Hi Ace, Can you verify that ipmitool works first on the jumphost to access every node? If so, then validate the same thing on the undercloud VM. That will rule out any connectivity issues between Undercloud Ironic and the IPMI access to each node. The errors still seem to show a problem with detecting power state of the nodes.
Thanks, Tim Rozet Red Hat SDN Team ----- Original Message ----- From: "liyin (F)" <liyi...@huawei.com> To: "Tim Rozet" <tro...@redhat.com> Cc: opnfv-tech-discuss@lists.opnfv.org Sent: Tuesday, January 10, 2017 3:56:58 AM Subject: RE: [opnfv-tech-discuss] Apex bare metel deploy problem Hi Tim, I have confirmed that IPMI is indeed not connected from our jumphost to the compute node. I have searched for the reasons. Finally I find out that it is because our ipmi switchboard is connected to external switchboard. To solve this problem I have connected the jumphost with ipmi switchboard and external switchboard. Then I use the iso of artifacts: opnfv-2016-12-21.iso to deploy apex. the results and log are shown as attached: 1) Ipmi_nova_list.png; 2) Ipmi_openstack_failure*.png (output of openstack stack failures list overcloud --long); 3) apex_log.txt The deployment of overcloud node still failed while the connection between jumphost and compute node is success. For more information please refer to the attachment. For the reasons of the unsuccessful deployment, my guesses are: 1)Those nodes have errors in network configuration. 2)The network_settings.yaml have some errors I wonder if my guesses are correct? Could you please provide me some solutions? ________________________________ Some information of the network configuration is provided as follows. 1.Attached my network configuration file: network_settings_normal.yaml 2. Jumphost informations could be found below [cid:image003.jpg@01D26B62.8A3DF910] NIC name IP switch Have external net access enp2s0f0 192.168.36.2 External switch yes enp2s0f1 10.10.10.2 External switch no The other nodes network configuration NIC name IP switch Have external net access enp2s0f0 no External switch * enp2s0f1 no External switch * The other nodes don’t have operation system, only two NICs connect to external switch. Both jumphost and nodes are connected with ipmi switch. ________________________________ There are some other issues during the deployment: During the deployment of apex, I have no access from external network to jumphost. After br-admin and br-external have bridged to NIC. I have access from external network to jumphost. But when the log shown: Executing overcloud deployment, this should run for an extended period without output. I couldn't connect to jumphost with external network. ________________________________ Thanks a lot, And waiting for your reply. Best Regards, Ace. -----Original Message----- From: Tim Rozet [mailto:tro...@redhat.com] Sent: Monday, January 09, 2017 11:48 PM To: liyin (F) <liyi...@huawei.com> Cc: opnfv-tech-discuss@lists.opnfv.org Subject: Re: [opnfv-tech-discuss] Apex bare metel deploy problem It looks like the problem might be IPMI connectivity from your jumphost to at least that compute node. Can you try from your jumphost issuing ipmitool cmdline to make sure you can connect to them? For example: ipmitool -I lanplus -H <host ip> -L ADMINISTRATOR -U <username> -P <password> power status Tim Rozet Red Hat SDN Team ----- Original Message ----- From: "liyin (F)" <liyi...@huawei.com> To: "Tim Rozet" <tro...@redhat.com> Cc: opnfv-tech-discuss@lists.opnfv.org Sent: Friday, January 6, 2017 10:50:35 PM Subject: RE: [opnfv-tech-discuss] Apex bare metel deploy problem Hi Tim, I could only get connect the jumphost by ipmi , so I only could provide you some picture . I think it's also a problem during deplovement, I have no access to this jumphost. By the way, this iso is master and the date is 2016.12.21. Stack_list.png is the output of 3. Nova_list.png is the output of 4. Thank you for you kindness. -----Original Message----- From: Tim Rozet [mailto:tro...@redhat.com] Sent: Friday, January 06, 2017 9:01 AM To: liyin (F) <liyi...@huawei.com> Cc: opnfv-tech-discuss@lists.opnfv.org Subject: Re: [opnfv-tech-discuss] Apex bare metel deploy problem Hi Ace, Can you please on your jumphost do: 1. opnfv-util undercloud 2. . stackrc 3. openstack stack failures list overcloud --long 4. nova list Please send me the output of 3 and 4. Thanks, Tim Rozet Red Hat SDN Team ----- Original Message ----- From: "liyin (F)" <liyi...@huawei.com> To: opnfv-tech-discuss@lists.opnfv.org Sent: Tuesday, December 27, 2016 3:41:57 AM Subject: [opnfv-tech-discuss] Apex bare metel deploy problem Hi all, We have an environment of bare metal pods. And we want to use apex to deploy openstack. I use the Centos iso from apex artifacts site to install jump server system. I have used several iso to deploy the environment and I get the same result as appendix showing. This log can’t help me to find where the problem is. And another thing is when I use opnfv-deploy os-nosdn-nofeature-ha.yaml to deploy, it will cost a lot of time. This puzzled me a lot, I need your help. Thanks in advance. Best Regards, Ace. _______________________________________________ opnfv-tech-discuss mailing list opnfv-tech-discuss@lists.opnfv.org https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss _______________________________________________ opnfv-tech-discuss mailing list opnfv-tech-discuss@lists.opnfv.org https://lists.opnfv.org/mailman/listinfo/opnfv-tech-discuss