Accepted neutron 2:29.0.0-4 (source) into unstable

Debian FTP Masters Sun, 11 Oct 2026 12:49:18 -0700
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Format: 1.8
Date: Sun, 11 Oct 2026 16:02:08 +0200
Source: neutron
Architecture: source
Version: 2:29.0.0-4
Distribution: unstable
Urgency: medium
Maintainer: Debian OpenStack <[email protected]>
Changed-By: Thomas Goirand <[email protected]>
Changes:
 neutron (2:29.0.0-4) unstable; urgency=medium
 .
   * Do not mutate the module level fake fixtures in test_agent
     (test_stale_update_ignored appending a port to the shared
     fake_network via NetworkCache.put_port and
     test_port_update_on_dhcp_agents_port_no_ip_change setting the device
     id on the shared fake_port1), which made `stestr run --dynamic`
     builds fail nondeterministically with an extra port_lookup entry in
     the TestNetworkCache tests and with create_dhcp_port never called in
     the TestDeviceManager tests.
   * Make the OVN maintenance thread pool test robust against foreign
     threads: it compared global threading.active_count() snapshots taken
     several seconds apart, so a fire-and-forget thread leaked by another
     test class (BatchNotifier batch threads, L3 HA notification servers,
     ...) exiting in the meantime made `stestr run --dynamic` builds fail
     nondeterministically with "MismatchError: 3 != 4" in
     TestMaintenanceThread.test_start_bounds_thread_pool_executor. The
     test now verifies that the single-worker executor never runs two
     periodic tasks concurrently and stops the MaintenanceThread worker
     deterministically in a cleanup.
   * Fix the segments plugin configuration cache leak between unit tests:
     neutron.services.segments.db._USER_CONFIGURED_SEGMENT_PLUGIN caches
     whether 'segments' is listed in cfg.CONF.service_plugins the first
     time an agent event fires in the test runner process, so a test
     class without the segments service plugin poisoned the check with
     False and every later segment host mapping test silently stopped
     processing agent reports (test_no_starting_agent_is_not_processed
     failing with 'agent_with_no_start_update' not in set(), and
     test_updated_agent_changed_physical_networks with 2 != 0), which
     made `stestr run --dynamic` builds fail nondeterministically. The
     test_segment.py module state leak fix now also resets the cache.
Checksums-Sha1:
 4c21c631874f742e3f6afd580b020691c427bfc9 4965 neutron_29.0.0-4.dsc
 04b5149b23ef8f6b923d83cf3b30aad58885e1df 48564 neutron_29.0.0-4.debian.tar.xz
 d3bb7984ad111be8285cee21b902647c09c94d24 22457 neutron_29.0.0-4_amd64.buildinfo
Checksums-Sha256:
 3b0f7e40aefbf3411117ef06c9414b79ea5211a712a4197440fc4fb0a070b3b5 4965 
neutron_29.0.0-4.dsc
 f709a0e45ecc698bda25b668756a60c798e440b79718874b42d0fe9ebe7096ea 48564 
neutron_29.0.0-4.debian.tar.xz
 7f6a8e1ef9be27affc97f17604fea778c02b51e44677565545da17af56f8e785 22457 
neutron_29.0.0-4_amd64.buildinfo
Files:
 3df4cea6a6763734b85e95ea282cfec8 4965 net optional neutron_29.0.0-4.dsc
 e3f9246dcc82ec971f63d134d790620c 48564 net optional 
neutron_29.0.0-4.debian.tar.xz
 b4035bc31a23a914c6ac8f614dc0c34f 22457 net optional 
neutron_29.0.0-4_amd64.buildinfo
pgp8ldT_QQQc1.pgp
Description: PGP signature
Accepted neutron 2:29.0.0-4 (source) into unstable

Reply via email to
