MTL-1980 - Configure a bonded HSN connection on an NCN

[spillerc-hpe]

Summary and Scope

Ansible role to configure a bonded HSN interface on NCNs.

This role is not enabled by default as it requires values that have to be supplied by the Slingshot administrator.

Issues and Related PRs

• Resolves MTL-1980
• Documentation changes to follow in MTL-1981

Testing

Tested on:

• surtur

Test description:

Rebuilt ncn-w005 with Playbook enabled with the following settings

hsn_bond_mac: "b2:00:00:00:00:01"
hsn_bond_ip: "10.253.254.1"
hsn_bond_netmask: '255.255.0.0'
hsn_bond_enable: true

Verified bond was configured as expected.

Running ncn_hsn_bonding.yml from repo https://api-gw-service-nmn.local/vcs/cray/csm-config-management.git
[WARNING]: Could not match supplied host pattern, ignoring: cfs_image

PLAY [Configure HSN bonding] ***************************************************

TASK [csm.ncn.hsn_bonding : Create routing table] ******************************
changed: [x3000c0s31b0n0]

TASK [csm.ncn.hsn_bonding : Create macvlan interface definitions] **************
changed: [x3000c0s31b0n0] => (item=['hsn0', 'macvlan0'])
changed: [x3000c0s31b0n0] => (item=['hsn1', 'macvlan1'])

TASK [csm.ncn.hsn_bonding : Create bond interface] *****************************
changed: [x3000c0s31b0n0]

TASK [csm.ncn.hsn_bonding : Set bond sysctl values] ****************************
changed: [x3000c0s31b0n0] => (item={'name': 'net.ipv4.conf.bond1.accept_local', 'value': 1})
changed: [x3000c0s31b0n0] => (item={'name': 'net.ipv4.conf.bond1.arp_filter', 'value': 1})
changed: [x3000c0s31b0n0] => (item={'name': 'net.ipv4.conf.bond1.arp_announce', 'value': 2})
changed: [x3000c0s31b0n0] => (item={'name': 'net.ipv4.conf.bond1.arp_ignore', 'value': 1})
changed: [x3000c0s31b0n0] => (item={'name': 'net.ipv4.conf.bond1.arp_notify', 'value': 0})
changed: [x3000c0s31b0n0] => (item={'name': 'net.ipv4.conf.bond1.arp_accept', 'value': 1})
changed: [x3000c0s31b0n0] => (item={'name': 'net.ipv4.conf.bond1.drop_gratuitous_arp', 'value': 1})
changed: [x3000c0s31b0n0] => (item={'name': 'net.ipv4.conf.bond1.rp_filter', 'value': 0})

TASK [csm.ncn.hsn_bonding : Set arp_filter sysctl for HSN NICs] ****************
changed: [x3000c0s31b0n0] => (item=hsn0)
changed: [x3000c0s31b0n0] => (item=hsn1)

TASK [csm.ncn.hsn_bonding : Set arp_ignore sysctl for HSN NICs] ****************
changed: [x3000c0s31b0n0] => (item=hsn0)
changed: [x3000c0s31b0n0] => (item=hsn1)

TASK [csm.ncn.hsn_bonding : Generate bond1 post-up script] *********************
changed: [x3000c0s31b0n0]
FAILED - RETRYING: Reload interfaces (3 retries left).

RUNNING HANDLER [csm.ncn.hsn_bonding : Reload interfaces] **********************
changed: [x3000c0s31b0n0]

PLAY RECAP *********************************************************************
x3000c0s31b0n0             : ok=11   changed=8    unreachable=0    failed=0    skipped=2    rescued=0    ignored=0

Rebooted ncn-w005, verified bond came up before CFS ran and that the Ansible playbook didn't perform any unnecessary configuration actions.

Running ncn_hsn_bonding.yml from repo https://api-gw-service-nmn.local/vcs/cray/csm-config-management.git
[WARNING]: Could not match supplied host pattern, ignoring: cfs_image

PLAY [Configure HSN bonding] ***************************************************

PLAY RECAP *********************************************************************
x3000c0s31b0n0             : ok=10   changed=0    unreachable=0    failed=0    skipped=2    rescued=0    ignored=0

Risks and Mitigations

• Role is not enabled out of the box, it must be explicitly enabled.

Pull Request Checklist

• Version number(s) incremented, if applicable
• Copyrights updated
• License file intact
• Target branch correct
• CHANGELOG.md updated
• Testing is appropriate and complete, if applicable
• HPC Product Announcement prepared, if applicable