Troubleshooting server registration

0001 - Disk not seen in the ILO5 servers

ILO sometimes does not detect the disks.

SOLUTION

This sometimes happens if the disks are NVME disks connected to the back-plane; this is expected behavior in MetalSoft. Add a regular disk to the server, in a controller. This allows the registration to complete. The OS still sees the NVMe disks.

0002 - NVME disks not seen in ILO5 if connected to the backplane

You see an error similar to this in the graph:

 failed - Error for PUT https:/ 10.123.240.30/redfish/v1/Systems/1/SmartStorageConfig/Settings - Request failed with status code 404 :
K"error": (" code": "iLO.0.10.ExtendedInfo", "message": "See (Message.ExtendedInfo for more information.",
"@Message.ExtendedInfo": [{"MessageArgs":[*/redfish/v1/Systems/1/SmartStorageConfig/Settings*],
"MessageId": "Base. 1.18. ResourceMissingAtURI

SOLUTION

Upgrade ILO and controller firmware to min. 3.11 version and retry the failed operation.

0003 - diskSizeGb must not be less than 1 during server registration

This happens if MetalSoft cannot determine the size of a disk.

SOLUTION

Reset the RAID controller manually.

0004 - Registration stuck at booting the BDK step

Registration stuck at booting BDK step

SOLUTION

First, check the servers’ console. There are two possible situations:

1. The BDK iso cannot be booted on the server. Check the BDK iso on the repository for corruption.
2. The BDK boots but remains in the booted state. This is often due to the inability of the Site Controller to use the SOL protocol to extract information from the BDK. a. If there is any firewall between the Site Controller and the server’s BMC, check that UDP port 623 is enabled. b. Check if the server’s SOL is enabled in ILO/iDRAC and that the serial port is correct (COM 2 on HP ILO5 is selected in the BIOS). MetalSoft sets these settings automatically during registration, but the BIOS might have been reset during registration by other means. Reconfigure and retry the failed operation.
3. Set the minimum number of interfaces in Site > site > Configuration to 0. This removes the BDK step when the system cannot determine LLDP information via redfish, and allows server registration to finish.

0005 - OS not installing, remains in manual install mode

OS not installing, remains in manual install mode as seen in the server’s console.

SOLUTION

1. Check the OS template by mounting the ISO in the image builder and looking at the kickstart file that is injected into the ISO. If the kickstart file is missing information, it becomes invalid, forcing the OS install to drop into manual mode. This is often due to the lack of a logical network attached to a server.
   1. Create a fabric and link a logical network to the server before deployment
   2. Modify the template via the Assets tab in the template to hardcode network information

0006 - Global Controller not responding (if deployed via OVA)

In certain circumstances the Global Controller is not responding on some pages but not all.

SOLUTION

OVA can run out of disk space while building many concurrent large ISOs, such as RedHat ISOs, which have 12GB.