You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: docs/guides/ha-cluster/readme.md
+26-18Lines changed: 26 additions & 18 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -5,10 +5,10 @@ categories: setup
5
5
6
6
# High Availability Cluster
7
7
8
-
Rabbit software supports provisioning of Red Hat GFS2 (Global File System 2) storage. Per Red Hat:
8
+
NNF software supports provisioning of Red Hat GFS2 (Global File System 2) storage. Per RedHat:
9
9
> GFS2 allows multiple nodes to share storage at a block level as if the storage were connected locally to each cluster node. GFS2 cluster file system requires a cluster infrastructure.
10
10
11
-
Therefore, in order to use GFS2, the Rabbit and its associated compute nodes must form a high availability cluster.
11
+
Therefore, in order to use GFS2, the NNF node and its associated compute nodes must form a high availability cluster.
12
12
13
13
## Cluster Setup
14
14
@@ -24,34 +24,42 @@ HPE hardware implements software known as the Hardware System Supervisor (HSS),
24
24
25
25
### Compute Fencing
26
26
27
-
!!! warning
28
-
Usage of the Redfish fencing agent is not yet verified
29
-
30
27
The [Redfish fencing agent](https://github.com/ClusterLabs/fence-agents/tree/main/agents/redfish) from [ClusterLabs](https://github.com/ClusterLabs/fence-agents) should be used for Compute nodes in the cluster. Configure the agent with the following parameters:
31
28
32
29
| Argument | Definition |
33
30
| -------- | ---------- |
34
-
|`--ip=[ADDRESS]`| The IP address or hostname of the compute node's HSS node controller |
35
-
|`--systems-uri=[URI]`| The URI of the Systems object. Must be `/redfish/v1/Systems/Node0`|
36
-
|`--ssl-insecure`| Instructs the use of an insecure SSL exchange |
31
+
|`ip=[ADDRESS]`| The IP address or hostname of the HSS controller |
32
+
|`port=80`| The Port of the HSS controller. Must be `80`|
33
+
|`systems-uri=/redfish/v1/Systems/Node0`| The URI of the Systems object. Must be `/redfish/v1/Systems/Node0`|
34
+
|`ssl-insecure=true`| Instructs the use of an insecure SSL exchange. Must be `true`|
35
+
|`username=[USER]`| The user name for connecting to the HSS controller |
36
+
|`password=[PASSWORD]`| the password for connecting to the HSS controller |
37
37
38
38
39
-
### Rabbit Fencing
39
+
### NNF Fencing
40
40
41
41
!!! info
42
-
Rabbit fencing agent is in active development; the description below is subject to change.
43
-
44
-
Since the Rabbit node is connected to 16 compute blades, careful coordination around fencing of a Rabbit node is required to minimize the impact of the outage. When a Rabbit node is fenced, the corresponding Kubernetes Storage resource (`storages.dws.cray.hpe.com`) is updated with a status of 'Fenced'. The workload manager must observe this change and handle the movement of resources off the Rabbit node and clear the 'Fenced' status before forcibly rebooting the node.
42
+
NNF fencing agent is in active development; the description below is subject to change.
45
43
46
-
Configure the Rabbit agent with the following parameters:
44
+
Configure the NNF agent with the following parameters:
47
45
48
46
| Argument | Definition |
49
47
| -------- | ---------- |
50
-
|`--kubernetes-service-host=[ADDRESS]`| The IP address of the kubeapi server |
51
-
|`--kubernetes-service-port=[PORT]`| The listening port of the kubeapi server |
52
-
|`--service-token-file=[PATH]`| The location of the service token file. The file must be present on all nodes within the cluster |
53
-
|`--service-cert-file=[PATH]`| The location of the service certificate file. The file must be present on all nodes within the cluster |
54
-
|`--nnf-node-name=[RABBIT-NODE-NAME]`| Name of the rabbit node |
48
+
|`kubernetes-service-host=[ADDRESS]`| The IP address of the kubeapi server |
49
+
|`kubernetes-service-port=[PORT]`| The listening port of the kubeapi server |
50
+
|`service-token-file=[PATH]`| The location of the service token file. The file must be present on all nodes within the cluster |
51
+
|`service-cert-file=[PATH]`| The location of the service certificate file. The file must be present on all nodes within the cluster |
52
+
|`nnf-node-name=[NNF-NODE-NAME]`| Name of the NNF node as it is appears in the System Configuration |
53
+
|`api-version=[VERSION]`| The API Version of the NNF Node resource. Defaults to "v1alpha1" |
54
+
55
+
Since the NNF node is connected to 16 compute blades, careful coordination around fencing of a NNF node is required to minimize the impact of the outage. When a Rabbit node is fenced, the corresponding DWS Storage resource (`storages.dws.cray.hpe.com`) status changes. The workload manager must observe this change and follow the procedure below to recover from the fencing status.
56
+
57
+
1. Observed the `storage.Status` changed and that `storage.Status.RequiresReboot == True`
58
+
2. Set the `storage.Spec.State := Disabled`
59
+
4. Wait for a change to the Storage status `storage.Status.State == Disabled`
0 commit comments