OpenCHAMI / deployment-recipes

Ochami deployment recipes
MIT License
8 stars 10 forks source link

[BUG] `boot-filename` option not getting served on fresh installations #61

Closed synackd closed 3 months ago

synackd commented 3 months ago

Describe the bug On fresh clones of deployment-recipes, boot-filename (the DHCP option containing the BSS URL) is not getting served in the DHCPOFFER.

To Reproduce Steps to reproduce the behavior:

  1. Fresh clone of deployment-recipes
  2. Configure .env, start Ochami containers.
  3. Populate SMD and BSS.
  4. Try to boot a node
  5. iPXE fails with: Nothing to boot: No such file or directory (https://ipxe.org/2d03e18e)

Expected behavior boot-filename is present in the DHCPOFFER and contains http://<bss_url>/boot/v1/bootscript?mac=<mac>. iPXE can contact BSS to receive the boot script and find the kernel/initramfs accordingly, e.g:

Next server: 172.16.0.253
Filename: http://172.16.0.253:8081/boot/v1/bootscript?mac=ec:e7:a7:05:9f:ec
http://172.16.0.253:8081/boot/v1/bootscript... ok
bootscript : 749 bytes [script]
http://10.100.0.1:9000/boot-images/efi-images/compute/slurm/vmlinuz-4.18.0-553.5.1.el8_10.x86_64... ok
http://10.100.0.1:9000/boot-images/efi-images/compute/slurm/initramfs-4.18.0-553.5.1.el8_10.x86_64.img... ok

Mitigations Restarting the Ochami services and repopulating SMD/BSS seems to mitigate this issue.

Notes

synackd commented 3 months ago

This seems to be fixed after testing after the merge of #63.