|
1 |
| -Central Kubernetes Cluster Config |
2 |
| -================================= |
| 1 | +Kubernetes Cluster Config |
| 2 | +========================= |
| 3 | + |
| 4 | +Cluster Options |
| 5 | +--------------- |
| 6 | + |
| 7 | +Three cluster topology approaches were considered for this project. |
| 8 | + |
| 9 | +- **Separate Kubernetes cluster per beamline**. This could be as simple as |
| 10 | + a single server and the K3S installation described in |
| 11 | + `setup_kubernetes` may be sufficient. The documentation at |
| 12 | + https://rancher.com/docs/k3s/ also details how to make a high availability |
| 13 | + cluster, requiring a minimum of 4 servers. |
| 14 | + |
| 15 | +- **A Central Facility Cluster**. A central facility cluster that runs |
| 16 | + all IOCs on its own nodes would keep everything centralized and provide |
| 17 | + economy of scale. However, there are significant issues with routing |
| 18 | + Channel Access, PVA and some device protocols to IOCs running in a |
| 19 | + different subnet to the beamline. DLS spent some time working around these |
| 20 | + issues but eventually abandoned this approach. |
| 21 | + |
| 22 | +- **Central Cluster with Beamline nodes**. This approach uses the central |
| 23 | + cluster but adds beamline nodes that sit in the beamline itself, |
| 24 | + connected to the beamline subnet. This has all the benefits of central |
| 25 | + management but is able to overcome the problems with protocol routing. |
| 26 | + The DLS argus cluster configuration described below is an example of |
| 27 | + how to achieve this. |
| 28 | + |
| 29 | + |
| 30 | + |
| 31 | +DLS Argus Cluster |
| 32 | +----------------- |
3 | 33 |
|
4 | 34 | **TODO** this section will give details of the topology and special
|
5 | 35 | configuration used by the DLS argus cluster to enable running
|
|
0 commit comments