Migrate2Rocky error on MegaRaid driver

Rocky Linux General
1

I am trying to use the migrate2rocky script on a CentOS8 machine. This is an older dell with the unsupported Megaraid drivers. I had to install it using a DUD driver, and then add the elrepo to my repos so I would find the drivers out there when it did a kernal update, else it would quit booting after a kernal update. The script errors out and the log shows:

Problem: cannot install the best update candidate for package kmod-megaraid_sas-07.714.04.00-3.el8_4.elrepo.x86_64

  • nothing provides kernel >= 4.18.0-372.9.1.el8 needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(_dev_info) = 0xbdc3f5dc needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(_dev_err) = 0x1cba9e84 needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(_dev_warn) = 0xaad33b64 needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(dev_printk) = 0x008c88fc needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(pci_find_capability) = 0x6e90426b needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(__num_online_cpus) = 0xc60d0620 needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(_dev_notice) = 0x7d4c60c7 needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(pci_hw_deprecated) = 0xb2930a63 needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
  • nothing provides kernel(pci_hw_unmaintained) = 0x82f2869a needed by kmod-megaraid_sas-07.719.03.00-1.el8_6.elrepo.x86_64
    (try to add ‘–skip-broken’ to skip uninstallable packages or ‘–nobest’ to use not only best candidate packages)

Not sure if I should try a --skip-broken or try a --nobest first, or if that will leave my system unbootable. I do have a complete backup of 16TB’s of data on this server and I can rebuild it from scratch fairly easily if it does go bad, but I would prefer it didn’t.

2

I tried adding the --nobest to the script. It installed, but no boot afterwards. I have reinstalled Rocky and I am reconfiguring. Luckily my 32TB storage array stayed in tact, so I just have to reinstall the OS and re-setup my Samba. I was on CentOS 8.0, so I think it has something to do with it. The DUD driver from the epel repository has a 8.6 version, but for some reason the script wouldn’t see it and it put the new kernel on without a driver. I know there was probably a way to boot from the old kernel and recreate a new one, but it was faster to just re-install the OS. Now I have 4 more servers to do this too.

3

Since this is not a “horizontal (8.6 → 8.6)” conversion, you’d have to update the kernel and the kmod at the same time. I wonder what happens if you update the kernel to 8.6 first (no boot here) and then run the script. Or, update the kernel AND the kmod, then run the script.