available on RGW systems.
http://tracker.ceph.com/issues/11864Fixes: #11864
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
The postrotate script in src/logrotate.conf uses the which and find utilities
to do its work. Although logrotate itself is only a Recommends, I think which
and find are so ubiquitous that it makes sense to have them as hard
dependencies.
Also, I checked and the package names which and findutils are the same on all
the RPM distros we are currently targeting in the spec file.
In Debian, find is also in a package called findutils while the package
containing 'which' is called debianutils.
Signed-off-by: Nathan Cutler <ncutler@suse.com>
Before this patch, the command 'logrotate -f /etc/logrotate.d/ceph'
was generating an error "Failed to reload ceph.target: Job type reload is not
applicable for unit ceph.target".
Before we issue systemctl reload, check that there is at least
one active ceph-* service. (The hyphen is significant.)
Since we use grep, make the grep package a dependency.
http://tracker.ceph.com/issues/12173Fixes: #12173
Signed-off-by: Tim Serong <tserong@suse.com>
Signed-off-by: Lars Marowsky-Bree <lmb@suse.com>
Signed-off-by: Nathan Cutler <ncutler@suse.com>
Given ceph's reputation for voluminous logs, it makes sense for
logrotate to be a hard dependency for all distros.
Signed-off-by: Nathan Cutler <ncutler@suse.com>
The semantic and interface of get_pg are the same, that avoids
duplication and the ceph-helpers.sh version is tested and documented.
Make the ceph-test package dependent on xmlstarlet because it is
needed by ceph-helpers.sh.
Signed-off-by: Loic Dachary <ldachary@redhat.com>
When ceph-objectstore-tool was moved from ceph-test to
ceph by 61cf5da0b5, the ceph package in
debian/control was updated accordingly, as recommended by
https://www.debian.org/doc/debian-policy/ch-relationships.html#s-replaces
The same must be done for the ceph-dbg package because
/usr/lib/debug/usr/bin/ceph-objectstore-too is no longer in
ceph-test-dbg.
Although the change was merged may 6th, 2015
8f23382064, teuthology jobs were not
always failing because packages were not systematically upgraded during
the installation. The missing dependencies that were responsible for
this upgrade problem were fixed by
f898ec1e4e on may 18th, 2015 and all
upgrade tests relying on ceph-*-dbg packages started to fail
systematically after this date.
http://tracker.ceph.com/issues/11546Fixes: #11546
Signed-off-by: Loic Dachary <ldachary@redhat.com>
Prior to this commit, if a user installed the "ceph-common" Debian
package without installing "ceph", then /usr/bin/ceph would crash
because it was missing the ceph_argparse library.
Ship the ceph_argparse library in "ceph-common" instead of "ceph". (This
was the intention of the original commit that moved argparse to "ceph",
2a23eac549)
http://tracker.ceph.com/issues/11388 Refs: #11388
Reported-by: Jens Rosenboom <j.rosenboom@x-ion.de>
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
Debian's debug packages ought to depend on their respective binary
packages. This was the case for many of our ceph packages, but it was
not the case for ceph-test-dbg or rest-bench-dbg.
Add the dependencies on the relevant binary packages, pinned to
"= ${binary:Version}" per convention.
http://tracker.ceph.com/issues/11673Fixes: #11673
Signed-off-by: Ken Dreyer <kdreyer@redhat.com>
