Thursday, December 4, 2025

OpenStack Epoxy : Add Compute Nodes (GPU)

 

OpenStack Epoxy : Add Compute Nodes (GPU)

 

Add GPU attached Compute Nodes to use GPU on virtual machine instances.

On this example, add [node02.srv.world] like follows.

------------+--------------------------+--------------------------+------------
            |                          |                          |
        eth0|10.0.0.30             eth0|10.0.0.50             eth0|10.0.0.51
+-----------+-----------+  +-----------+-----------+  +-----------+-----------+
|   [ dlp.srv.world ]   |  | [ network.srv.world ] |  |  [ node01.srv.world ] |
|     (Control Node)    |  |     (Network Node)    |  |     (Compute Node)    |
|                       |  |                       |  |                       |
|  MariaDB    RabbitMQ  |  |      Open vSwitch     |  |        Libvirt        |
|  Memcached  Nginx     |  |     Neutron Server    |  |      Nova Compute     |
|  Keystone   httpd     |  |      OVN-Northd       |  |      Open vSwitch     |
|  Glance     Nova API  |  |         Nginx         |  |   OVN Metadata Agent  |
|                       |  |                       |  |     OVN-Controller    |
+-----------------------+  +-----------------------+  +-----------------------+

------------+------------
            |
        eth0|10.0.0.52
+-----------+-----------+
|  [ node02.srv.world ] |
|  (Compute Node (GPU)) |
|                       |
|        Libvirt        |
|      Nova Compute     |
|      Open vSwitch     |
|   OVN Metadata Agent  |
|     OVN-Controller    |
+-----------------------+

[1]

On the new Node with GPU, Configure GPU Passthrough setting, refer to here for [1].

[2]

On the new Node with GPU, add it to Openstack Cluster as a Compute Node, refer to here.

[3]On the new Node with GPU, Configure additional settings for Nova-Compute.
root@node02:~# 
lspci -nn | grep -i nvidia

81:00.0 VGA compatible controller [0300]: NVIDIA Corporation GA104 [GeForce RTX 3060] [10de:2487] (rev a1)
81:00.1 Audio device [0403]: NVIDIA Corporation GA104 High Definition Audio Controller [10de:228b] (rev a1)

root@node02:~# 
vi /etc/nova/nova.conf
# add to last line
# add [vendor_id], [product_id] of passthrough device
[pci]
passthrough_whitelist = { "vendor_id": "10de", "product_id": "2487" }

root@node02:~# 
systemctl restart nova-compute
[4]On Control Node, Change Nova settings.
root@dlp ~(keystone)# 
vi /etc/nova/nova.conf
# add to last line
# add [vendor_id], [product_id] that are the same one on compute node
# for the [name], set any name you like
[pci]
alias: { "vendor_id":"10de", "product_id":"2487", "device_type":"type-PCI", "name":"RTX-3060" }

[filter_scheduler]
enabled_filters = PciPassthroughFilter

root@dlp ~(keystone)# 
systemctl restart nova-api nova-scheduler

root@dlp ~(keystone)# 
su -s /bin/bash nova -c "nova-manage cell_v2 discover_hosts"
# create [flavor] for GPU instance

root@dlp ~(keystone)# 
openstack flavor create --id 6 --vcpus 4 --ram 8192 --disk 20 --property "pci_passthrough:alias"="RTX-3060:1" gpu1.small

+----------------------------+------------------------------------+
| Field                      | Value                              |
+----------------------------+------------------------------------+
| OS-FLV-DISABLED:disabled   | False                              |
| OS-FLV-EXT-DATA:ephemeral  | 0                                  |
| description                | None                               |
| disk                       | 20                                 |
| id                         | 6                                  |
| name                       | gpu1.small                         |
| os-flavor-access:is_public | True                               |
| properties                 | pci_passthrough:alias='RTX-3060:1' |
| ram                        | 8192                               |
| rxtx_factor                | 1.0                                |
| swap                       | 0                                  |
| vcpus                      | 4                                  |
+----------------------------+------------------------------------+

root@dlp ~(keystone)# 
openstack flavor list

+----+------------+-------+------+-----------+-------+-----------+
| ID | Name       |   RAM | Disk | Ephemeral | VCPUs | Is Public |
+----+------------+-------+------+-----------+-------+-----------+
| 1  | m1.tiny    |  2048 |   10 |         0 |     1 | True      |
| 2  | m1.small   |  4096 |   10 |         0 |     2 | True      |
| 3  | m1.medium  |  8192 |   10 |         0 |     4 | True      |
| 4  | m1.large   | 16384 |   10 |         0 |     8 | True      |
| 5  | m2.medium  |  8192 |   10 |        10 |     4 | True      |
| 6  | gpu1.small |  8192 |   20 |         0 |     4 | True      |
+----+------------+-------+------+-----------+-------+-----------+
[5]Verify settings to create a GPU instance with any Openstack user.
debian@dlp ~(keystone)$ 
openstack image list

+--------------------------------------+----------+--------+
| ID                                   | Name     | Status |
+--------------------------------------+----------+--------+
| 33def798-3361-483b-9de1-8c2c6e1c840c | Debian13 | active |
+--------------------------------------+----------+--------+

debian@dlp ~(keystone)$ 
openstack network list

+--------------------------------------+---------+--------------------------------------+
| ID                                   | Name    | Subnets                              |
+--------------------------------------+---------+--------------------------------------+
| 7de3878f-814f-4909-b4e6-d4dc1c740577 | public  | 319013fd-5412-4cce-bb87-49f5a0c91b0e |
| d442015a-b6f6-4349-890b-c08eb5366a4d | private | bb5efd0a-ea4d-42ee-99ca-97cee2f56ca2 |
+--------------------------------------+---------+--------------------------------------+

debian@dlp ~(keystone)$ 
netID=$(openstack network list | grep private | awk '{ print $2 }')

debian@dlp ~(keystone)$ 
openstack server create --flavor gpu1.small --image Debian13 --security-group secgroup01 --nic net-id=$netID --key-name mykey Debian-13GPU
debian@dlp ~(keystone)$ 
openstack server list

+--------------------------------------+--------------+---------+------------------------------------+----------+------------+
| ID                                   | Name         | Status  | Networks                           | Image    | Flavor     |
+--------------------------------------+--------------+---------+------------------------------------+----------+------------+
| 8b270701-c516-4f7e-979b-53d7186a6e4d | Debian-13GPU | ACTIVE  | private=192.168.100.97             | Debian13 | gpu1.small |
| 2deabceb-e220-4e5c-b8dd-320683d0bf0b | Debian-13    | SHUTOFF | private=10.0.0.249, 192.168.100.39 | Debian13 | m1.small   |
+--------------------------------------+--------------+---------+------------------------------------+----------+------------+

debian@dlp ~(keystone)$ 
openstack floating ip create public

+---------------------+--------------------------------------+
| Field               | Value                                |
+---------------------+--------------------------------------+
| created_at          | 2025-08-29T05:06:04Z                 |
| description         |                                      |
| dns_domain          |                                      |
| dns_name            |                                      |
| fixed_ip_address    | None                                 |
| floating_ip_address | 10.0.0.251                           |
| floating_network_id | 7de3878f-814f-4909-b4e6-d4dc1c740577 |
| id                  | ec9e221a-68a6-4ce7-a149-47cf4f62eaef |
| name                | 10.0.0.251                           |
| port_details        | None                                 |
| port_id             | None                                 |
| project_id          | ecfa98ba82de421e8f16c3d862b5ab04     |
| qos_policy_id       | None                                 |
| revision_number     | 0                                    |
| router_id           | None                                 |
| status              | DOWN                                 |
| subnet_id           | None                                 |
| tags                | []                                   |
| updated_at          | 2025-08-29T05:06:04Z                 |
+---------------------+--------------------------------------+

debian@dlp ~(keystone)$ 
openstack server add floating ip Debian-13GPU 10.0.0.251

debian@dlp ~(keystone)$ 
ssh debian@10.0.0.251

The authenticity of host '10.0.0.251 (10.0.0.251)' can't be established.
ED25519 key fingerprint is SHA256:Vl3vZ+KUqZ0fiuiDUR7gsOQ1fnyCFAsQ833tz59TVQY.
This key is not known by any other names.
Are you sure you want to continue connecting (yes/no/[fingerprint])? yes
Warning: Permanently added '10.0.0.251' (ED25519) to the list of known hosts.
Linux debian-13gpu 6.12.41+deb13-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.12.41-1 (2025-08-12) x86_64

The programs included with the Debian GNU/Linux system are free software;
the exact distribution terms for each program are described in the
individual files in /usr/share/doc/*/copyright.

Debian GNU/Linux comes with ABSOLUTELY NO WARRANTY, to the extent
permitted by applicable law.
debian@debian-13gpu:~$
debian@debian-13gpu:~$ 
lspci | grep -i nvidia

00:05.0 VGA compatible controller: NVIDIA Corporation GA104 [GeForce RTX 3060] (rev a1)
Matched Content

No comments:

Post a Comment