CentOS7/8 image is broken now

Hi, guys,
Today i am trying to deploy centos 7(download inside MAAS) on my node, but the centos images stuck after install, during the reboot:

Same image has successful deploy a few days ago, so I think that is because images auto update?
Please help, thanks!

Could you grab the deployment logs and file a bug request at https://bugs.launchpad.net/maas, we’re trying to minimise the number of channels in use.
Can you additionally verify the version you deployed when it worked vs when it didn’t, might be able to pin it down to your suspicion of an image update

1 Like

Thank you for your reply.
I found this problem is related to the subnet.
I created 2 subnets, the first one is where the maas is located, and the second one is in a different subnet, and I assigned IPs for MAAS in both subnets.
So the problem is, if I deploy centos for the machine in subnet 1, everything is fine, but it is not work in subnet 2.

Did I miss anything?

From screenshot, it seems machine cant get ip address from cloud-init after installation.

@raindrop, what do you mean:

I’m sorry I didn’t explain it clearly. I will provide my subnet settings in the screenshot below.
If the machine is connected to the first subnet shown in the screenshot, you can deploy both Ubuntu and CentOS. If the machine is connected to the second subnet shown in the picture, you can still deploy Ubuntu, but you cannot deploy CentOS. It will get stuck in the same place as the picture I sent at first.

Hiya raindrop,

Thank you for providing additional information about your deployment issue. Based on your explanation and the screenshots you shared, it seems that the problem is related to the subnets you have configured in MAAS.

From your description, it appears that you have two subnets set up in MAAS: one where MAAS is located (subnet 1) and another subnet (subnet 2) that is in a different network. Additionally, you mentioned that you have assigned IP addresses for MAAS in both subnets.

The issue you are experiencing is that when you deploy CentOS on a machine connected to subnet 2, the deployment process gets stuck during the reboot and the machine is unable to obtain an IP address from cloud-init.

To troubleshoot this further, here are a few suggestions:

  1. Verify your subnet configurations: Double-check the settings and configurations for subnet 2, including the network range, DHCP configuration, gateway settings, and any relevant firewall rules. Ensure that the subnet settings are correctly configured and aligned with the network infrastructure.
  2. Check DHCP and DNS: It’s always DNS! Make sure that DHCP and DNS services are properly configured for subnet 2. Ensure that the DHCP server is running and capable of assigning IP addresses to machines in that subnet. Do some packet captures if you have to, just to be sure DORA is happening on that subnet. Also confirm that the DNS resolution is functioning correctly for the machines in subnet 2. You can use the usual litany of command-line networking tools to do this.
  3. Network connectivity: Probably not an issue, but do make sure that there is proper network connectivity between the MAAS server and the machines in subnet 2. Check for any potential network issues, especially firewall rules blocking communication or misconfigured routing.
  4. Debug cloud-init: Since the issue occurs during the cloud-init phase after installation, you can try to debug cloud-init on the deployed CentOS machine. Check the cloud-init logs (/var/log/cloud-init.log) and cloud-init configuration files for any errors or misconfigurations that might prevent the machine from obtaining an IP address.
  5. MAAS troubleshooting: Consider reviewing the MAAS logs for any related errors or warnings. The MAAS logs can provide unexpected insight into the deployment process and help identify any potential issues. I think you know where to find those, but if not, bounce back.

If the issue persists or you need further assistance, we’re here, ready to help you resolve the problem on a best-effort basis. Good luck!