* cmake/modules/CephChecks.cmake: detect the existence of struct msghdr,
define HAVE_MSGHDR if it is found
* src/common/CMakeLists.txt: do not compile journald.cc if HAVE_MSGHDR
is FALSE. as in that case, we cannot use sendmsg() to write to
journald unix domain socket
* src/test/CMakeLists.txt, src/test/common/CMakeLists.txt: disable test
exercising journald logging backend if HAVE_MSGHDR is not defined
* src/common/Journald.h: define a dummy JournaldLogger and a dummy
JournaldClusterLogger when HAVE_MSGHDR is not defined, in order to
minimize the change in Log.h and Log.cc, otherwise the source code of
Log.h and Log.cc would be segmented into smaller chunks by
`ifdef HAVE_MSGHDR` macros.
* src/include/config-h.in.cmake: define a new macro named
HAVE_MSGHDR.
Signed-off-by: Kefu Chai <kchai@redhat.com>
acting.size() >= pool.info.min_size is meant to check min_size against
acting set participants, but acting is a vector with placeholders.
actingset is the representation with placeholders removed.
The upshot of this bug is that the activation process will basically
ignore min_size for an ec pool allowing writes in cases where it
shouldn't. PastIntervals::check_new_interval, however, performs
the check correctly, and will therefore discount intervals in which
we really did serve writes as not writeable. This can trigger many
different problem conditions including but not limited to:
- Unfound objects due to accepting a last_update with insufficient
osds
- Lost writes
- Crashes due to peering rules being violated
This bug was originally introduced with recovery below min_size in
e5a96fd, and then preserved through refactors in 749a13d and 95bec9.
7cb818a exposed it with with expansion of recovery below min_size
to include ec pools (acting.size() is sufficient for replicated
pools).
Fixes: https://tracker.ceph.com/issues/48613
Fixes: https://tracker.ceph.com/issues/48417
Signed-off-by: Samuel Just <sjust@redhat.com>
* refs/pull/40025/head:
cephadm: use split cgroup strategy for podman
cephadm: use class to represent container engine
Reviewed-by: Adam King <adking@redhat.com>
This PR fixes https://tracker.ceph.com/issues/50138 by stopping the kafka consumer if the test fails.
Signed-off-by: Kalpesh Pandya <kapandya@redhat.com>
de6c8250a6 added an explicit %dir directive for
a new directory added to the ceph-common package, but -- due to a typo --
neglected to include the "%". As a result, RPM builds started to fail with:
Processing files: ceph-common-17.0.0-2787.gde6c8250.el8.x86_64
error: File must begin with "/": {_libdir}/ceph/denc/
RPM build errors:
File must begin with "/": {_libdir}/ceph/denc/
Fixes: de6c8250a6
Signed-off-by: Nathan Cutler <ncutler@suse.com>
* refs/pull/40500/head:
doc/cephadm: remove keepalived_user from haproxy docs
Reviewed-by: Juan Miguel Olmo <jolmomar@redhat.com>
Reviewed-by: Adam King <adking@redhat.com>
to address the failure like
src/tcmalloc.cc:332] Attempt to free invalid pointer 0x5627da496130
Aborted
we need to link and load libtcmalloc before it is loaded by a shared
library, otherwise the memory chunk allocated by libc will be freed
by tcmalloc after it is loaded.
Signed-off-by: Kefu Chai <kchai@redhat.com>
marker is potional, however, it may also be provided empty
Fixes: https://tracker.ceph.com/issues/50135
Signed-off-by: Yuval Lifshitz <ylifshit@redhat.com>
resurrect 835f2f5f51, which was reverted by
1f885b6c2f
so all tests for `pacific` will be tested on the latest ubuntu version
Signed-off-by: Kefu Chai <kchai@redhat.com>
2d3c6561b4 introduced a new library directory
"%{_libdir}/ceph/denc/" in ceph-common but did not explicitly state that it
should be owned by the package. This caused OBS builds to fail as follows:
[ 5515s] ceph-common-17.0.0-2786.1.x86_64.rpm: directories not owned by a package:
[ 5515s] - /usr/lib64/ceph/denc
Fixes: 2d3c6561b4
Signed-off-by: Nathan Cutler <ncutler@suse.com>
detect the libpmem libraries' version using pkg-config
please note, the version is not identical to the ones defined in
libpmem.h or libpmemobj/base.h
Signed-off-by: Kefu Chai <kchai@redhat.com>
Using subscription-manager will fail in a container so use dnf
config-manager which should work on bare metal as well as in a
container.
Fixes: https://tracker.ceph.com/issues/50118
Signed-off-by: Brad Hubbard <bhubbard@redhat.com>
it can be replaced with entity_name_t::parse(string_view)
also refactor entity_name_t::parse(string_view) a little bit, to
embed the logic of `entity_name_t::parse(const char*...)` in it.
Signed-off-by: Kefu Chai <kchai@redhat.com>
in 6147c09171, "lo" is also skipped when
daemon is trying to find an address to bind. but that change reverts the
fix of 201b592043, to address the problem.
an option named "ms_bind_exclude_lo_iface" is added, it defaults to
"true". but it can be changed to false to allow daemon to bind on "lo".
Fixes: https://tracker.ceph.com/issues/50012
Signed-off-by: Kefu Chai <kchai@redhat.com>
