kuryr-daemon default port can be used as local port by another process

[Alexey Perevalov]

Now-days most linux based OSes have following local port range
net.ipv4.ip_local_port_range = 32768 60999
The default port for CNI Daemon is 50036 and it's in local port range.
If default port was used during deployment we have non-zero probability what that
port port (50036) already was used as local port.

For example:

[root@ccm ~]# netstat -ltapn |grep 50036
tcp 0 0 127.0.0.1:50036 127.0.0.1:2379 ESTABLISHED 7924/kube-apiserver
tcp 0 0 127.0.0.1:2379 127.0.0.1:50036 ESTABLISHED 7783/etcd

In this case daemon won't start, due to address already in use error.

It's not always issue and bug can be worked around just by adding bind_address option into daemon section.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.opendev.org/660152

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/660152
Committed: https://git.openstack.org/cgit/openstack/kuryr-kubernetes/commit/?id=c8a87bf2fd65212d5e0a00db0435dea521b8b97e
Submitter: Zuul
Branch: master

commit c8a87bf2fd65212d5e0a00db0435dea521b8b97e
Author: Alexey Perevalov <email address hidden>
Date: Mon May 20 18:11:42 2019 +0300

    Use non local port for CNI Daemon

    This port 50036 can be already used in the system, due
    to it's in range for egress ports.

    Closes-Bug: 1829188

    Change-Id: Ieee291893ca342867eb408d65402a8576327ac6d
    Signed-off-by: Alexey Perevalov <email address hidden>

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/kuryr-kubernetes 1.1.0 release.