| Summary: | [ocp-on-osp] /etc/resolv.conf was reset after rebooting the instances on RHEL-7.3 | ||
|---|---|---|---|
| Product: | OpenShift Container Platform | Reporter: | Gan Huang <ghuang> |
| Component: | Reference Architecture | Assignee: | scollier |
| Status: | CLOSED WONTFIX | QA Contact: | Gan Huang <ghuang> |
| Severity: | high | Docs Contact: | |
| Priority: | high | ||
| Version: | 3.4.0 | CC: | aos-bugs, jokerman, mmccomas, tsedovic |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | If docs needed, set a value | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2018-02-22 14:44:37 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
"skip_dns: true" also can reproduce this issue. "skip_dns: true" will install and config dnsmqsq in OpenShift cluster, thus it will trigger to restart NetworkManager during installing OpenShift, and /etc/resolv.conf would be reset as well. this might be related to docker version: docker-1.10.3-46.el7.10.x86_64 (rhel72 - works) root@test-ocp34-master-0: ~ # nmcli con show docker0|grep -i dns ipv4.dns: ipv4.dns-search: ipv4.ignore-auto-dns: no ipv6.dns: ipv6.dns-search: ipv6.ignore-auto-dns: no docker-selinux-1.10.3-57.el7.x86_64 (rhel73 - networkmanager overwrites resolv.conf) [root@bug-openshift-node-15ok67md etc]# nmcli con show 5fb06bd0-0bb0-7ffb-45f1-d6edd65f3e03|grep -i dns ipv4.dns: ipv4.dns-search: ipv4.dns-options: (default) ipv4.dns-priority: 0 ipv4.ignore-auto-dns: yes ipv6.dns: ipv6.dns-search: ipv6.dns-options: (default) ipv6.dns-priority: 0 ipv6.ignore-auto-dns: no It's unrelated to docker interface. We set "PEERDNS=no" for all interfaces but it seems that for new NetworkManager its meaning is ""never add automatic (DHCP, PPP, VPN, etc.) nameservers to resolv.conf" instead of the expected "never touch resolv.conf". Putting "dns=none" into [main] section of /etc/NetworkManager/NetworkManager.conf has desired effect. This is related to NetworkManager indeed. #RHEL-7.2 /etc/resolv.conf won't be overided in below version: NetworkManager-1.0.6-27.el7.x86_64 NetworkManager-libnm-1.0.6-27.el7.x86_64 NetworkManager-team-1.0.6-27.el7.x86_64 NetworkManager-config-server-1.0.6-27.el7.x86_64 NetworkManager-tui-1.0.6-27.el7.x86_64 /etc/resolv.conf would be overided in below version: NetworkManager-tui-1.0.6-31.el7_2.x86_64 NetworkManager-1.0.6-31.el7_2.x86_64 NetworkManager-team-1.0.6-31.el7_2.x86_64 NetworkManager-config-server-1.0.6-31.el7_2.x86_64 NetworkManager-libnm-1.0.6-31.el7_2.x86_64 #RHEL-7.3 /etc/resolv.conf would be overided in RHEL-7.3 NetworkManager-1.4.0-12.el7.x86_64 NetworkManager-team-1.4.0-12.el7.x86_64 NetworkManager-libnm-1.4.0-12.el7.x86_64 NetworkManager-tui-1.4.0-12.el7.x86_64 NetworkManager-config-server-1.4.0-12.el7.x86_64 Team, Closing this as the heat templates have been deprecated. Future funcionality and integration capabilities will be moving to openshift-ansible moving forward. |
Description of problem: Creating a heat stack which specify "volume_quota: 3" in env file with RHEL-7.3 hosts, that option would result in rebooting the instances during creating the stack. Then the stack would failed due to the dnsname server of those instances were missing because of the reboot. Version-Release number of selected component (if applicable): openshift-on-openstack-0.9.5-1.el7.centos.noarch How reproducible: always Steps to Reproduce: 1. Specify "volume_quota: 3" and RHEL-7.3 image in env file 2. Creat a stack 3. Actual results: Stack failed when running the ansible playbook: Ansible log on bastion host: "msg": "http://xxx.redhat.com/rhel/server/7/7Server/x86_64/os/rhel-7-server-rpms/Packages/iptables-services-1.4.21-17.el7.x86_64.rpm: [Errno 14] curl#6 - \"Could not resolve host: xxx.redhat.com; Unknown error\"\nTrying other mirror.\n\n\nError downloading packages:\n iptables-services-1.4.21-17.el7.x86_64: [Errno 256] No more mirrors to try.\n\n", "rc": 1, "results": [ "iptables-1.4.21-17.el7.x86_64 providing iptables is already installed", "Loaded plugins: search-disabled-repos\nResolving Dependencies\n--> Running transaction check\n---> Package iptables-services.x86_64 0:1.4.21-17.el7 will be installed\n--> Finished Dependency Resolution\n\nDependencies Resolved\n\n================================================================================\n Package Arch Version Repository Size\n================================================================================\nInstalling:\n iptables-services x86_64 1.4.21-17.el7 rhel7 50 k\n\nTransaction Summary\n================================================================================\nInstall 1 Package\n\nTotal download size: 50 k\nInstalled size: 24 k\nDownloading packages:\n" ] } NO MORE HOSTS LEFT ************************************************************* to retry, use: --limit @/var/lib/ansible/playbooks/main.retry PLAY RECAP ********************************************************************* flannel-73-openshift-infra-0.example.com : ok=61 changed=7 unreachable=0 failed=0 flannel-73-openshift-master-0.example.com : ok=74 changed=10 unreachable=0 failed=1 flannel-73-openshift-node-h7kt27x6.example.com : ok=61 changed=7 unreachable=0 failed=0 localhost : ok=21 changed=15 unreachable=0 failed=0 Check on flannel-73-openshift-master-0.example.com, nameservers were missing. $ cat /etc/resolv.conf # Generated by NetworkManager search example.com Expected results: Additional info: