For tests using the QEMU backend the networking type used is controlled by the
NICTYPE variable. If unset or empty NICTYPE defaults to user, i.e.
QEMU User Networking which requires no further configuration.
For more advanced setups or tests that require multiple jobs to be in the same networking the TAP or VDE based modes can be used.
Other backends can be treated just the same as bare-metal setups. Tests can be triggered in parallel same as for QEMU based ones and synchronization primitives can be used. For the physical network according separation needs to be ensured externally where needed as means for machines to be able to access each other.
With QEMU user networking each jobs gets its own isolated network with
TCP and UDP routed to the outside. DHCP is provided by QEMU. The MAC address of
the machine can be controlled with the NICMAC variable. If not set, it is
52:54:00:12:34:56.
os-autoinst can connect QEMU to TAP devices of the host system to
leverage advanced network setups provided by the host by setting NICTYPE=tap.
The TAP device to use can be configured with the TAPDEV variable. If not
defined, it is automatically set to "tap" + ($worker_instance - 1), i.e.
worker1 uses tap0, worker 2 uses tap1 and so on.
For multiple networks per job (see NETWORKS variable), the following numbering
scheme is used:
worker1: tap0 tap64 tap128 ...
worker2: tap1 tap65 tap129 ...
worker3: tap2 tap66 tap130 ...
...The MAC address of each virtual NIC is controlled by the NICMAC variable or
automatically computed from $worker_id if not set.
In TAP mode the system administrator is expected to configure the network, required internet access, etc. on the host as described in the next section.
The complete multi-machine test setup can be provided from the script
os-autoinst-setup-multi-machine provided by "os-autoinst". The script can be
also found online on
https://github.com/os-autoinst/os-autoinst/blob/master/script/os-autoinst-setup-multi-machine
The configuration is applicable for openSUSE and will use Open vSwitch for virtual switch, firewalld (or SuSEfirewall2 for older versions) for NAT and wicked or NetworkManager as network manager. Keep in mind that a firewall is not strictly necessary for operation. The operation without firewall is not covered in all necessary details in this documentation.
|
Note
|
Another way to setup the environment with iptables and firewalld is described on the Fedora wiki. |
|
Note
|
Alternatively
salt-states-openqa contains
necessities to establish such a setup and configure it for all workers with the
tap worker class. They also cover GRE tunnels (that are explained in the next
section).
|
The script os-autoinst-setup-multi-machine can be run like this:
# specify the number of test VMs to run on this host
instances=30 bash -x $(which os-autoinst-setup-multi-machine)The script will install and configure Open vSwitch as well as a service called os-autoinst-openvswitch.service.
|
Note
|
os-autoinst-openvswitch.service is a support service that sets the
vlan number of Open vSwitch ports based on NICVLAN variable - this separates
the groups of tests from each other. The NICVLAN variable is dynamically
assigned by the openQA scheduler.
|
The name of the bridge (default: br1) will be set in
/etc/sysconfig/os-autoinst-openvswitch.
The script will add the bridge device and the tap devices for every multi-machine worker instance.
|
Note
|
The bridge device will also call a script at
/etc/wicked/scripts/gre_tunnel_preup.sh on PRE_UP.
This script needs manual touch if you want to set up multiple
multi-machine worker hosts. Refer to the GRE tunnels section below
for further information.
|
By default all multi-machine workers have to be on a single physical machine. You can join multiple physical machines and its OVS bridges together by a GRE tunnel.
If the workers with TAP capability are spread across multiple hosts, the network must be connected. See Open vSwitch documentation for details.
Create a gre_tunnel_preup script (change the remote_ip value correspondingly
on both hosts):
cat > /etc/wicked/scripts/gre_tunnel_preup.sh <<EOF
#!/bin/sh
action="$1"
bridge="$2"
ovs-vsctl set bridge $bridge rstp_enable=true
ovs-vsctl --may-exist add-port $bridge gre1 -- set interface gre1 type=gre options:remote_ip=<IP address of other host>
EOFAnd call it by PRE_UP_SCRIPT="wicked:gre_tunnel_preup.sh" entry:
# /etc/sysconfig/network/ifcfg-br1
<..>
PRE_UP_SCRIPT="wicked:gre_tunnel_preup.sh"Ensure to make gre_tunnel_preup.sh executable.
|
Note
|
When using GRE tunnels keep in mind that virtual machines inside the ovs bridges have to use MTU=1458 for their physical interfaces (eth0, eth1). If you are using support_server/setup.pm the MTU will be set automatically to that value on support_server itself and it does MTU advertisement for DHCP clients as well. |
Allow worker instances to run multi-machine jobs:
# /etc/openqa/workers.ini
[global]
WORKER_CLASS = qemu_x86_64,tap|
Note
|
The number of tap devices should correspond to the number of the running worker instances. For example, if you have set up 3 worker instances, the same number of tap devices should be configured. |
Enable worker instances to be started on system boot:
systemctl enable openqa-worker@{1..3}Simply run a MM test scenario. For openSUSE, you can find many relevant tests on
o3, e.g. look for networking-related tests like
ping_server/ping_client or wicked_basic_ref/wicked_basic_sut.
To test GRE tunnels, you may want to change the jobs worker classes so the
different jobs are executed on different workers. So you could call
openqa-clone-job like this:
openqa-clone-job \
--skip-chained-deps \ # assuming assets are present
--max-depth 0 \ # clone the entire parallel cluster
--within-instance # create new jobs on the same instance
https://openqa.opensuse.org/tests/3886213 \ # arbitrary job in cluster to clone
_GROUP=0 BUILD+=test-mm-setup \ # avoid interfering with production jobs
WORKER_CLASS:wicked_basic_ref+=,worker_foo \ # ensure one job runs on `worker_foo`
WORKER_CLASS:wicked_basic_sut+=,worker_bar # ensure other job runs on `worker_bar`Also be sure to reboot the worker host to make sure the setup is actually persistent.
You may also start VMs manually to verify the setup.
First, download a suitable image and launch a VM in the same way os-autoinst
would do for MM jobs:
wget http://download.opensuse.org/tumbleweed/appliances/openSUSE-Tumbleweed-Minimal-VM.x86_64-Cloud.qcow2
qemu-system-x86_64 -m 2048 -enable-kvm -vnc :42 -snapshot \
-netdev tap,id=qanet0,ifname=tap40,script=no,downscript=no \
-device virtio-net,netdev=qanet0,mac=52:54:00:13:0b:4a \
openSUSE-Tumbleweed-Minimal-VM.x86_64-Cloud.qcow2The image used here is of course just an example. You need to make sure to assign a unique MAC address (e.g. by adjusting the last two figures in the example; this will not conflict with MAC addresses used by os-autoinst) and use a tap device not used at the same time by a SUT-VM.
Within the VM configure the network like this (you may need to adjust concrete IP addresses, subnets and interface names):
ip link set dev eth0 up mtu 1380
ip a add dev eth0 10.0.2.15/24
ip r add default via 10.0.2.2
echo 'nameserver 8.8.8.8' > /etc/resolv.confThe MTU is chosen in accordance with what the openSUSE test distribution uses
for MM tests and should be below the MTU set on the Open vSwitch bridge device
(e.g. via os-autoinst-setup-multi-machine).
After this it should be possible to reach other hosts. You may also launch a 2nd
VM to see whether the VMs can talk to each other. You may conduct ping tests
similar to the ping_client test mentioned in the previous section (see the
utility function in openSUSE tests
for details). When running ping you can add/remove machines to/from the GRE
network to bisect problematic hosts/connections (via ovs-vsctl add-port … and
ovs-vsctl del-port …).
Boot sequence with wicked (version 0.6.23 and newer):
-
openvswitch (as above)
-
wicked - creates the bridge
br1and tap devices, adds tap devices to the bridge, -
firewalld (or SuSEfirewall2 in older setups)
-
os-autoinst-openvswitch - installs openflow rules, handles vlan assignment
The configuration and operation can be checked with the following commands:
cat /proc/sys/net/ipv4/conf/{br1,eth0}/forwarding # check whether IP forwarding is enabled
ovs-vsctl show # shows the bridge br1, the tap devices are assigned to it
ovs-ofctl dump-flows br1 # shows the rules installed by os-autoinst-openvswitch in table=0
ovs-dpctl show # show basic info on all datapaths
ovs-dpctl dump-flows # displays flows in datapaths
ovs-appctl rstp/show # show rstp information
ovs-appctl fdb/show br1 # show MAC address tableWhen everything is ok and the machines are able to communicate, the ovs-vsctl should show something like the following:
Bridge "br0"
Port "br0"
Interface "br0"
type: internal
Port "tap0"
Interface "tap0"
Port "tap1"
tag: 1
Interface "tap1"
Port "tap2"
tag: 1
Interface "tap2"
ovs_version: "2.11.1"|
Note
|
Notice the tag numbers are assigned to tap1 and tap2. They should have the same number. |
|
Note
|
If the balance of the tap devices is wrong in the workers.ini the tag cannot be assigned and the communication will be broken. |
To list the rules which are effectively configured in the underlying netfilter
(nftables or iptables) use one of the following commands depending on which
netfilter is used.
|
Note
|
Whether firewalld is using nftables or iptables is determined by the
setting FirewallBackend in /etc/firewalld/firewalld.conf. SuSEfirewall2 is
always using iptables.
|
nft list tables # list all tables
nft list table firewalld # list all rules in the specified tableiptables --list --verbose # list all rules with package countsCheck the flow of packets over the network:
-
packets from tapX to br1 create additional rules in table=1
-
packets from br1 to tapX increase packet counts in table=1
-
empty output indicates a problem with os-autoinst-openvswitch service
-
zero packet count or missing rules in table=1 indicate problem with tap devices
As long as the SUT has access to external network, there should be a non-zero packet count in the forward chain between the br1 and external interface.
|
Note
|
To list the package count when nftables is used one needed to use
counters (which can
be added to existing rules).
|
# Enable ip forwarding
sysctl -w net.ipv4.ip_forward=1
sysctl -w net.ipv6.conf.all.forwarding=1
# Install and enable openvswitch
zypper in openvswitch3
systemctl enable --now openvswitch| Host | Network address | Bridge address | Remote IP |
|---|---|---|---|
A |
192.0.2.1/24 |
192.168.42.1/24 |
192.0.2.2 |
B |
192.0.2.2/24 |
192.168.43.1/24 |
192.0.2.1 |
|
Note
|
instead of having two /24 networks, it is also possible to assign addresses from one bigger network (which have the benefit of not needing explicit route assignment). |
Two servers with a single bridge on each side connected with GRE tunnel.
# Create bridge and tunnel
nmcli con add type bridge con.int br0 bridge.stp yes ipv4.method manual ipv4.address "$bridge_address" ipv4.routes 192.168.42.0/23
nmcli con add type ip-tunnel mode gretap con.int gre1 master br0 remote "$remote_ip"
# Test the tunnel with ping
# -M do -- prohibit fragmentation
# -s xxxx -- set packet size
ping -c 3 -M do -s 1300 192.168.42.1
ping -c 3 -M do -s 1300 192.168.43.1Two servers with a one virtual bridge connected with GRE tunnel.
# Create bridge, port and interface
nmcli con add type ovs-bridge con.int br0 ovs-bridge.rstp-enable yes
nmcli con add type ovs-port con.int br0 con.master br0
nmcli con add type ovs-interface con.int br0 con.master br0 ipv4.method manual ipv4.address "$bridge_address" ipv4.routes 192.168.42.0/23
# Create GRE tunnel
nmcli con add type ovs-port con.int gre1 con.master br0
nmcli con add type ip-tunnel mode gretap con.int gre1 master gre1 remote "$remote_ip"
# Test the tunnel
ping -c 3 -M do -s 1300 192.168.42.1
ping -c 3 -M do -s 1300 192.168.43.1# ovs-vsctl show
de1f31e9-1b51-4cc3-954a-4e037191ac07
Bridge br0
Port br0
Interface br0
type: internal
Port gre1
Interface gre1
type: system
ovs_version: "3.1.0"
openvswitch uses flow-based GRE tunneling, i.e. one interface gre_sys for all tunnels, the tunnel can be created by ovs-vsctl. After that, everything works as expected.
# Create bridge, port and interface
nmcli con add type ovs-bridge con.int br0 ovs-bridge.rstp-enable yes
nmcli con add type ovs-port con.int br0 con.master br0
nmcli con add type ovs-interface con.int br0 con.master br0 ipv4.method manual ipv4.address "$bridge_address" ipv4.routes 192.168.42.0/23
# Create GRE tunnel
ovs-vsctl add-port br0 gre1 -- set interface gre1 type=gre options:remote_ip="$remote_ip"
# Test the tunnel
ping -c 3 -M do -s 1300 192.168.42.1
ping -c 3 -M do -s 1300 192.168.43.1# ovs-vsctl show
de1f31e9-1b51-4cc3-954a-4e037191ac07
Bridge br0
Port br0
Interface br0
type: internal
Port gre1
Interface gre1
type: gre
options: {remote_ip="192.0.2.2"}
ovs_version: "3.1.0"
Virtual Distributed Ethernet provides a software switch that runs in user space. It allows to connect several QEMU instances without affecting the system’s network configuration.
The openQA workers need a vde_switch instance running. The workers reconfigure the switch as needed by the job.
To start with a basic configuration like QEMU user mode networking, create a machine with the following settings:
-
VDE_SOCKETDIR=/run/openqa -
NICTYPE=vde -
NICVLAN=0
Start the switch and user mode networking:
systemctl enable --now openqa-vde_switch
systemctl enable --now openqa-slirpvdeWith this setting all jobs on the same host would be in the same network and share the same SLIRP instance.