CHANGELOG.md

v1.0.8

Feature

• Support reserve ip for all workloads #107, please check custom resource workloads
• Support configuring multiple device of vlan cni plugin #114

Bug fix

• Fix veth device name conflict when multi network enabled #115
• Disable rp_filter in container #119
• Fix release IP API race condition #120
• Fix list ip API lists reserved IPs [#121] #121

v1.0.7

Feature

• Support allocating multiple ips, please check usage docs
• Add mtu config in ipvlan and macvlan mode

Enhance

• Fix unbind pod is slow if lots of pod exits at the same time
• Check if device is veth before deleting eth0
• change log level if bridge-nf-call-iptables module is not installed
• Make galaxy pods critical guaranteed scheduling

Clean up

• Remove subnet from floatingip spec

Bug fix

• Fix a race condition that resync released ip due to pod uid missmatch

v1.0.6

Feature

• Release ip of completed pod
• Release ip fully concurrently
• Add prometheus metrics: galaxy_ip_counter, galaxy_schedule_latency, galaxy_cloud_provider_latency
• K8s-vlan Ipvlan cni plugin supports l2, l3s mode

Bug fix

• Ensure crd created before start list and watch crd
• Resync release ip if pod recreates with the same name but different uid

Clean up

• Remove second ipam

v1.0.5

Feature

• Support floatingip for pod with all kinds of workloads or with no owner reference

Bug fix

• Fix race condition of bind/unbind with cluster provider enabled