CentOS 7 deploy succeeds from MAAS but fails from Juju (3.2)?

Hey there!
Any reason you guys can think of that MAAS>Take action>Deploy CentOS 7 would produce different results than juju add-machine --base centos@7?

The machine fails (30min timeout) if I use Juju, but succeeds normally from the MAAS GUI.

This machine has an XFS partition in addition to the root EXT4, and if I remove that XFS partition it works from both sides…
I get the same issue if I have commissioning OS set to 22.04, so I think somehow Juju is overriding the commissioning OS when it requests the deploy o_O?

Anyone experienced this?

Hmmm… It does look like it is requesting 20.04:

 Thu, 02 Nov. 2023 23:52:36	HTTP Request - /images/ubuntu/amd64/ga-20.04/focal/stable/squashfs

Technically the commissioning image is always ubuntu. In short, we use an ubuntu image to install centos on the disk.

Since this sounds more like an integration problem between juju → maas, I’d suggest to reach out the juju community

Hey r00ta,
Ahh gotcha. Yeah I have a parallel conversation on the Juju Mattermost, but they were hoping to see the cloud-init-output.log, but since it’s not finishing deployment and it’s got an old BMC with only Java KVM on a weird port, I can’t get access to it as I’m remote… Trying to get a network exception.

haha, off-topic, your username reminds me of an old friend I lost contact with, but missed talking with. :slight_smile:

nice :slight_smile: must be a good guy :wink:

1 Like