Kefu Chai [Wed, 24 Dec 2025 05:55:26 +0000 (13:55 +0800)]
debian/control: add iproute2 to build dependencies
Test scripts like qa/tasks/cephfs/mount.py expect the ip command to be
available in the container environment. Without it, tests fail with:
```
/bin/bash: line 1: ip: command not found
File "/ceph/qa/tasks/cephfs/mount.py", line 96, in cleanup_stale_netnses_and_bridge
p = remote.run(args=['ip', 'netns', 'list'],
...
teuthology.exceptions.CommandFailedError: Command failed with status 127: 'ip netns list'
```
Add iproute2 to the debian package build dependencies when the
<pkg.ceph.check> build profile is enabled. This ensures the package is
available during container-based builds, since buildcontainer-setup.sh
→ script/run-make.sh → install-deps.sh → debian/control → generated
dependency package chain respects build profiles configured via
`FOR_MAKE_CHECK` and `WITH_CRIMSON` environment variables set in
Dockerfile.build.
David Galloway [Tue, 16 Dec 2025 22:08:00 +0000 (17:08 -0500)]
install-deps: Replace apt-mirror
apt-mirror.front.sepia.ceph.com has happened to always work because we set up CNAMEs to gitbuilder.ceph.com.
That host is making its way to a new home upstate (literally and figuratively) so we'll get rid of the front subdomain since it's publicly accessible anyway and add TLS while we're at it.
Thomas Lamprecht [Wed, 31 Jul 2024 07:48:08 +0000 (09:48 +0200)]
debian/control: record python3-packaging dependency for ceph-volume
Commit 0985e201 added `from packaging import version` to
ceph_volume/util/encryption.py. On Debian and its derivatives, the
packaging module is provided by the python3-packaging package.
Add python3-packaging to ceph-volume's runtime dependencies to ensure
this import is satisfied.
Signed-off-by: Thomas Lamprecht <t.lamprecht@proxmox.com> Signed-off-by: Max R. Carrara <m.carrara@proxmox.com>
(cherry picked from commit a6ab6c63853d071f7383b0ce51d278de92fc3960)
Afreen Misbah [Tue, 21 Oct 2025 16:37:46 +0000 (22:07 +0530)]
mgr/dashboard: Carbonize the Change Password Form
Fixes https://tracker.ceph.com/issues/73193
- using carbon based stylings, typography and components
- used grid layout for form arrangement
- breadcrumb is slightly off, which needs to be fixed by applying grid layout to the app shell
Nitzan Mordechai [Wed, 22 Oct 2025 05:41:56 +0000 (05:41 +0000)]
tasks/cbt_performance: Tolerate exceptions during performance data updates
If an exception occurs during the POST request to update CBT performance,
log the error instead of failing the entire job. This ensures that
intermittent update failures do not block the main workflow.
Afreen Misbah [Mon, 17 Nov 2025 05:01:45 +0000 (10:31 +0530)]
mgr/dashboard: Set max subsystem count to 512 rather than 4096
Fixes https://tracker.ceph.com/issues/73867
- regression from https://github.com/ceph/ceph/pull/64477/files
- removing frontend valdations as this values are volatiel and require changes every release. Nvmeof is seeting these and validating as well.
Kefu Chai [Tue, 21 Oct 2025 03:25:00 +0000 (11:25 +0800)]
debian: Use system packages for cephadm bundled dependencies
Configure the Debian build to use CEPHADM_BUNDLED_DEPENDENCIES=deb,
which instructs the cephadm build script to bundle dependencies from
system-installed Debian packages instead of downloading from PyPI.
This change addresses build failures in restricted network environments
where Debian build tools do not permit internet access. By leveraging
the Debian package support added in commit 9378a2988e1, the build now
uses python3-markupsafe, python3-jinja2, and python3-yaml packages
that are already installed as build dependencies.
This approach mirrors the existing RPM packaging workflow, ensuring
consistent behavior across different distribution package formats.
Kefu Chai [Tue, 21 Oct 2025 03:26:25 +0000 (11:26 +0800)]
cephadm/tests: Add tests for deb bundled dependencies
Add container definitions and test cases for building cephadm with
Debian package dependencies. The new test_cephadm_build_from_debs
function mirrors the existing RPM test structure, verifying that:
- Build succeeds when required Debian packages are installed
- Build fails when packages are missing
- Bundled packages are correctly identified as sourced from 'deb'
- All expected packages (Jinja2, MarkupSafe, PyYAML) are included
- The zipapp contains expected package directories
Test environments include Ubuntu 22.04 and 24.04 with and without
the required python3-jinja2, python3-yaml, and python3-markupsafe
packages.
Kefu Chai [Tue, 14 Oct 2025 13:04:42 +0000 (21:04 +0800)]
cephadm/build: Add Debian package support for bundled dependencies
Extends the cephadm build script to support bundling dependencies from
Debian packages in addition to pip and RPM packages. This allows building
cephadm on Debian-based distributions using system packages.
Key changes:
- Add 'deb' to DependencyMode enum to enable Debian package mode
- Implement _setup_deb() to configure Debian dependency requirements
- Add _install_deb_deps() to orchestrate Debian package installation
- Add _gather_deb_package_dirs() to parse Debian package file listings
and locate Python package directories (handles both site-packages and
dist-packages directories used by Debian)
- Add _deps_from_deb() to extract Python dependencies from installed
Debian packages using dpkg/apt-cache tools
- Fix variable reference bug in _install_deps() (deps.mode -> config.deps_mode)
The Debian implementation follows a similar pattern to the existing RPM
support, using dpkg-query and dpkg -L to locate installed packages and
their files, with special handling for Debian naming conventions
(e.g., PyYAML -> python3-yaml).
Kefu Chai [Mon, 10 Nov 2025 04:11:08 +0000 (12:11 +0800)]
cephadm: fix zip_root_entries population in version command
The 'cephadm version --verbose' command was returning an empty
zip_root_entries list because it relied on the private '_files'
attribute of zipimport.zipimporter, which is not reliably populated
across Python versions.
This commit fixes the issue by using the zipfile module to properly
read the archive contents via the loader.archive path. This ensures
that zip_root_entries is correctly populated with the root-level
directories in the zipapp.
This fix is necessary for the cephadm build tests to properly validate
that all expected packages and modules are included in the built zipapp.
Kefu Chai [Mon, 10 Nov 2025 04:10:46 +0000 (12:10 +0800)]
cephadm/tests: fix _dist_info function logic error
The _dist_info helper function had a logic error where it was checking
if 'entry.startswith(entry)' instead of 'entry.startswith(name)'. This
caused the function to always evaluate incorrectly when checking for
.dist-info or .egg-info entries in the zipapp.
This bug was preventing the test assertions from properly validating
that package metadata directories are included in the built cephadm
zipapp.
Kefu Chai [Wed, 15 Oct 2025 07:46:26 +0000 (15:46 +0800)]
debian/control: Add libxsimd-dev build dependency for vendored Arrow
In commit e8460cbd, we introduced the "pkg.ceph.arrow" build profile to
support building with system Arrow packages. However, neither Debian nor
Ubuntu currently ships Arrow packages.
Since WITH_RADOSGW_SELECT_PARQUET is always enabled in debian/rules,
Arrow support is required for all builds. When the pkg.ceph.arrow profile
is not selected, the build uses vendored Arrow. With the recent change to
use AUTO mode for xsimd detection, Arrow will attempt to find system xsimd
>= 9.0.1. Adding libxsimd-dev as a build dependency ensures it's available
for Arrow to detect and use, reducing build time on supported distributions.
On distributions with insufficient xsimd versions (< 9.0.1), Arrow will
automatically fall back to its bundled version.
Kefu Chai [Wed, 15 Oct 2025 07:46:22 +0000 (15:46 +0800)]
cmake/BuildArrow: Use AUTO mode for xsimd dependency detection
Arrow requires xsimd >= 9.0.1 according to arrow/cpp/thirdparty/versions.txt.
Previously, we unconditionally set -Dxsimd_SOURCE=BUNDLED, forcing the use
of Arrow's vendored xsimd regardless of system package availability.
This commit changes to -Dxsimd_SOURCE=AUTO, which allows Arrow's
resolve_dependency mechanism to automatically:
1. Try to find system xsimd package
2. Check if version >= 9.0.1
3. Use system version if found and sufficient
4. Fall back to bundled version otherwise
This reduces build time and dependencies on systems with sufficient xsimd,
while maintaining compatibility with older distributions.
Distribution availability:
- Ubuntu Noble (24.04): libxsimd-dev 12.1.1 (✓ will use system)
- Ubuntu Jammy (22.04): libxsimd-dev 7.6.0 (✗ will use bundled)
- Debian Trixie (13): libxsimd-dev 13.2.0 (✓ will use system)
- CentOS Stream 9: xsimd-devel 7.4.9 (✗ will use bundled)
Edwin Rodriguez [Mon, 18 Aug 2025 20:17:07 +0000 (16:17 -0400)]
mds: add ceph.dir.subvolume get vxattr
This introduces handling for the "ceph.dir.subvolume" virtual xattr on directories.
- Returns ASCII "1" when the directory is a subvolume root, "0" otherwise.
QA:
- Extend CephFS subvolume tests to validate vxattr retrieval and behavior around setting/removal.
Conflicts:
qa/tasks/cephfs/test_subvolume.py
- resolved conflicts of https://github.com/ceph/ceph/commit/3d5dd1235316775a9dbfdd225c76da09efe02e31
not backported to tentacle yet.
mds: fix rank 0 marked damaged if stopping fails after Elid flush and log trimmed
steps to reproduce
../src/vstart.sh --debug --new -x --localhost --bluestore
./bin/ceph tell mds.<rank 0> config set mds_kill_shutdown_at 10
./bin/ceph fs set <fs name> down true
wait for a few seconds and will see the following log from take-over mds
and rank 0 is marked damaged
2025-09-11T16:47:24.591+0800 785dabeaa6c0 -1 log_channel(cluster) log [ERR] : No subtrees found for root MDS rank!
2025-09-11T16:47:24.591+0800 785dabeaa6c0 5 mds.beacon.b set_want_state: up:rejoin -> down:damaged
During shutdown_pass after submitting Elid and trimming mdlog, mds log
will now have only ELid event which does nothing at replay.
After replay, no subtree is found.
Fix this by checking whther MDLog contains only one event.
If so, skip the subtree check for rank 0, and allow it to request
STATE_STOPPED just like the other ranks.
mds/FSMap: fix join_fscid being incorrectly reset for active MDS during filesystem removal
Fix bug where active MDS daemons in remaining filesystems incorrectly
have their join_fscid cleared to FS_CLUSTER_ID_NONE when any other
filesystem is removed.
The issue was caused by variable name shadowing in erase_filesystem()
where the loop variable 'fscid' shadowed the function parameter 'fscid':
Inside loop: if (info.join_fscid == fscid) compared against the
loop variable (remaining FS ID) instead of parameter (removed FS ID)
Renamed the loop variable to 'remaining_fscid' to eliminate the shadowing
and ensure the comparison uses the correct filesystem ID.
Reproducer:
../src/vstart.sh --new -x --localhost --bluestore
FS=b
./bin/ceph osd pool create cephfs.${FS}.meta 64 64 replicated
./bin/ceph osd pool create cephfs.${FS}.data 64 64 replicated
./bin/ceph fs new ${FS} cephfs.${FS}.meta cephfs.${FS}.data
./bin/ceph config set mds.a mds_join_fs a
./bin/ceph config set mds.b mds_join_fs a
./bin/ceph fs fail ${FS}
./bin/ceph fs rm ${FS} --yes-i-really-mean-it
Then from ./bin/ceph fs dump
We can see join_fscid in all active mds filesystem 'a' are reset.
Since there are standby mds with join_fscid=1
MDSMonitor think they have better affinity and trigger switch over.
John Mulligan [Thu, 13 Feb 2025 20:59:42 +0000 (15:59 -0500)]
ceph.spec.in: use rpm macro for python shebang pathfix
To support EL 10 distros, update the source of the pathfix tool (on EL
9+ distros) and use the macro for updating python shebangs that has been
available since at least EL 9.
Casey Bodley [Tue, 19 Aug 2025 13:44:52 +0000 (09:44 -0400)]
rpm: require gcc >= 13.3 regardless of gts_version
when gts_version is not set, bump the required version of gcc-c++ >= 13.3.
move this into a `%if 0%{?gts_version} == 0` block to prevent that from
applying to builds using gts, because the distro probably doesn't
provide a recent enough gcc-c++
John Mulligan [Fri, 27 Jun 2025 15:08:39 +0000 (11:08 -0400)]
ceph.spec.in: conditionalize crimson gts version on el10
EL10 distros come with GCC 14. When crimson was enabled it was always
trying to set gts_version to 13 (gcc-toolset version). Make the use of
gts version conditional on using el versions lower than 10.
Casey Bodley [Sat, 7 Jun 2025 01:43:33 +0000 (21:43 -0400)]
valgrind: wildcard glibc version for dlopen() leak suppression
the original suppression for "dlopen@@GLIBC_2.2.5" is very similar to
several later suppressions for "dlopen@@GLIBC_2.34". add a wildcard to
the original suppression so the rest can be removed
this also helps suppress a new leak, seen with gcc-13:
{
<insert_a_suppression_name_here>
Memcheck:Leak
match-leak-kinds: reachable
fun:malloc
fun:UnknownInlinedFun
fun:decompose_rpath
fun:_dl_map_object
fun:dl_open_worker_begin
fun:_dl_catch_exception
fun:dl_open_worker
fun:_dl_catch_exception
fun:_dl_open
fun:dlopen_doit
fun:_dl_catch_exception
fun:_dl_catch_error
fun:_dlerror_run
fun:dlopen@@GLIBC_2.34
fun:_sub_I_65535_0.0
fun:call_init
fun:call_init
fun:_dl_init
obj:/usr/lib64/ld-linux-x86-64.so.2
obj:*
obj:*
obj:*
obj:*
obj:*
obj:*
obj:*
}
Casey Bodley [Sat, 7 Jun 2025 01:27:20 +0000 (21:27 -0400)]
valgrind: update rocksdb ObjectLibrary leak suppression for gcc-13
the suppression for gcc-13 only differs on two lines, so add wildcards
to match either. the diff between the current suppression and the new
one follows:
rpm: reenable lto for gcc-toolset-13 by requiring 13.3
referenced gcc bug https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113359
shows that it was resolved for 13.3. reenable lto for gcc-toolset-13 by
requiring 13.3 or later
mgr/vol: fix subvolume removal with retained snapshots when osd is full
The order of operation done for subvolume removal with retain-snapshot
option set to true, is reversed. The metadata is updated first
followed by a rename operation on the uuid directory. If the metadata
update operation fails, then the remove operations is failed thereby,
keeping the subvolume metadata consistent with the uuid path.
projects / ceph.git / log