Thanks for the update. So far here's what I've seen.
the storage config for the node isn't quite right.
First the nvme devices is marked as grub_boot device, but it's not being used at the root; that's typically what we see. That itself isn;'t fatal, in this case the fatal error is the fact that the node is booting under UEFI (curtin is installing grub-efi packages) and then that fails to work as the storage config does not contain a /boot/efi partition and mountpoint. This typically happens if the node hasn't been commissioned after upgrading to a maas with custom storage config; or generally if the node is out of sync w.r.t maas's view of it's uefi status.
For the reporting/event/install log bug, we can see that the curtin config sent includes a reporting configuration:
The python-requests post is from cloud-init; which IIUC, those messages are being returned
the 'Curtin/0.1' are from curtin but are not appearing in maas UI.
At this point, I don't think there is a curtin bug here.
Thanks for the update. So far here's what I've seen.
the storage config for the node isn't quite right.
First the nvme devices is marked as grub_boot device, but it's not being used at the root; that's typically what we see. That itself isn;'t fatal, in this case the fatal error is the fact that the node is booting under UEFI (curtin is installing grub-efi packages) and then that fails to work as the storage config does not contain a /boot/efi partition and mountpoint. This typically happens if the node hasn't been commissioned after upgrading to a maas with custom storage config; or generally if the node is out of sync w.r.t maas's view of it's uefi status.
For the reporting/ event/install log bug, we can see that the curtin config sent includes a reporting configuration:
install: 10.246. 64.50/MAAS/ metadata/ status/ 4y3h7p', xWnakTM3uzwQCLq Kj,
log_file: /tmp/install.log
post_files: [/tmp/install.log]
reporting:
maas: {consumer_key: XfeDkMX7pqfMWRx8bL, endpoint: 'http://
token_key: w3M3BpvTecqBGxKCS6, token_secret: LYF5FwgGHwkPN79
type: webhook}
After examining the maas regiond log, we can see that curtin does send events back to maas:
016-06-10 19:45:22 [-] 127.0.0.1 - - [10/Jun/ 2016:19: 45:21 +0000] "POST /MAAS/metadata/ status/ 4y3h7p HTTP/1.1" 200 2 "-" "python- requests/ 2.9.1" 2016:19: 45:21 +0000] "POST /MAAS/metadata/ status/ 4y3h7p HTTP/1.1" 200 2 "-" "python- requests/ 2.9.1" 2016:19: 45:22 +0000] "POST /MAAS/metadata/ status/ 4y3h7p HTTP/1.1" 200 2 "-" "Curtin/0.1" 2016:19: 45:22 +0000] "POST /MAAS/metadata/ status/ 4y3h7p HTTP/1.1" 200 2 "-" "Curtin/0.1"
2016-06-10 19:45:22 [-] 127.0.0.1 - - [10/Jun/
2016-06-10 19:45:23 [-] 127.0.0.1 - - [10/Jun/
2016-06-10 19:45:23 [-] 127.0.0.1 - - [10/Jun/
The python-requests post is from cloud-init; which IIUC, those messages are being returned
the 'Curtin/0.1' are from curtin but are not appearing in maas UI.
At this point, I don't think there is a curtin bug here.