The host list order seen during Ansible handlers may differ to the usual
play host list order, due to race conditions in notifying handlers. This
means that restart_services.yml for RabbitMQ may be included in a
different order than the rabbitmq group, resulting in a node other than
the 'first' being restarted first. This can cause some nodes to fail to
join the cluster. The include_tasks loop was introduced in [1].
This change fixes the issue by splitting the handler into two tasks, and
restarting the first node before all others.
[1] https://review.opendev.org/c/openstack/kolla-ansible/+/763137
Change-Id: I1823301d5889589bfd48326ed7de03c6061ea5ba
Closes-Bug: #1930293
On machines with many cores, we were seeing excessive CPU load on systems
that were not very busy. With the following Erlang VM argument we saw
RabbitMQ CPU usage drop from about 150% to around 20%, on a system with
40 hyperthreads.
+S 2:2
By default RabbitMQ starts N schedulers where N is the number of CPU
cores, including hyper-threaded cores. This is fine when you assume all
your CPUs are dedicated to RabbitMQ. Its not a good idea in a typical
Kolla Ansible setup. Here we go for two scheduler threads.
More details can be found here:
https://www.rabbitmq.com/runtime.html#scheduling
and here:
https://erlang.org/doc/man/erl.html#emulator-flags
+sbwt none
This stops busy waiting of the scheduler, for more details see:
https://www.rabbitmq.com/runtime.html#busy-waiting
Newer versions of rabbit may need additional flags:
"+sbwt none +sbwtdcpu none +sbwtdio none"
But this patch should be back portable to older versions of RabbitMQ
used in Train and Stein.
Note that information on this tuning was found by looking at data from:
rabbitmq-diagnostics runtime_thread_stats
More details on that can be found here:
https://www.rabbitmq.com/runtime.html#thread-stats
Related-Bug: #1846467
Change-Id: Iced014acee7e590c10848e73feca166f48b622dc
This change enables the use of Docker healthchecks for rabbitmq services.
Implements: blueprint container-health-check
Depends-On: https://review.opendev.org/c/openstack/kolla/+/784562
Change-Id: I23a2c2efab858b9ed39c6ce0ec4a82df10e7f93d
This reverts commit 9cae59be51e8d2d798830042a5fd448a4aa5e7dc.
Reason for revert: This patch was found to introduce issues with fluentd customisation. The underlying issue is not currently fully understood, but could be a sign of other obscure issues.
Change-Id: Ia4859c23d85699621a3b734d6cedb70225576dfc
Closes-Bug: #1906288
Main plays are action-redirect-stubs, ideal for import_tasks.
This avoids 'include' penalty and makes logs/ara look nicer.
Fixes haproxy and rabbitmq not to check the host group as well.
Change-Id: I46136fc40b815e341befff80b54a91ef431eabc0
Partially-Implements: blueprint performance-improvements
Config plays do not need to check containers. This avoids skipping
tasks during the genconfig action.
Ironic and Glance rolling upgrades are handled specially.
Swift and Bifrost do not use the handlers at all.
Partially-Implements: blueprint performance-improvements
Change-Id: I140bf71d62e8f0932c96270d1f08940a5ba4542a
This change adds support for encryption of communication between
OpenStack services and RabbitMQ. Server certificates are supported, but
currently client certificates are not.
The kolla-ansible certificates command has been updated to support
generating certificates for RabbitMQ for development and testing.
RabbitMQ TLS is enabled in the all-in-one source CI jobs, or when
The Zuul 'tls_enabled' variable is true.
Change-Id: I4f1d04150fb2b5af085b762890092f87ae6076b5
Implements: blueprint message-queue-ssl-support
Including tasks has a performance penalty when compared with importing
tasks. If the include has a condition associated with it, then the
overhead of the include may be lower than the overhead of skipping all
imported tasks. For unconditionally included tasks, switching to
import_tasks provides a clear benefit.
Benchmarking of include vs. import is available at [1].
This change switches from include_tasks to import_tasks where there is
no condition applied to the include.
[1] https://github.com/stackhpc/ansible-scaling/blob/master/doc/include-and-import.md#task-include-and-import
Partially-Implements: blueprint performance-improvements
Change-Id: Ia45af4a198e422773d9f009c7f7b2e32ce9e3b97
Previously we mounted /etc/timezone if the kolla_base_distro is debian
or ubuntu. This would fail prechecks if debian or ubuntu images were
deployed on CentOS. While this is not a supported combination, for
correctness we should fix the condition to reference the host OS rather
than the container OS, since that is where the /etc/timezone file is
located.
Change-Id: Ifc252ae793e6974356fcdca810b373f362d24ba5
Closes-Bug: #1882553
Including tasks has a performance penalty when compared with importing
tasks. If the include has a condition associated with it, then the
overhead of the include may be lower than the overhead of skipping all
imported tasks. In the case of the check-containers.yml include, the
included file only has a single task, so the overhead of skipping this
task will not be greater than the overhead of the task import. It
therefore makes sense to switch to use import_tasks there.
Partially-Implements: blueprint performance-improvements
Change-Id: I65d911670649960708b9f6a4c110d1a7df1ad8f7
The common role was previously added as a dependency to all other roles.
It would set a fact after running on a host to avoid running twice. This
had the nice effect that deploying any service would automatically pull
in the common services for that host. When using tags, any services with
matching tags would also run the common role. This could be both
surprising and sometimes useful.
When using Ansible at large scale, there is a penalty associated with
executing a task against a large number of hosts, even if it is skipped.
The common role introduces some overhead, just in determining that it
has already run.
This change extracts the common role into a separate play, and removes
the dependency on it from all other roles. New groups have been added
for cron, fluentd, and kolla-toolbox, similar to other services. This
changes the behaviour in the following ways:
* The common role is now run for all hosts at the beginning, rather than
prior to their first enabled service
* Hosts must be in the necessary group for each of the common services
in order to have that service deployed. This is mostly to avoid
deploying on localhost or the deployment host
* If tags are specified for another service e.g. nova, the common role
will *not* automatically run for matching hosts. The common tag must
be specified explicitly
The last of these is probably the largest behaviour change. While it
would be possible to determine which hosts should automatically run the
common role, it would be quite complex, and would introduce some
overhead that would probably negate the benefit of splitting out the
common role.
Partially-Implements: blueprint performance-improvements
Change-Id: I6a4676bf6efeebc61383ec7a406db07c7a868b2a
Some services look for /etc/timezone on Debian/Ubuntu, so we should
introduce it to the containers.
In addition, added prechecks for /etc/localtime and /etc/timezone.
Closes-Bug: #1821592
Change-Id: I9fef14643d1bcc7eee9547eb87fa1fb436d8a6b3
Both include_role and import_role expect role's name to be given
via "name" param instead of "role".
This worked but caused errors with ansible-lint.
See: https://review.opendev.org/694779
Change-Id: I388d4ae27111e430d38df1abcb6c6127d90a06e0
We assume that all groups are present in the inventory, and quite obtuse
errors can result if any are not.
This change adds a precheck that checks for the presence of all expected
groups in the inventory for each service. It also introduces a common
service-precheck role that we can use for other common prechecks.
Change-Id: Ia0af1e7df4fff7f07cd6530e5b017db8fba530b3
Partially-Implements: blueprint improve-prechecks
Make it require uniqueness of resolution as well to avoid later
issues with RabbitMQ going crazy.
Change-Id: I000ba6c62ab44eac0abdf8d5d1f069adfbc6552f
Closes-bug: #1863363
For the CentOS 7 to 8 transition, we will have a period where both
CentOS 7 and 8 images are available. We differentiate these images via a
tag - the CentOS 8 images will have a tag of train-centos8 (or
master-centos8 temporarily).
To achieve this, and maintain backwards compatibility for the
openstack_release variable, we introduce a new 'openstack_tag' variable.
This variable is based on openstack_release, but has a suffix of
'openstack_tag_suffix', which is empty except on CentOS 8 where it has a
value of '-centos8'.
Change-Id: I12ce4661afb3c255136cdc1aabe7cbd25560d625
Partially-Implements: blueprint centos-rhel-8
deploy rabbitmq cluster by train with ipv6 report:
unable to connect to epmd (port 4369) on control-1: address (cannot connect to host/port)
Closes-Bug: #1856725
Change-Id: I36ebb4e196ece8a304269e8c85e39dda72faae50
Signed-off-by: yj.bai <bai.yongjun@99cloud.net>
Adds rabbitmq_server_additional_erl_args variable which
is appended to RABBITMQ_SERVER_ADDITIONAL_ERL_ARGS
environment variable to RabbitMQ server startup script.
This can be used to configure the schedulers.
Docs attached.
Change-Id: Id683c8cc6dac61354ffd94f3b460335b42136ba2
Co-authored-by: Radosław Piliszek <radoslaw.piliszek@gmail.com>
Related-bug: #1846467
Introduce kolla_address filter.
Introduce put_address_in_context filter.
Add AF config to vars.
Address contexts:
- raw (default): <ADDR>
- memcache: inet6:[<ADDR>]
- url: [<ADDR>]
Other changes:
globals.yml - mention just IP in comment
prechecks/port_checks (api_intf) - kolla_address handles validation
3x interface conditional (swift configs: replication/storage)
2x interface variable definition with hostname
(haproxy listens; api intf)
1x interface variable definition with hostname with bifrost exclusion
(baremetal pre-install /etc/hosts; api intf)
neutron's ml2 'overlay_ip_version' set to 6 for IPv6 on tunnel network
basic multinode source CI job for IPv6
prechecks for rabbitmq and qdrouterd use proper NSS database now
MariaDB Galera Cluster WSREP SST mariabackup workaround
(socat and IPv6)
Ceph naming workaround in CI
TODO: probably needs documenting
RabbitMQ IPv6-only proto_dist
Ceph ms switch to IPv6 mode
Remove neutron-server ml2_type_vxlan/vxlan_group setting
as it is not used (let's avoid any confusion)
and could break setups without proper multicast routing
if it started working (also IPv4-only)
haproxy upgrade checks for slaves based on ipv6 addresses
TODO:
ovs-dpdk grabs ipv4 network address (w/ prefix len / submask)
not supported, invalid by default because neutron_external has no address
No idea whether ovs-dpdk works at all atm.
ml2 for xenapi
Xen is not supported too well.
This would require working with XenAPI facts.
rp_filter setting
This would require meddling with ip6tables (there is no sysctl param).
By default nothing is dropped.
Unlikely we really need it.
ironic dnsmasq is configured IPv4-only
dnsmasq needs DHCPv6 options and testing in vivo.
KNOWN ISSUES (beyond us):
One cannot use IPv6 address to reference the image for docker like we
currently do, see: https://github.com/moby/moby/issues/39033
(docker_registry; docker API 400 - invalid reference format)
workaround: use hostname/FQDN
RabbitMQ may fail to bind to IPv6 if hostname resolves also to IPv4.
This is due to old RabbitMQ versions available in images.
IPv4 is preferred by default and may fail in the IPv6-only scenario.
This should be no problem in real life as IPv6-only is indeed IPv6-only.
Also, when new RabbitMQ (3.7.16/3.8+) makes it into images, this will
no longer be relevant as we supply all the necessary config.
See: https://github.com/rabbitmq/rabbitmq-server/pull/1982
For reliable runs, at least Ansible 2.8 is required (2.8.5 confirmed
to work well). Older Ansible versions are known to miss IPv6 addresses
in interface facts. This may affect redeploys, reconfigures and
upgrades which run after VIP address is assigned.
See: https://github.com/ansible/ansible/issues/63227
Bifrost Train does not support IPv6 deployments.
See: https://storyboard.openstack.org/#!/story/2006689
Change-Id: Ia34e6916ea4f99e9522cd2ddde03a0a4776f7e2c
Implements: blueprint ipv6-control-plane
Signed-off-by: Radosław Piliszek <radoslaw.piliszek@gmail.com>
This is to avoid split-brain.
This change also adds relevant docs that sort out the
HA/quorum questions.
Change-Id: I9a8c2ec4dbbd0318beb488548b2cde8f4e487dc1
Closes-Bug: #1837761
Co-authored-by: Radosław Piliszek <radoslaw.piliszek@gmail.com>
Sometimes as cloud admins, we want to only update code that is running
in a cloud. But we dont need to do anything else. Make an action in
kolla-ansible that allows us to do that.
Change-Id: I904f595c69f7276e71692696471e32fd1f88e6e8
Implements: blueprint deploy-containers-action
This allows the install type for the project to be different than
kolla_install_type
This can be used to avoid hitting bug 1786238, since kuryr only supports
the source type.
Change-Id: I2b6fc85bac092b1614bccfd22bee48442c55dda4
Closes-Bug: #1786238
The RabbitMQ role supports namespacing the service via the
project_name. For example, if you change the project_name, the
container name and config directory will be renamed accordingly. However
the log folder is currently fixed, even though the service tries to
write to one named after the project_name. This change fixes that.
Whilst you might generally use vhosts, running multiple RabbitMQ
services on a single node is useful at the very least for testing,
or for running 'outward RabbitMQ' on the same node.
This change is part of the work to support Cells v2.
Partially Implements: blueprint support-nova-cells
Change-Id: Ied2c24c01571327ea532ba0aaf2fc5e89de8e1fb
Docker has no restart policy named 'never'. It has 'no'.
This has bitten us already (see [1]) and might bite us again whenever
we want to change the restart policy to 'no'.
This patch makes our docker integration honor all valid restart policies
and only valid restart policies.
All relevant docker restart policy usages are patched as well.
I added some FIXMEs around which are relevant to kolla-ansible docker
integration. They are not fixed in here to not alter behavior.
[1] https://review.opendev.org/667363
Change-Id: I1c9764fb9bbda08a71186091aced67433ad4e3d6
Signed-off-by: Radosław Piliszek <radoslaw.piliszek@gmail.com>
We don't add extra volumes support for all services in patch [1].
In order to unify the management of the volume, so we need add extra volumes
support for these services.
[1] 12ff28a693
Change-Id: Ie148accdd8e6c60df6b521d55bda12b850c0d255
Partially-Implements: blueprint support-extra-volumes
Signed-off-by: ZijianGuo <guozijn@gmail.com>
Many tasks that use Docker have become specified already, but
not all. This change ensures all tasks that use the following
modules have become:
* kolla_docker
* kolla_ceph_keyring
* kolla_toolbox
* kolla_container_facts
It also adds become for 'command' tasks that use docker CLI.
Change-Id: I4a5ebcedaccb9261dbc958ec67e8077d7980e496
Since Ansible 2.5, the use of jinja tests as filters has been
deprecated.
I've run the script provided by the ansible team to 'fix' the
jinja filters to conform to the newer syntax.
This fixes the deprecation warnings.
Change-Id: I844ecb7bec94e561afb09580f58b1bf83a6d00bd
Closes-bug: #1827370
Several config file permissions are incorrect on the host. In general,
files should be 0660, and directories and executables 0770.
Change-Id: Id276ac1864f280554e98b937f2845bb424d521de
Closes-Bug: #1821579
Since Id724b44a3edd951fa8b06c9f2c347e9ed8c5ffd9, there is a reference to a
non-existent variable, rabbitmq_confs, that causes deployment to fail if
rabbitmq configuration other than config.json is changed.
I'm taking this opportunity to simplify the role, since we can use the Ansible
handler notification system to determine when handlers need to run, without
registering and checking variables. This simpler approach was used in the
haproxy refactor.
Change-Id: Ibe0e7fda93afff741243ff9c350db1c8c6e1e6d3
Closes-Bug: #1816053
With this change, an operator may be able to stop a
service container without stopping all services in a host.
This change is the starting point to start
fast-forward upgrades support.
In next changes new flags will be introducced to disable
stop dataplane services during upgrades.
Change-Id: Ifde7a39d7d8596ef0d7405ecf1ac1d49a459d9ef
Implements: blueprint support-stop-containers
Having all services in one giant haproxy file makes altering
configuration for a service both painful and dangerous. Each service
should be configured with a simple set of variables and rendered with a
single unified template.
Available are two new templates:
* haproxy_single_service_listen.cfg.j2: close to the original style, but
only one service per file
* haproxy_single_service_split.cfg.j2: using the newer haproxy syntax
for separated frontend and backend
For now the default will be the single listen block, for ease of
transition.
Change-Id: I6e237438fbc0aa3c89a3c8bd706a53b74e71904b
