DEL can be called multiple times, a plugin should return no error if
the device is already removed, and other errors should be returned. It
was the opposite for vlan plugin. This PR fixes it.
Signed-off-by: Quan Tian <qtian@vmware.com>
bump the go-iptables module to v0.4.5 to avoid
concurrency issues with the portmap plugin and
errors related to iptables not able to hold the
lock.
Signed-off-by: Antonio Ojea <antonio.ojea.garcia@gmail.com>
Add the following idempotent functions to iptables utils:
DeleteRule: idempotently delete an iptables rule
DeleteChain: idempotently delete an iptables chain
ClearChain: idempotently flush an iptables chain
Signed-off-by: Antonio Ojea <antonio.ojea.garcia@gmail.com>
It turns out that the portmap plugin is not idempotent if its
executed in parallel.
The errors are caused due to a race of different instantiations
deleting the chains.
This patch does that the portmap plugin doesn't fail if the
errors are because the chain doesn't exist on teardown.
Signed-off-by: Antonio Ojea <antonio.ojea.garcia@gmail.com>
Use a Describe container for the It code block of the
portmap port forward integration test.
Signed-off-by: Antonio Ojea <antonio.ojea.garcia@gmail.com>
Concurrent use of the `portmap` and `firewall` plugins can result in
errors during iptables chain creation:
- The `portmap` plugin has a time-of-check-time-of-use race where it
checks for existence of the chain but the operation isn't atomic.
- The `firewall` plugin doesn't check for existing chains and just
returns an error.
This commit makes both operations idempotent by creating the chain and
then discarding the error if it's caused by the chain already
existing. It also factors the chain creation out into `pkg/utils` as a
site for future refactoring work.
Signed-off-by: Tim Gross <tim@0x74696d.com>
This change sends gratuitous ARP when MAC address is changed to
let other devices to know the MAC address update.
Signed-off-by: Tomofumi Hayashi <tohayash@redhat.com>
When running in a user namespace created by an unprivileged user the
owner of /var/run will be reported as the unknown user (as defined in
/proc/sys/kernel/overflowuid) so any access to the directory will
fail.
If the XDG_RUNTIME_DIR environment variable is set, check whether the
current user is also the owner of /var/run. If the owner is different
than the current user, use the $XDG_RUNTIME_DIR/netns directory.
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
This change introduce priorities for IPs input among CNI_ARGS,
'args' and runtimeConfig. Fix #399.
Signed-off-by: Tomofumi Hayashi <tohayash@redhat.com>
test.sh doesn't exists now as it was separated into two OS-specific
scripts in 4e1f7802db08570702ebbbc0bfa4b82e3b800a78.
Signed-off-by: Sho SHIMIZU <sho.shimizu@gmail.com>
The CNI spec states that for DEL implementations, "when CNI_NETNS and/or
prevResult are not provided, the plugin should clean up as many resources as
possible (e.g. releasing IPAM allocations) and return a successful response".
This change results in the firewall plugin conforming to the spec by not
returning an error whenever the del method is not provided a prevResult.
Signed-off-by: Erik Sipsma <sipsma@amazon.com>
Previously, if an IPAM plugin provided DNS settings in the result to the PTP
plugin, those settings were always lost because the PTP plugin would always
provide its own DNS settings in the result even if the PTP plugin was not
configured with any DNS settings.
This was especially problematic when trying to use, for example, the host-local
IPAM plugin's support for retrieving DNS settings from a resolv.conf file on
the host. Before this change, those DNS settings were always lost when using the
PTP plugin and couldn't be specified as part of PTP instead because PTP does not
support parsing a resolv.conf file.
This change checks to see if any fields were actually set in the PTP plugin's
DNS settings and only overrides any previous DNS results from an IPAM plugin in
the case that settings actually were provided to PTP. In the case where no
DNS settings are provided to PTP, the DNS results of the IPAM plugin (if any)
are used instead.
Signed-off-by: Erik Sipsma <sipsma@amazon.com>
