vcswatch reports that
this package seems to have a new changelog entry (version
43.0.8, distribution
unstable) and new commits
in its VCS. You should consider whether it's time to make
an upload.
Here are the relevant commit messages:
commit f6b87efcde8407e830232524bcf6e5092244d6e9
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 19 10:41:02 2025 +0200
Do not set use_queue_manager, as this creates conflict.
commit 8724adb573309da7008fa580566afb2f92b06475
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 19 10:36:12 2025 +0200
Revert "controller.pp: do not set use_queue_manager and rabbit_stream_fanout"
This reverts commit 0f4abbbb959a41bd191a8a1e8ea435119a27aa78.
commit 0f4abbbb959a41bd191a8a1e8ea435119a27aa78
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 19 10:31:52 2025 +0200
controller.pp: do not set use_queue_manager and rabbit_stream_fanout
These options are already set by puppet-module-oslo, so this makes a double-defined value.
commit d3c8c98f7bd4cb0b4650b61714e5e033f58ba9df
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 19 10:25:56 2025 +0200
controller.pp: use glance_api_config resource
The glance_config thingy seems gone.
commit e4ee6f201a1fd05639667f2e7e6b93ab37e19780
Author: Thomas Goirand <zigo@debian.org>
Date: Wed Jun 18 16:26:53 2025 +0200
hyperconverged: fix setting-up compute.
commit 438397b5ffba2d815f730fe812dc5861236bc4cc
Author: Thomas Goirand <zigo@debian.org>
Date: Wed Jun 18 16:26:03 2025 +0200
Blacklist test_delete_security_group_rule_with_non_existent_id
commit 5a42ca18658756ba0e906c0c0a01750932701d9f
Author: Thomas Goirand <zigo@debian.org>
Date: Wed Jun 18 14:56:08 2025 +0200
PoC: correctly add ext-net1-v4subnet1 to ext-net1-router1.
commit a5e2c50180de307d7b852a1d44a578625cf0ef28
Author: Thomas Goirand <zigo@debian.org>
Date: Wed Jun 18 14:02:22 2025 +0200
controller.pp: doubled-defined cloudkitty's orchestrator/coordination_url
commit 22ddd54b2589d784bd4183e0e4e4032d6af6bbcc
Author: Thomas Goirand <zigo@debian.org>
Date: Wed Jun 18 13:19:41 2025 +0200
libvirt: Always use disk_cachemodes=['network=writeback', ]
commit f706bfa5de294d738eccbcdea9de9d30d3d4879f
Author: Thomas Goirand <zigo@debian.org>
Date: Wed Jun 18 10:36:34 2025 +0200
oci-cluster-upgrade-openstack-release: add --partial option
This new option makes it possible to perform an upgraded skipping
Neutron upgrades on network nodes and compute, and skipping puppet
activation on these nodes too. The reason is: this makes it possible
to do these operations manually, as there's the risk that rabbitmq
gets a shitstorm of message when restarting Neutron. This way,
Operators can do this in a slowly, manual way, while monitoring the
number of RabbitMQ messages.
commit 0ded0faf607335d3b82634fc5f9c7fe05612024b
Author: Thomas Goirand <zigo@debian.org>
Date: Tue Jun 17 12:29:51 2025 +0200
d/changelog
commit 6af47fe1e7dfc8d5bda1c7edfdbb731e7fd4170b
Author: Thomas Goirand <zigo@debian.org>
Date: Tue Jun 17 12:29:06 2025 +0200
api_rate_whitelist understands IPv6.
commit d6a30579e58558e22083dbf03a161ac99aeb607b
Author: Thomas Goirand <zigo@debian.org>
Date: Tue Jun 17 10:43:29 2025 +0200
d/changelog
commit 67ca99c90889cbc250da06225ac51fc0938713ef
Author: Thomas Goirand <zigo@debian.org>
Date: Tue Jun 17 10:41:18 2025 +0200
Number of gnocchi-api and cloudkitty-processor workers configurable.
This patch makes the number of gnocchi-api process and cloudkitty-processor
workers configurable on the command line.
commit ba5869ab9a6c6e27cf5476461a44b7d3cf95fa87
Author: Thomas Goirand <zigo@debian.org>
Date: Tue Jun 17 10:12:23 2025 +0200
d/changelog
commit 706040c1469a01285327c12ad4839a9aaa13b4bc
Author: Thomas Goirand <zigo@debian.org>
Date: Tue Jun 17 10:11:16 2025 +0200
compute.pp: instance poller not separated.
commit 49c97c1a4c10ad610089a295c0ce773a931d2345
Author: Thomas Goirand <zigo@debian.org>
Date: Tue Jun 17 09:15:27 2025 +0200
d/changelog
commit 2211bedfef110031701cd27f69c9c967a5fe358c
Author: Thomas Goirand <zigo@debian.org>
Date: Fri Jun 13 14:20:38 2025 +0200
Using multiple threads for ceilometer-polling
This patch separates Ceilometer central-polling into multiple threads
for swift, glance, cinder and neutron.
On the compute, ceilometer-instance-poller runs on its own thread.
commit 1033776ebfaddbfa7da9285f556475d71d7c8530
Author: Olivier Chaze <olivier.chaze@infomaniak.com>
Date: Wed Apr 23 15:48:43 2025 +0200
Enable S3 Versioning
commit c4043687b6388d86cd9f2dc4c261dd1d8d4e79cd
Author: Jim Scadden <jims@silentasylum.co.uk>
Date: Mon Jun 16 10:33:37 2025 +0100
Updated LLDP neighbour switchport name regex
Added 'vnet' to the list of permitted prefix. This fixes and issue where
auto-racking was failing on KVM/libvirtd VMs.
commit d396783b84438dea5a27444c4f7a6fa95cfd1605
Author: Thomas Goirand <zigo@debian.org>
Date: Mon Jun 16 11:23:47 2025 +0200
Adding memcache_use_advanced_pool for gnocchi & cloudkitty
commit f0aaab8e45c08858e28a6c8e393ac902211abca4
Author: Thomas Goirand <zigo@debian.org>
Date: Fri Jun 13 14:31:18 2025 +0200
Make ceilometer-middleware optional.
In some cases, it may be desirable to disable ceilometer-middleware
completely, because it can be *very* resource intensive on the billing
side if there's too much traffic on the proxies.
This patch makes it possible to disable the feature if needed, even
if the rest of the billing system is deployed.
commit ea8bfa00c6b07c3fc3c3ccf4d2e235dd9e1aa5d4
Author: Thomas Goirand <zigo@debian.org>
Date: Fri Jun 13 10:37:50 2025 +0200
Fix controller_iptables.pp to use 'jump' in the ipv6 section also.
commit c7332221c1c944cd6e8c6ca867c246a6ec948518
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 12 22:19:47 2025 +0200
oci-disable-puppet: add ConnectTimeout=5
Add -o ConnectTimeout=5 to ssh command and allow to fail
with || true (if a compute was decomitionned for example).
commit 7364fc4a7c729f213858a69e122be952e9857228
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 12 22:03:40 2025 +0200
d/changelog.
commit 25f9837e6e215b25be4fdae033a5b845ddd10730
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 12 22:02:41 2025 +0200
messaging.pp: fix wrong section edit.
commit d1be6b467a348586571607da8a8c6630938af776
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 12 09:37:41 2025 +0200
PoC: Add and run oci-poc-provision-cloudkitty
commit b122f5e21a3135b36deecc64c8b6c85ea60d021f
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Jun 12 08:13:13 2025 +0200
Fix oci-poc-provision-image after wrong --use-daily-image commit.
commit 5c5d559653b5bef686a0da7c72b43122cc4de4ac
Author: Thomas Goirand <zigo@debian.org>
Date: Wed Jun 11 14:24:56 2025 +0200
poc: ocicli cluster-set cl1 --gnocchi-incoming-sacks-use-redis yes
commit 2f7e6086ea8d861f955bf58b6b230336b090d459
Author: Theo Gindre <theo.gindre@infomaniak.com>
Date: Fri Apr 11 11:09:26 2025 +0200
Incoming sacks with redis
commit 9f4174ccab58a205b7ba48d1c16a21e564f46109
Author: Thomas Goirand <zigo@debian.org>
Date: Tue Jun 10 10:53:51 2025 +0200
Initial Trixie support.
commit 669cffb000aa470f78dcd9033b1be41cd1e49cc3
Author: Evan Bebi <evan.bebi@infomaniak.com>
Date: Wed May 21 17:19:09 2025 +0200
Better config for global_request_id
This patch adds logging_context_format_string and
logging_default_format_string to every daemon files,
plus it adds:
'http-request' => 'capture req.hdr(X-Openstack-Request-ID) len 64'
to the haproxy frontend configuration.
This way, we should get a much better tracking of the
X-Openstack-Request-ID in our logs.
commit 997740b0fcf0c13085b3c12748523732317c60fe
Author: Thomas Goirand <zigo@debian.org>
Date: Tue May 27 11:41:35 2025 +0200
oci-first-boot: Set puppet MAX_RUN=9
commit d203b381cc182937f97377643844d6f890024cb7
Author: Thomas Goirand <zigo@debian.org>
Date: Tue May 27 09:19:58 2025 +0200
Add trixie-epoxy key.
commit 1d33142ca28504e8849c3591ad3cfa976b905689
Author: Thomas Goirand <zigo@debian.org>
Date: Mon May 26 15:04:03 2025 +0200
Fix a value for parallel_operations in Gnocchi.
commit c3d5ef17f26a666a8f83473804b454785ce1d5d8
Author: Valentin Chassignol <valentin.chassignol@infomaniak.com>
Date: Thu May 22 10:53:38 2025 +0200
fix: enc.php: SNI certificate is generated for vip-hostname instead of
machine name
commit 3d57ee799b3de4f42513da538e4549cc98f1dd0f
Author: Thomas Goirand <zigo@debian.org>
Date: Wed May 21 08:56:27 2025 +0200
compute.pp: fix cluster directive if not using multiple Ceph AZ.
commit 82f2d096b8396ffa333f92788a07d6d2338165ad
Author: Thomas Goirand <zigo@debian.org>
Date: Tue May 20 14:37:07 2025 +0200
compute.pp: (bugfix) only use Ceilometer quorum queues for Caracal and up.
commit f7a1ba16bb3b176a85102f9002bb814f029b9736
Author: Thomas Goirand <zigo@debian.org>
Date: Tue May 20 14:28:06 2025 +0200
network.pp: And again, for Rocky.
commit aa53b0a7c1291bd35608f0ffe2ad38f818f26af6
Author: Thomas Goirand <zigo@debian.org>
Date: Tue May 20 14:25:45 2025 +0200
network.pp: rpc_response_timeout param sent twice.
commit 05fc9958707936a527c08506a3c3d2611cad5f68
Author: Thomas Goirand <zigo@debian.org>
Date: Tue May 20 10:13:52 2025 +0200
Call ::keystone::logging before calling ::keystone.
commit 1d0eb71b1a3992823884d66487fd554abe60bb9e
Author: Thomas Goirand <zigo@debian.org>
Date: Tue May 20 09:21:44 2025 +0200
Set haproxy_timeout_server to 10m by default
Experience in production shows this is a nicer value than 1m.
commit 26260c8963a196fad692da3c095696b37e6de8d8
Author: Thomas Goirand <zigo@debian.org>
Date: Tue May 20 09:10:34 2025 +0200
Setup "cluster" variable for Ceph
This is needed, so that volumes aren't bound to a single
cinder-volume service in Ceph mode (ie: a volume can be
managed by any volume node).
commit 938491a39c9eb785c8cbd7a223416166633a5853
Author: Thomas Goirand <zigo@debian.org>
Date: Thu Feb 6 10:49:33 2025 +0100
Set new RabbitMQ defaults.
commit 8837b37284217b2fe9e859f804ba353e70ed76e9
Author: Thomas Goirand <zigo@debian.org>
Date: Thu May 15 11:45:19 2025 +0200
Some quick fixes for Trixie
commit e10fd9a4b1ba3a7b5fb418815805681f8954ea44
Author: Evan Bebi <evan.bebi@infomaniak.com>
Date: Thu May 15 11:32:33 2025 +0200
update for keystone use json
![[bug history graph]](https://qa.debian.org/data/bts/graphs/o/openstack-cluster-installer.png){kind=link}