The existing example had an implicit overlap between the host network
and some tenant network, use a different network for the host.
Peering network resized to /24 because usually we would have several
network nodes on it instead of creating each their own transfer/peering
network.
Diagrams adjusted accordingly with omnigraffle.
PNGs reexported with 160% size to approximate the previous ones.
Router IPs in Diagrams adjusted to reflect documentation.
Peering Network IPs added to Peering Session Box.
Change-Id: I8778df02330196a8999afd3bbf9a3b87f0e14f7b
While the focus of the routed provider networks documentation
is on the Neutron side of things, this document provides a useful
location to point at relevant Nova configuration.
This reference relates to the Nova scheduler and ensuring that
instance move operations don't result in instances being scheduled
to hosts which do not have access to the required network(s).
Change-Id: Idd7382bd38d70383107d2997205ff880bede8d96
It has been suggested in the Neutron CI meeting to include a section in
the documentation advicing against blind rechecks. It turns out that
such section already exists. What this change does is to move the
section to the first level of the contributors guide, to make it more
visible. This change also improves some wording and adds some examples
of proper recheck comments.
Change-Id: Ib0a00d13a28f98b0a0f26c7233365d04453db4e0
Based on discussions during the Zed PTG (see [1]) mark Prefix Delegation
as experimental.
Add warning to cfg option ipv6_pd_enabled and to IPv6 documentation's
PD section that the feature has low test coverage, and dibbler is
concluded (the tool used for notifing Neutron that a prefix was changed)
see: [2].
[1]: https://etherpad.opendev.org/p/neutron-zed-ptg#L208
[2]: https://github.com/tomaszmrugalski/dibbler#project-status
Related-Bug: #1916428
Change-Id: Ia712eb6f5b6f947aeab89ba7c018771eb4104e37
Noticed a typo while reading the docs, fixed other grammatical
issues while I was here.
Trivial-fix
Change-Id: I9d62281e095688bcbcb1fbab3d1cd5101ae7ad37
The validation is intended mostly for tests and don't make much sense
when running the migration in production because likely there are
already running workloads. This patch changes the default to False so
migration validation must be explicitly asked for.
Change-Id: I5470f61a5e0b55bf682526208c3f57dc0ca6ffd5
Signed-off-by: Jakub Libosvar <libosvar@redhat.com>
tripleo-ansible-inventory stopped working in Wallaby. However, TripleO
now stores the needed ansible-inventory on the undercloud filesystem.
This patch switches from dynamic generation of the Ansible inventory to
use of the already existing inventory file. Fortunately, the format of
the file remained the same as the generated one, so no other changes in
parsing are required.
Closes-Bug: #1966099
Change-Id: I3bdf878617fbe962d56ebb66d59ae7edeb9b7c38
Signed-off-by: Jakub Libosvar <libosvar@redhat.com>
Added support for filtering the QoS rule type list command.
Two new filter flags are added:
- all_supported: if True, the listing call will print all QoS rule
types supported by at least one loaded mechanism driver.
- all_rules: if True, the listing call will print all QoS rule types
supported by the Neutron server.
Both filter flags are exclusive and not required.
Depends-On: https://review.opendev.org/c/openstack/neutron-lib/+/827533
Closes-Bug: #1959749
Change-Id: I41eaab177e121316c3daec34b309c266e2f81979
Table 59 will be used for pps limitation, the pipeline change is:
all original flows with ``goto table 60`` will be changed to
``goto table 59``, while table 59 has a default rule is goto
table 60. Then we can add pps flows to table 59 for all ports.
Basic limit pipeline is:
Ingress: packets get into br-int table 0, before send to table 60,
in table 59, check the destanation MAC and local_vlan ID, if the
dest is resident in this host, do the meter pps action and send
to table 60.
Egress: match src MAC and in_port, before send to table 60,
in table 59, do the meter pps action and send to table 60.
Why table 59? Because for ovs-agent flow structure, all packets
will be send to table 60 to do next actions such as security group.
Between table 0 and table 60, there are tables for ARP poison/spoofing
prevention rules and MAC spoof filtering. We want similar security
checks to take effect first, so it can drop packets before filling
our limit queues (pps limitation based on data forwarding queue).
And we do not want packets go through the long march of security group
flows, in case of performance side effect when there are large amount
of packets try to send, so limit it before goto security group flows.
Partially-Implements: bp/packet-rate-limit
Related-Bug: #1938966
Related-Bug: #1912460
Change-Id: I943f610c3b6bcf05e2e752ca3b57981f523f88a8
Read the chassis bandwidth configuration (stored in the "Chassis"
registers) only once, in the maintenance worker. The SB synchronizer
will call the OVN client Placement extension
"read_initial_chassis_config" method.
This new approach changes how the Placement information is stored. The
Placement extension does not store anymore a local cache of the
resource providers. Instead of this, in future patches, when this
information is required, the Placement extension will retrieve this
information from the SB DB, reading the content from the "Chassis"
registers and parsing the values.
Partial-Bug: #1578989
Change-Id: I160b1dda85596277125c532ea4ce4df8e4d25b63
Dashboard urls syntax changed after the latest Gerrit upgrades. This
change fixes the urls to the correct latest syntax in the
documentation.
Change-Id: I8883eac81c6db4d6bcd96d08c072d8378b07e6e6
Today the sriov qos service plugin blindly blocks creating ports
with minimum bandwidth qos and with direct_physical vnic_type. This was
originally added when only dataplane enforcement was the scope of the
qos service plugin. However in the last many releases we created
placement enforcement for this qos rule regardless of the vnic_type.
So now blindly blocking the port creation is now preventing using the
placement enforcement for this rule for direct_physical ports.
This patch removes this limitation by marking minimum bandwidth as
supported rule for the sriov qos service plugin. The limitation that
data plane enforcement is not supported for this rule remains. The agent
will not even try to apply any kind of rules to these ports as port
binding is not forwarded for the sriov agent at all.
The documentation is extended to explain that placement enforcement now
works while data plane enforcement still not supported.
This is somewhat similar to the case when the support for egress
direction is added to the minimum bandwidth rule, while the sriov data
plane enforcement was not (could not) been implemented for this
direction in the sriov agent. Today the sriov agent simply ignores the
egress direction rules in the minimum bandwidth qos rule during applying
the data plane enforcement.
Closes-Bug: #1949877
Change-Id: I20ad32eac414ff90b551bff940d92cbcfa848101
These are leftover from opendev migration and
without this fix the links results into Not Found.
Mainly need to use "src" in place of "tree" to
get these links working with opendev.org.
Also used git tags where line references are used
as branched references do not persist. And for line
references use #L in place of #n as that's where it
get's redirected.
Also update references for zuul and use of
devstack-vm-gate-wrap.sh in neutron functional jobs.
Change-Id: I92d11c99a17dab80d4b91da49f341f9ba202bcfe
Without this port binding fails with below error:-
Network <nw> is type of vxlan but agent <host> or mechanism
driver only support ['gre', 'local', 'flat', 'vlan'].
Also fix permissions of /opt/stack/devstack in ml2 ovs testing
documentation and added these files to irrelevant-files to skip
running functional jobs as these files are not used in those jobs.
Related-Bug: #1934466
Change-Id: I3ca2ea19bf5e316e580669caab4c607447034a11