This document provides Edge Conductor system requirements, release location, issues and limitations, and legal information. To learn more, refer to the following sections.
Click for Release Notes details
- New in this Release v0.4.0
- Feature & Version List for Cluster Type v0.4.0
- Known Issues v0.4.0
- Related Documentation v0.4.0
- Where to find the release v0.4.0
- Hardware and Software Requirements v0.4.0
- Legal Disclaimers
Edge Conductor Release 0.4.0 includes the following:
-
DEK (Development Experience Kit) framework and DEK instances created for CAPI (K8s Cluster API), RKE, and KIND clusters.
-
DCE (Distributed Command Executor) framework enabled for operation automation.
-
Offline deployment solution for KIND DEK, RKE DEK (only cluster deployment for RKE) and CAPI DEK.
-
Intel ESP integrated for OS provisioning.
-
BYOH (Bring Your Own Host) as CAPI provider integrated with ESP.
-
SR-IOV NIC supported end-2-end software stack integrated for container workloads. (Support for VMs is planned for a future release.)
-
CRI-O as an optional supported CRI, in addition to container.
-
CPU manager, Prometheus stack and Grafana are integrated.
-
Rook operator and Ceph as the integrated backend storage solution.
-
TLS enabled for Orchestrated service endpoints.
-
Akri supported both USB and ONVIF camera.
-
NFD integrated for node feature discovery.
-
Edge Conductor Tutorials for first-time users.
-
RT Linux Detection on node.
The following table lists available features and versions for each cluster type provided in the Edge Conductor v0.4.0 release.
| No. | EC v0.4.0 Feature | Version | Kind | RKE | CAPI |
|---|---|---|---|---|---|
| 1 | CPU Manager |
N/A | Yes | Yes | No |
| 2 | CRI-O |
1.23.2 | No | No | Yes |
| 3 | Edge Software Provisioner |
2.0.3 | No | Yes | Yes† |
| 4 | Grafana Dashboard |
8.3.6 | Yes | Yes | Yes |
| 5 | Intel-GPU-Plugin |
0.23 | No | Yes | No |
| 6 | Multus |
v3.8 | Yes | Yes | Yes |
| 7 | Offline deployment |
N/A | No | Yes | No |
| 8 | Portainer-ce |
1.0.22 | Yes | Yes | Yes |
| 9 | Prometheus |
2.33.4 | Yes | Yes | Yes |
| 10 | Rook-ceph |
1.8.5 | No | Yes | Yes |
| 11 | Rook-ceph-cluster |
1.8.5 | No | Yes | Yes |
| 12 | SR-IOV |
N/A | No | Yes | Yes |
| 13 | Akri |
0.8.4 | Yes | Yes | Yes |
| 14 | NFD |
0.11.0 | No | Yes | Yes |
| 15 | rt-linux-detection |
N/A | Yes | Yes | Yes |
† Note that Edge Software Provisioner can be run on CAPI clusters when using the BYOH deployment framework.
EPJ-2133
Description: It is not possible to modify the node list in the Experience Kit, after executing conductor init.
Workaround: Add all nodes in the Experience Kit configuration file before executing conductor init.
EPJ-2126
Description: When using ESP as the OS provider to profile the OS of the nodes, the user name and password is not set from the Experience Kit config file, but set from the ESP profile.
Workaround: Use the user name and password set in the official ESP profile to config the node list in a Experience Kit config.
EPJ-2397
Description: For capi-metal3, in some cases, the work node provisioning fails, and the command "kubectl get machine -n metal3" shows that the worker node status is always in provisioning state.
Workaround: Run the command "kubectl delete machine -n metal3" to delete this machine and wait for capi-metal3 to provision the worker node again.
EPJ-2376
Description: When using CAPI deploy cluster with crio as container runtime, there is error message "overlayfs: unrecognized mount option "volatile" or missing value" on the provisioned nodes. This message is by design of the upstream project. Refer to the known issue of CRI-O.
Workaround: No workaround required.
EPJ-2155
Description: The NO_PROXY variable existing on the OS deployed node is not the same as that of the NO_PROXY variable set in EK.
Workaround: Add all private networks in OS deployed node to "global_settings" -> "no_proxy" in EK yml.
EPJ-2849
Description: failed to pull image "10.10.10.1:9000/docker.io/library/rt-linux-detection:latest on CAPI cluster
Workaround: No workaround, it will be fixed in next release.
EPJ-2432
Description: On ClusterAPI clusters, Prometheus helm deployment failed during the first and second deployments.
Workaround: There is no workaround.
EPJ-2436
Description: Edge Conductor, version 0.4.0 does not include the latest functional and security updates. Customers should update to the latest version as it becomes available.
| Image | #Critical | #High | #Medium | #Low | Action |
|---|---|---|---|---|---|
| quay.io/cephcsi/cephcsi:v3.5.1 | 0 | 18 | 103 | 80 | No action |
| rook/ceph:v1.8.5 | See rook-ceph-188 | ||||
| k8s.gcr.io/ingress-nginx/controller:v0.47.0 | 0 | 8 | 4 | 1 | See nginx-0.14.0 |
| docker.io/nfvpe/multus:stable | 0 | 4 | 243 | 288 | No action |
| ghcr.io/k8snetworkplumbingwg/multus-cni:stable | 0 | 4 | 243 | 288 | No action |
| quay.io/ceph/ceph:v16.2.7 | See ceph-ceph-v17.1 | ||||
| k8s.gcr.io/build-image/debian-base:buster-v1.7.2 | 0 | 1 | 2 | 45 | No action |
| k8s.gcr.io/ingress-nginx/controller:v1.1.1 @sha256:0bc88eb15f9e7f84e8e56c14fa5735aaa48 8b840983f87bd79b1054190e660de |
0 | 1 | 0 | 0 | No action |
| docker.io/calico/node:v3.22.1 | 0 | 0 | 16 | 7 | No action |
| quay.io/kubevirt/virt-api:v0.41.0 | 0 | 0 | 1 | 26 | No action |
| quay.io/kubevirt/virt-controller:v0.41.0 | 0 | 0 | 1 | 26 | No action |
| quay.io/kubevirt/virt-operator:v0.41.0 | 0 | 0 | 1 | 26 | No action |
| grafana/grafana:8.3.6 | 0 | 0 | 0 | 1 | No action |
| quay.io/kiwigrid/k8s-sidecar:1.15.6 | 0 | 0 | 0 | 1 | No action |
EPJ-1965
Description: Prometheus Alertmanager exposes a cluster service while implementing a web application on port 80, which does not use TLS.
EPJ-1973
Description: Prometheus server enables web user interface as a cluster service on port 80 without TLS authentication.
EPJ-2403
Description: When deploying rook ceph on CAPI cluster, there is a possibility that pod ceph-objectstore might fail to launch.
Workaround: No workaround
The README file has an overview of the Edge Conductor tool and its capabilities.
Use the Get Started guide to try out Edge Conductor.
Next, follow the Edge Conductor Tutorials to learn how to deploy some simple applications on the Kubernetes cluster you built with the Get Started guide.
See the Guides index for a current list of Edge Conductor user guides.
Download the package from Release Tag for v0.4.0.
Be sure your host meets the following requirements.
Hardware:
- CPU: 2+ cores
- Memory: 2+ GB
OS and System:
- Ubuntu 18.04+ LTS
- DockerCE
- 20.10.3+ (for DockerCE v20)
Copyright (C) 2022 Intel Corporation
SPDX-License-Identifier: Apache-2.0