[devstack][neutron] DNS Resolution Issue After Server Reboot with DevStack
Hi everyone, I am reaching out to seek your assistance regarding an issue I am experiencing with my DevStack installation. After successfully installing DevStack stable/2024.2 on ubuntu 22.04, I configured the network and set up DNS for private subnet. stack@controller:~$ openstack subnet list +--------------------------------------+---------------------+--------------------------------------+---------------------+ | ID | Name | Network | Subnet | +--------------------------------------+---------------------+--------------------------------------+---------------------+ | 2b6411d5-9c19-4a45-bd47-58dc748f5e00 | public-subnet | 45502b6a-773b-40cf-866a-e8b18397ffcb | 192.168.1.0/24 | | 69629170-2e0e-49c0-b741-61560558001c | ipv6-public-subnet | 45502b6a-773b-40cf-866a-e8b18397ffcb | 2001:db8::/64 | | c74bf062-b67a-45cb-8a50-73583498dfcd | ipv6-private-subnet | 8fb37e61-bd59-4624-a4be-1fc1841987d7 | fd3a:2f86:ca98::/64 | | f2e10d3d-f785-4cf9-8804-1969269623c2 | private-subnet | 8fb37e61-bd59-4624-a4be-1fc1841987d7 | 10.0.0.0/26 | +--------------------------------------+---------------------+--------------------------------------+---------------------+ stack@controller:~$ openstack subnet set --dns-nameserver 10.12.255.254 --dns-nameserver 192.168.1.95 f2e10d3d-f785-4cf9-8804-1969269623c2 10.12.255.254 is the real DNS server address and 192.168.1.95 is my server IP address. Initially, when I launched a VM instance, it was able to resolve the proxy and connect to the external network without any issues. jenkins@np0000000531:~$ curl opendev.org <a href="https://opendev.org/">Temporary Redirect</a>. However, after rebooting the server and launching the instance again, the instance is unable to resolve the proxy. jenkins@np0000000531:~$ curl opendev.org curl: (5) Could not resolve proxy: my.proxy.com I have verified that the systemd-resolved service is in an active state, but the logs show a large number of entries with the following message: Nov 13 11:53:51 np0000000531 systemd-resolved[568]: Using degraded feature set UDP instead of TCP for DNS server 10.12.255.254. Nov 13 11:54:01 np0000000531 systemd-resolved[568]: Using degraded feature set UDP instead of TCP for DNS server 192.168.1.95. Nov 13 11:54:01 np0000000531 systemd-resolved[568]: Using degraded feature set TCP instead of UDP for DNS server 10.12.255.254. Nov 13 11:54:11 np0000000531 systemd-resolved[568]: Using degraded feature set TCP instead of UDP for DNS server 192.168.1.95. Nov 13 11:54:21 np0000000531 systemd-resolved[568]: Using degraded feature set UDP instead of TCP for DNS server 10.12.255.254. I have checked the network and routing table of the VM instance, as well as the bridge settings on the server. Everything appears to be the same as it was before the reboot. Could you please help me identify the cause of this issue and suggest any potential solutions? If you need any additional information, please let me know, and I will be happy to provide it. Thank you for your time and assistance. Best regards, Inori.
Hi everyone, I reinstalled devstack by running commands "./unstack.sh" and "./stack.sh", and applied the same network configurations. When I launch an instance, it was able to resolve the proxy and connect to the external network. Upon investigation, I discovered that in previous versions of devstack, there was a rejoin-stack.sh script that could be used to restart services after a reboot. However, it seems that this script has been removed in the later versions. Given that reinstalling devstack and configuring the necessary services is quite time-consuming, I would like to inquire if there is an alternative method available in the 2024.2 version that could replace the need for a complete reinstallation. Thanks and Regards, Inori
On 14/11/2024 09:04, xuq.fnstxz@fujitsu.com wrote:
Hi everyone,
I reinstalled devstack by running commands "./unstack.sh" and "./stack.sh", and applied the same network configurations.
When I launch an instance, it was able to resolve the proxy and connect to the external network.
Upon investigation, I discovered that in previous versions of devstack, there was a rejoin-stack.sh script that could be used to restart services after a reboot. However, it seems that this script has been removed in the later versions.
Given that reinstalling devstack and configuring the necessary services is quite time-consuming, I would like to inquire if there is an alternative method available in the 2024.2 version that could replace the need for a complete reinstallation.
so it was removed because it should nto be required anymore. you shoudl be able to just reboot the vms and everything should come back up now that we use systemd unit files instead of screen to run openstack services. as this is not tested i ci this occasionally breaks but it was working on ubuntu at least in the last 12 months i fixed it on fedora/centos a while back. my guess is you hit a bug/regression like this https://github.com/openstack/devstack/commit/80c3ffe154fd79e03d8c4258b500b77... the first thing i woudl try is restartign all the devstack servvices after teh reboot `sudo systemctl restart devstack@*` sometimes ordering can cause it to fail. Often the issue is with apache or simialr startign before rabbit or mariadb which causes keystone or similar not to start properly.
Thanks and Regards, Inori
Hi Sean, Sorry for the delay in my reply. And Thanks for your suggestions on my problem. I have carefully followed the steps you outlined, but unfortunately, the issue persists. After I reboot the devstack server and run `sudo systemctl restart devstack@*`, the nova instance still can not the my proxy server. I still need to uninstall and re-install the devstack and set the dns to solve this problem. Could you please advise on any further steps I might take or if there are alternative solutions we could explore? If you need any additional information, please let me know, and I will be happy to provide it. Thanks and Regards, Inori.
participants (2)
-
Sean Mooney
-
xuq.fnstxz＠fujitsu.com