Pxe boot before comission hanging problem, dns can't resolve problem

We are going to install MAAS 3.3.1 version and use it for bare metal provisioning.

# snap list maas
Name Version Rev Tracking Publisher Notes
maas 3.3.1-13169-g.94920eb1e 26658 - canonical✓ -

During pxe booting at a specific R750 node, we are in a problem that we cannot move on as we are freezing in the following situation.


Booting PXE over IPv4.
Station IP address is 172.20.192.1
Server IP address is 172.20.1.27
NBP filename is bootx64.efi
NBP filesize is 955656 Bytes
Downloading NBP file…
NBP file downloaded successfully.
Fetching Netboot Image
Booting under MAAS direction…


(hanging)

Other nodes operate normally and are registered in the MAAS “Machine” list in a properly ready state. Even nodes of the model, such as nodes being frozen.
Is there a solution to the problem?

In addition, there is another problem that dns resolve is not possible in the machine where deploy is completed.
It seems to be properly resolved only when applying “$ sudo netplan apply” command, but it does not resolve again when rebooting the node.
Is there a way to resolve the dns record properly without additional command after machine deployment?

we use ubuntu 18.04 bionic and 20.04 focal (both controller server and deployed server)

what do you mean by ‘dns resolve is not possible in the machine’ is maas not acting like dns in your case?

After the ubuntu deploy using maas is over, the dns server does not seem to be applied to the deployed server.

assuming that the test1, test2 nodes are deployed,

root@test1:/home/ubuntu# ping test2
ping: bptest18: Temporary failure in name resolution
root@test1:/home/ubuntu# netplan apply
root@test1:/home/ubuntu# ping test2
PING test2.maas (172.20.60.10) 56(84) bytes of data.

This is my situation right now.

is your machine in deployed state? Cause if it is not deployed then you wont be able ping the machines.

yes. test1, test2 both nodes are in deployed state.

ok, can u show me machine log and dns settings in maas

172.20.1.27: MAAS Server
172.20.50.120: test1 Server
172.20.60.10: test2 Server

maas dns setting
–Upstream DNS used to resolve domains not managed by this MAAS: 172.20.1.27
–Enable DNSSEC validation of upstream zones: Automatic (use default root key)
–List of external networks: (empty)

and where could i get machine logs?

thats ok, for upstream you can even use any dns of your choice there, but i wanted to ask what machines are there in resource record.
you can find both in maas ui
dns : maas -> dns -> domain -> there will be machines in that domain
logs : maas -> machine -> select your machine -> logs

1 Like

Domain summary

Name
maas

TTL
(default)

Authoritative
Yes

Resource records

NAME TYPE TTL DATA ACTIONS
maas-server A (default) 172.20.1.27
test1 A (default) 172.20.50.120
test2 A (default) 172.20.60.10

I think it’s hard to show the log because it’s too much. But there are no specific logs or errors related to the network or dns.
If you tell me the information you want to find, I will search it in the log.
Thank you for your help.

@seochokid i have one concern from the logs you posted in the question is,
Booting PXE over IPv4.
Station IP address is 172.20.192.1
Server IP address is 172.20.1.27
NBP filename is bootx64.efi
NBP filesize is 955656 Bytes
Downloading NBP file…
NBP file downloaded successfully.
Fetching Netboot Image
Booting under MAAS direction…
and hanging after that
this usually happen when your deployment fails. In traditional deployment process there are lots of logs after that, So do check that.

Other than that, there is a workaround for DNS. You can add deployment script at the time of deployment.
This link (MaaS DNS issue) might also be useful.

Oh, the log I left in the question is the log for certain nodes that are not deployed. The dns problem occurs at nodes that have been deployed.
Deploy hanging problem was solved by upgrading the NIC firmware version.
Let’s take a look at the link you left. Thank you!

@seochokid, how do you know DNS is not resolving on your deployed hosts?

I searched for the host provisioned through maas as the hostname using the nslookup and ping command, but I couldn’t find it.

I think the DNS issue is similar to the following.


Is there any update on the issue or bug?