зеркало из https://github.com/Azure/azurehpc.git
Merge pull request #108 from Azure/js-peer_network_example
Added an example for a peer network cluster setup
This commit is contained in:
Jon Shelley
GitHub
Коммит
1bfd93546c
|
|
@ -0,0 +1,119 @@
|
|||
{
|
||||
"location": "variables.location",
|
||||
"resource_group": "variables.resource_group",
|
||||
"install_from": "headnode",
|
||||
"admin_user": "hpcadmin",
|
||||
"variables": {
|
||||
"hpc_image": "OpenLogic:CentOS-HPC:7.7:latest",
|
||||
"location": "southcentralus",
|
||||
"resource_group": "<NOT-SET>",
|
||||
"vm_type": "Standard_D8_v3",
|
||||
"compute_vm_type": "Standard_HB60rs",
|
||||
"compute_instances": 2,
|
||||
"vnet_resource_group": "variables.resource_group",
|
||||
"peer_network_resource_group": "<NOT-SET>",
|
||||
"peer_vnet_name": "<NOT_SET>"
|
||||
},
|
||||
"vnet": {
|
||||
"resource_group": "variables.vnet_resource_group",
|
||||
"name": "hpcvnet",
|
||||
"address_prefix": "10.2.0.0/20",
|
||||
"subnets": {
|
||||
"compute": "10.2.4.0/22"
|
||||
},
|
||||
"peer": {
|
||||
"Network1": {
|
||||
"resource_group": "variables.peer_network_resource_group",
|
||||
"vnet_name": "variables.peer_vnet_name"
|
||||
}
|
||||
}
|
||||
},
|
||||
"resources": {
|
||||
"headnode": {
|
||||
"type": "vm",
|
||||
"vm_type": "variables.vm_type",
|
||||
"public_ip": true,
|
||||
"image": "variables.hpc_image",
|
||||
"subnet": "compute",
|
||||
"tags": [
|
||||
"cndefault",
|
||||
"nfsserver",
|
||||
"pbsserver",
|
||||
"loginnode",
|
||||
"localuser",
|
||||
"disable-selinux"
|
||||
]
|
||||
},
|
||||
"compute": {
|
||||
"type": "vmss",
|
||||
"vm_type": "variables.compute_vm_type",
|
||||
"instances": "variables.compute_instances",
|
||||
"image": "variables.hpc_image",
|
||||
"subnet": "compute",
|
||||
"tags": [
|
||||
"nfsclient",
|
||||
"pbsclient",
|
||||
"cndefault",
|
||||
"localuser",
|
||||
"disable-selinux"
|
||||
]
|
||||
}
|
||||
},
|
||||
"install": [
|
||||
{
|
||||
"script": "disable-selinux.sh",
|
||||
"tag": "disable-selinux",
|
||||
"sudo": true
|
||||
},
|
||||
{
|
||||
"script": "cndefault.sh",
|
||||
"tag": "cndefault",
|
||||
"sudo": true
|
||||
},
|
||||
{
|
||||
"script": "nfsserver.sh",
|
||||
"tag": "nfsserver",
|
||||
"sudo": true
|
||||
},
|
||||
{
|
||||
"script": "nfsclient.sh",
|
||||
"args": [
|
||||
"$(<hostlists/tags/nfsserver)"
|
||||
],
|
||||
"tag": "nfsclient",
|
||||
"sudo": true
|
||||
},
|
||||
{
|
||||
"script": "localuser.sh",
|
||||
"args": [
|
||||
"$(<hostlists/tags/nfsserver)"
|
||||
],
|
||||
"tag": "localuser",
|
||||
"sudo": true
|
||||
},
|
||||
{
|
||||
"script": "pbsdownload.sh",
|
||||
"tag": "loginnode",
|
||||
"sudo": false
|
||||
},
|
||||
{
|
||||
"script": "pbsserver.sh",
|
||||
"copy": [
|
||||
"pbspro_19.1.1.centos7/pbspro-server-19.1.1-0.x86_64.rpm"
|
||||
],
|
||||
"tag": "pbsserver",
|
||||
"sudo": false
|
||||
},
|
||||
{
|
||||
"script": "pbsclient.sh",
|
||||
"args": [
|
||||
"$(<hostlists/tags/pbsserver)"
|
||||
],
|
||||
"copy": [
|
||||
"pbspro_19.1.1.centos7/pbspro-execution-19.1.1-0.x86_64.rpm"
|
||||
],
|
||||
"tag": "pbsclient",
|
||||
"sudo": false
|
||||
}
|
||||
]
|
||||
}
|
||||
|
|
@ -0,0 +1,65 @@
|
|||
# Build a PBS compute cluster
|
||||
|
||||
Visualisation: [config.json](https://azurehpc.azureedge.net/?o=https://raw.githubusercontent.com/Azure/azurehpc/master/examples/simple_hpc_pbs_peer_network/config.json)
|
||||
|
||||
This example will create an HPC cluster ready to run with PBS Pro.
|
||||
|
||||
## Initialise the project
|
||||
|
||||
To start you need to copy this directory and update the `config.json`. Azurehpc provides the `azhpc-init` command that can help here by compying the directory and substituting the unset variables. First run with the `-s` parameter to see which variables need to be set:
|
||||
|
||||
```
|
||||
azhpc-init -c $azhpc_dir/examples/simple_hpc_pbs_peer_network -d simple_hpc_pbs_peer_network -s
|
||||
```
|
||||
|
||||
The variables can be set with the `-v` option where variables are comma separated. The output from the previous command as a starting point. The `-d` option is required and will create a new directory name for you. Please update to whatever `resource_group` you would like to deploy to:
|
||||
|
||||
```
|
||||
azhpc-init -c $azhpc_dir/examples/simple_hpc_pbs_peer_network -d simple_hpc_pbs_peer_network -v resource_group=azurehpc-cluster,peer_network_resource_group=some_name,peer_vnet_name=some_name
|
||||
```
|
||||
|
||||
> Note: You can still update variables even if they are already set. For example, in the command below we change the region to `westus2` and the SKU to `Standard_HC44rs`:
|
||||
|
||||
```
|
||||
azhpc-init -c $azhpc_dir/examples/simple_hpc_pbs_peer_network -d simple_hpc_pbs_peer_network -v location=westus2,vm_type=Standard_HC44rs,resource_group=azhpc-cluster,peer_network_resource_group=some_name,peer_vnet_name=some_name
|
||||
```
|
||||
|
||||
## Create the cluster
|
||||
|
||||
```
|
||||
cd simple_hpc_pbs_peer_network
|
||||
azhpc-build
|
||||
```
|
||||
|
||||
Allow ~10 minutes for deployment. You are able to view the status VMs being deployed by running `azhpc-status` in another terminal.
|
||||
|
||||
## Log in the cluster
|
||||
|
||||
Connect to the headnode and check PBS and NFS
|
||||
|
||||
```
|
||||
$ azhpc-connect -u hpcuser headnode
|
||||
Fri Jun 28 09:18:04 UTC 2019 : logging in to headnode (via headnode6cfe86.westus2.cloudapp.azure.com)
|
||||
[hpcuser@headnode ~]$ pbsnodes -avS
|
||||
vnode state OS hardware host queue mem ncpus nmics ngpus comment
|
||||
--------------- --------------- -------- -------- --------------- ---------- -------- ------- ------- ------- ---------
|
||||
compuc407000003 free -- -- 10.2.4.8 -- 224gb 60 0 0 --
|
||||
compuc407000002 free -- -- 10.2.4.7 -- 224gb 60 0 0 --
|
||||
[hpcuser@headnode ~]$ sudo exportfs -v
|
||||
/share/apps <world>(sync,wdelay,hide,no_subtree_check,sec=sys,rw,secure,root_squash,no_all_squash)
|
||||
/share/data <world>(sync,wdelay,hide,no_subtree_check,sec=sys,rw,secure,root_squash,no_all_squash)
|
||||
/share/home <world>(sync,wdelay,hide,no_subtree_check,sec=sys,rw,secure,root_squash,no_all_squash)
|
||||
/mnt/resource/scratch
|
||||
<world>(sync,wdelay,hide,no_subtree_check,sec=sys,rw,secure,root_squash,no_all_squash)
|
||||
[hpcuser@headnode ~]$
|
||||
```
|
||||
|
||||
To check the state of the cluster you can run the following commands
|
||||
|
||||
```
|
||||
azhpc-connect -u hpcuser headnode
|
||||
qstat -Q
|
||||
pbsnodes -avS
|
||||
df -h
|
||||
```
|
||||
|
||||
Загрузка…
Ссылка в новой задаче