v1.2.0

Summary

• Add cnivpctl binary, which is used to manage and view the SecondaryIP assigned by ipamd.
• For safety, remove GC-related code in the main logic, and move the GC recovery process to cnivpctl.
• Use BoltDB to persist ipamd's pool and cooling pool to avoid ip leakage.
• Some bugfixes.

Docker Images

• uhub.service.ucloud.cn/uk8s/cni-vpc-node:1.2.0
• uhub.service.ucloud.cn/uk8s/cni-vpc-ipamd:1.2.0
• uhub.service.ucloud.cn/uk8s/vip-controller:1.2.0

What's Changed

• feat(cnivpc): take snapshot after adding route fails by @fioncat in #20
• feat(cni): remove gc in main cni logic by @fioncat in #21
• [REBUILD] feat: use boltdb database, discard the cache to avoid ip leaks by @fioncat in #22
• feat(cnivpctl): rebuild cli tools, running in master node by @fioncat in #23
• feat(cnivpctl): support get unused ip by @fioncat in #24
• feat(cnivpctl): use workerpool to speed up command by @fioncat in #25
• style(ipamd): use generic method to put ip to database by @fioncat in #26
• feat(cnivpc): ignore route add exists error by @fioncat in #27
• fix(pool): use correct condition to check cooldown by @fioncat in #28
• feat(cnivpctl): add version command by @fioncat in #29
• feat(pool): add recycle ipamd status to avoid cr leakage by @fioncat in #30

Full Changelog: v1.1.0...v1.2.0