Age | Commit message (Collapse) | Author | Files | Lines |
|
The per port memory model provides a more transparent memory usage model
and avoids pool exhaustion due to competing memory requirements for
interfaces. (http://docs.openvswitch.org/en/latest/topics/dpdk/memory/)
Change-Id: I5add0f49cdcdf2fc3d24affee10a275abe3ca46a
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
Fix broken systemd service unit dependecies:
- OVS should start before networking service;
- OVS ports & bridges should not be automatically ifup-ed by
networking service to avoid races, so drop 'auto' for both
(OVS ports are automatically handled when part of an OVS bridge);
- explicitly ifup OVS bridges as part of networking service, but
after all Linux interfaces have been handled;
- use 'allow-ovs br-prv' to let OVS handle br-prv and avoid another
race condition;
While at it, fix some other related issues:
- make OVS service start after DPDK service (if present);
- bump OVS-DPDK compute VMs RAM since since switching from MTU 1500
to jumbo frames for virtual PODs a while ago failed to do so [1];
- avoid creating conflicting reclass linux.network.interfaces entries
for OVS ports by using their name (drop 'ovs_port_' prefix):
* for untagged networks they will override existing common defs;
* for tagged networks, they will create separate entries;
- DPDK scenarios: make gtw01 br-prv members OVS ports to avoid race
conditions after node reboot by letting OVS handle them;
[1] https://developers.redhat.com/blog/2018/03/16/\
ovs-dpdk-hugepage-memory/
Change-Id: I0266ba67f3849b6f7e331a758146b331730bae55
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
The ovs port remains in down state after reboot if "auto" is off.
Also turn off no_wait option for odl-noha scenarios.
Change-Id: I0121b3190869528e5f2e9985f9e9299ac6c6724e
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
Handle noifupdown option for all cmd.run states
with explicit ifup call as well.
Change-Id: Ie855a0810bcfe4a856cf9d29bd0755643d71ff4d
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
- cmp, gtw: bump RAM allocation to accomodate hugepages/VPP;
for now we overcommit, gtw01 resources can probably be lowered;
- submodule: add salt-formula-neutron so we can locally patch it;
- repo:
* FD.IO repos for VPP packages;
* networking-vpp PPA for python-networking-vpp Neutron driver;
- use vpp-router for L3, disable neutron-l3-agent;
- baremetal_init: apply repo config before network (otherwise UCA
repo is missing when trying to install DPDK on baremetal nodes);
- arm64: iommu.passthrough=1 is required on ThunderX for VPP on
newer kernels;
Design quirks:
- vpp service runs as 'neutron' user, which does not exist at the
time VPP is installed and initially started, hence the need to
restart it before starting the vpp-agent service;
- gtw01 node has DPDK, yet to configure it via IDF we use the
compute-specific OVS-targeted parameters like
`compute_ovs_dpdk_socket_mem`, which is a bit misleading;
- vpp-agent requires ml2_conf.ini on ALL compute AND network nodes
to parse per-node physnet-to-real interface names;
- vpp process is bound to core '1' (not parameterized via IDF);
Change-Id: I659f7dbebcab7b154e7b1fb829cd7159b4372ec8
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
* bump salt-formula-maas git submodule;
* sync AArch64 initial salt config with the x86_64 default config;
* bump Pharos git submodule to sync `power_pass` MaaS configuration
paramater naming;
Change-Id: Ic59dd8becb6d83a9e67004c38d51681c88c4be7c
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
JIRA: FUEL-392
Change-Id: Ia21840c7561a14a5eeed3d08bf89eb2dbf9acc3a
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
* ship prebuilt salt master conf for better readability:
- enable x509.sign_remote_certificate (for prx VCP nodes);
* refactor Salt master CA handling:
- preinstall `salt_minion_dependency_packages` and
`salt_minion_reclass_dependencies` inside docker image;
- persistent /etc/pki;
- run salt.minion on cfg01 to generate master keys;
* bump container formulas to 1 Sep 2018 versions or newer:
- inject date into Docker makefile, forcing a fresh fetch of all
salt formulas from upstream git repos;
* workaround broken salt-formula-designate's meta/sphinx.yml:
- the DEB package version of salt-formula-designate uses `cmd.shell`
to query dpkg on the minion, while the git repo version still
uses `cmd.run`, running into parsing issues;
- temporarily disable sphinx metadata generation for designate until
upstream git repo syncs with the DEB version;
* upstream: salt-formula-salt AArch64 salt.control.virt support:
- retire salt-formula-salt git submodule and related patches;
* skip installing reclass distro package (already installed via pip
inside the container);
* limit initial pillar_refresh call to nodes on jumphost;
* remove unused salt-formula-nova git submodule;
JIRA: FUEL-383
Change-Id: I883b825e556f887a5e31f8a43676dcd8ece6dfde
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
* shift MTU from public bridge to physical interface
* add neutron related settings
Change-Id: Ia57d1ca7976968d6e7ee23f58a0abae1a1a256c0
Signed-off-by: Michael Polenchuk <mpolenchuk@mirantis.com>
|
|
Since we reboot all nodes, applying the network configuration via
Salt before reboot is pointless and creates a race condition with
OVS.
While at it, add `--ignore-errors` to ifup call for OVS bridge to
prevent a race condition during linux.network state apply.
Change-Id: I22fe0afaffecd7b850a6b77d7b810ed296bfc9ca
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|
|
* Refactor OPNFV salt-formulas mechanism to resemble upstream git
structure:
- git submodules: add new submodule for each formula we patch;
- create salt-formula-x directories for OPNFV formulas;
- move mcp/metadata/service contents to their each formula subdir;
- use `make patches-import` for patches previously handled by
patch.sh;
- retire patch.sh
* states: add virtual_init:
- mostly based on old salt.sh, which is now obsolete;
- exclude salt-master service restart (it would kill the container);
* scenarios: cleanup (rm cfg01 virtual node def), adopt virtual_init;
* reclass: align our model with prebuilt container's Salt config:
- drop linux:network pillar data (handled by Docker);
- stop applying linux.system state on cfg01;
- align salt user homedir;
- drop salt-formula packages (preprovisioned);
* minor plumbing in deploy.sh and lib.sh;
JIRA: FUEL-383
Change-Id: I28708a9b399d3f19012212c71966ebda9d6fc0ac
Signed-off-by: Alexandru Avadanii <Alexandru.Avadanii@enea.com>
|