v1.3.8 hpe.7

Base tag: v1.3.8-beta.0

Base branch: release-1.3

Tag: v1.3.8-hpe.7

Branch: hpe-release-1.3.8

This build (v1.3.8 hpe.7) has the following new changes/backports (since 1.3.8-hpe.6):

✔ When a security group does not exist, do not try and delete all the rules in the security group PR#23

Debian Packages:

Located in 10.247.81.77 (Harbor jumpbox in FTC): /home/davide/hcp_kubernetes_builds/v1.3.8-hpe.7/

This build (v1.3.8 hpe.6) has the following new changes/backports (since 1.3.8-hpe.5):

✔ Openstack lb not updating member based on nodeport changes: PR#22

**Note: this build did not include the following commit:

Validating node-ip by having kubelet check its list of IPs bade640575da6e2de273abf02a85dc6553a44b96

Debian Packages:

Located in 10.247.81.77 (Harbor jumpbox in FTC): /home/davide/hcp_kubernetes_builds/v1.3.8-hpe.6/

This build (v1.3.8 hpe.5) has the following new changes/backports (since 1.3.8-hpe.4):

✔ Validating node-ip by having kubelet check its list of IPs bade640575da6e2de273abf02a85dc6553a44b96

These commmits are related to volume attach/detach bug fixes:

✔ Fix issue in updating device path when volume is attached multiple times 9e8edf6baf06cdbf662dab3399c61fb5da6861ca

Use strongly-typed types.NodeName for a node name - too large 54195d590f03a544d78b4449b2fbafaa258fd6df

✔ Fix race conditino in setting node statusUpdateNeeded flag 14cad206f5b4a56276ba7f35ce2710266abb8c0f

✔ Fix race conditino in setting node statusUpdateNeeded flag - already back-ported in 1.3 efaceb28cce760e6be3f7d0b9abf880e92a5bf82

✔ Post event message for volume attachment b9157b7524f51635761f11b875800e15552fbb5a

✔ Fix collisions issues / timeouts for mounts c884297990712c346daa4ff6938d5df7df54167b

Make volume unmount more robust using exclusive mount w/ O_EXCL - too large e13c678e3b04becd6131442e069f1d910241992d

This change supports robust kubelet volume cleanup - too large f19a1148db1b7584be6b6b60abaf8c0bd1503ed3

This build (1.3.8-hpe.4) has the following changes/backports (since 1.3.8-hpe.3):

OpenStack LBaaSV2: EnsureLoadBalancer now updates instead of recreates existing LBs

https://github.com/kubernetes/kubernetes/pull/33276

This build (1.3.8-hpe.3) has the following changes/backports (since 1.3.8-hpe.2):

Corect filtering of OpenStack LBaaS resources to delete

https://github.com/hpcloud/kubernetes/pull/20

This build (1.3.8-hpe.2) has the following changes/backports (since 1.3.8-hpe.1):

Replace Load Balancer Security Groups

PR: https://github.com/hpcloud/kubernetes/pull/19

Source: https://github.com/grahamhayes/kubernetes/commit/7f4b2547bd7d17f1c8794c906f42ebe1cb5c49dc

This build (1.3.8-hpe.1) has the following changes/backports (since 1.3.7-hpe.1):

Node-ip arg in kubelet not used when using cloud provider

Issue: https://github.com/kubernetes/kubernetes/issues/23568

Upstream PR: https://github.com/kubernetes/kubernetes/pull/33123

Downstream PR: https://github.com/hpcloud/kubernetes/pull/18

AWS mounting the wrong volume

Underlying issue: https://github.com/kubernetes/kubernetes/issues/29324

1.3 backport PR: https://github.com/kubernetes/kubernetes/pull/33149

1.3 backport PR: https://github.com/kubernetes/kubernetes/pull/33142

1.3 backport PR: https://github.com/kubernetes/kubernetes/pull/33150

OpenStack: Security Group Rule Generation for Load Balancer

Upstream PR: https://github.com/kubernetes/kubernetes/pull/31921

Downstream Cherry-Pick: https://github.com/hpcloud/kubernetes/pull/14

CAPS-990 - Cinder device reported incorrectly

Upstream PR: https://github.com/kubernetes/kubernetes/pull/33270

Downstream PR: https://github.com/hpcloud/kubernetes/pull/16

This build (1.3.7-hpe.1) has the following new changes (from release-1.3 in Kubernetes):

vSphere zones fix

https://github.com/kubernetes/kubernetes/pull/32147

OpenStack config-drive and Zones fix

https://github.com/kubernetes/kubernetes/pull/31671