-
Feature Request
-
Resolution: Done
-
Major
-
None
-
Important
-
False
-
False
-
-
8
-
Known Issue
-
-
Done
NOTE: This issue started as a Bug, but was moved to be a Feature because we realized it is functionality that is no currently supported with either os-net-config provider.
Description:
This feature deals with "thin footprint" cases where compute nodes are equipped with very less resource (NIC in this case). In such cases, Customers have to run all traffic (OSP provision, control plane, data plane, ) over shared NIC resource. NIC Partition (GAed feature) comes in rescue here. Customer can take advantage of NIC Partition to partition NIC to isolate each traffic over created VFs.
However, challenge here is, NIC ports would be used for Provisioning as well (Which is before NIC Part come into the picture of deployment) and its provision network config needs to remain same same time running on newly created VFs. This ticket covers this case. A successful delivery of this mandate below "Acceptance Criteria"
1) After successful deployment, compute node should have provision traffic running over bond (active-backup) with member interfaces having VF from each Port
2) If PIXe/Provision traffic runs over vlan, make sure same VLAN configured over bond after deployment
3) If customer has configured qos values for VF, they must be configured accodingly on the VFs
4) Provide appropriate failure logs in os-net-config logging system.
5) This feature should support both pre-provisioned and un-provisioned node
6) This feature is applicable only for "Green field" deployments. No support for upgrade from 17.1.
7) All persistent cases (reboot, update, link redundancy) should pass
Definition Done:
1) Document all the limitations encountered while implementing/testing this feature
2) Document "config guide" for this feature (along with topology if possible)
3) All QE test should have been automated, executed and passed in CI
Got this error when deploying openstackdataplanenodeset.dataplane.openstack.org/openstack-edpm:
[zuul@controller-0 ~]$ oc get openstackdataplanenodeset.dataplane.openstack.org/openstack-edpm
NAME STATUS MESSAGE
openstack-edpm False Deployment error occurred in configure-network service error backoff limit reached for execution.name configure-network-edpm-deployment-openstack-edpm execution.namespace openstack execution.status.jobstatus: Failed
[zuul@controller-0 ~]$ oc get pod |grep configure-net
configure-network-edpm-deployment-openstack-edpm-b4wl6 0/1 Error 0 13m
configure-network-edpm-deployment-openstack-edpm-jtrfr 0/1 Error 0 12m
configure-network-edpm-deployment-openstack-edpm-lr8sp 0/1 Error 0 29m
configure-network-edpm-deployment-openstack-edpm-mtktc 0/1 Error 0 13m
configure-network-edpm-deployment-openstack-edpm-q7b4v 0/1 Error 0 2m46s
configure-network-edpm-deployment-openstack-edpm-qphmp 0/1 Error 0 8m13s
configure-network-edpm-deployment-openstack-edpm-rfr8h 0/1 Error 0 11m
Identity added: /runner/artifacts/configure-network-edpm-deployment-openstack-edpm/ssh_key_data (cifmw_reproducer_key)
[WARNING]: Collection osp.edpm does not support Ansible version 2.15.12
[WARNING]: Invalid characters were found in group names but not replaced, use
-vvvv to see detailsPLAY [Deploy EDPM Network] *****************************************************TASK [Gathering Facts] *********************************************************
fatal: [compute-1]: UNREACHABLE! => {"changed": false, "msg": "Failed to connect to the host via ssh: ssh: connect to host compute-1 port 22: No route to host", "unreachable": true}
ok: [compute-0]NO MORE HOSTS LEFT *************************************************************NO MORE HOSTS LEFT *************************************************************PLAY RECAP *********************************************************************
compute-0 : ok=1 changed=0 unreachable=0 failed=0 skipped=0 rescued=0 ignored=0
compute-1 : ok=0 changed=0 unreachable=1 failed=0 skipped=0 rescued=0 ignored=0
compute-0 is accessible via control plane network but compute-1 is not:
$ ssh -i /tmp/k -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null cloud-admin@192.168.122.101 Warning: Permanently added '192.168.122.101' (ED25519) to the list of known hosts. Register this system with Red Hat Insights: insights-client --register Create an account or view all your systems at https://red.ht/insights-dashboard Last login: Tue Jul 23 09:51:51 2024 from 192.168.122.30 [cloud-admin@compute-0 ~]$ [cloud-admin@compute-0 ~]$ ip a sh ens3f0np0 8: ens3f0np0: <BROADCAST,MULTICAST,PROMISC,UP,LOWER_UP> mtu 9000 qdisc mq state UP group default qlen 1000 link/ether 0c:42:a1:5d:a0:14 brd ff:ff:ff:ff:ff:ff altname enp65s0f0np0 inet 192.168.122.101/24 brd 192.168.122.255 scope global noprefixroute ens3f0np0 valid_lft forever preferred_lft forever inet6 fe80::e42:a1ff:fe5d:a014/64 scope link valid_lft forever preferred_lft forever [cloud-admin@compute-0 ~]$ ip r default via 192.168.122.1 dev ens3f0np0 proto static metric 150 172.23.0.0/24 dev eno34 proto kernel scope link src 172.23.0.64 metric 139 192.168.122.0/24 dev ens3f0np0 proto kernel scope link src 192.168.122.101 metric 150 $ ssh -i /tmp/k -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null cloud-admin@192.168.122.100 ssh: connect to host 192.168.122.100 port 22: No route to host
This error present in /var/log/os-net-config.log:
[cloud-admin@compute-0 ~]$ grep ERROR /var/log/os-net-config.log 2024-07-23 09:26:52.498 ERROR os_net_config.impl_ifcfg.apply Failure(s) occurred when applying configuration 2024-07-23 09:26:52.498 ERROR os_net_config.impl_ifcfg.apply stdout: ERROR : [/etc/sysconfig/network-scripts/ifup-eth] Error, some other host (0C:42:A1:5D:A0:14) already uses address 192.168.122.101.
The equivalent thing can be seen on compute-1 but for 192.168.122.100 (via iDRAC Virtual Console).
- mentioned in
-
Page Loading...
