My Razor server is offline and I can't boot any machines in my PE deployment. What can I do to fix this?
Version and Installation Information
PE version: 3.7 to 2019.2
OS: Any RHEL, CentOS
If your Razor server cannot respond to requests, nodes in your deployment will not be able to complete the boot-up cycle.
In a functional Razor deployment, the node boots and requests an iPXE script from the TFTP server. Razor identifies the node and issues instructions for provisioning, booting into the microkernel, or booting locally, and the node follows the instructions.
As a temporary fix when Razor cannot respond to requests, replace
/var/lib/tftpboot/bootstrap.ipxe on your TFTP server with the following script, which will boot your nodes locally:
#!ipxe echo Razor server down, booting locally instead sleep 3 echo forcing local booting with sanboot 0x80 sanboot --no-describe --drive 0x80 echo exiting iPXE and continuing to next firmware boot device exit 0
During the first 29 minutes that Razor is unresponsive, nodes in your deployment will continue to retry the Razor server. If you replace the script or if the Razor service resumes proper functionality during that time, reboot the nodes manually or via IPMI.
Once this script is in place and the nodes are booting successfully, contact Support to troubleshoot the Razor failure and safely add Razor back into the workflow.