Proposed validation test case: root on LVM thinp

Chris Murphy lists at colorremedies.com
Tue Dec 17 08:09:35 UTC 2013


On Dec 17, 2013, at 12:57 AM, Chris Murphy <lists at colorremedies.com> wrote:

> 
> On Dec 16, 2013, at 6:33 PM, Adam Williamson <awilliam at redhat.com> wrote:
> 
>> Actually - it'd basically just be the 'guided installation' table from
>> https://fedoraproject.org/wiki/User:Adamwill/Draft_storage_matrix#Guided_installationwithout all the other bits, basically.
> 
> I basically get this. I expect there'll be separate matrices or pages for the archs?
> 
> Question for EFI as an arch, we haven't been testing every layout presumably because of overlap with x86_64. But I'd guess the QA:Testcase_dualboot_with_windows tests for existing GPT, and EFI System partition, rather than actually creating them (correctly). I'll bet the code that does that is called once for all of the guided partition schemes. So it seems like if we check the Windows case to make sure it doesn't break Windows, we also get a "use existing GPT and ESP" test of code; but still need one that "creates new GPT and ESP" layout test of the code. Yes? I did that on Mac EFI, but it's got slightly different code.

Another one I'm thinking of is software RAID. On BIOS, it must be the case that every member drive is getting a copy of grub2 core.img in the MBR gap, because when I yank a virtio drive, the system still boots (slowly). I haven't tested this for EFI or how that could even work because if a member drive goes away and yet the NVRAM entry is pointing to that particular (bad luck) drive, now what? I guess there's a fallback NVRAM entry that would be needed pointing to one of the other drives's EFI System partitions. I don't know that this works. I assume we want it to work one day.

Might also look at the prebaked grubx64.efi file for what modules are being put into it. Unlike BIOS which can find all grub modules, it's not working that way on x86 EFI due to Secure Boot, the grubx64.efi has to be signed. But that package isn't baking all modules in it. Why I don't know. But the missing ones might indicate things that work on BIOS that probably break on EFI. The only one I ran into a while ago was LVM, but funny enough now that lvm.mod is baked into the grub efi bootloader, we dropped /boot on LVs in anaconda for F20.


Chris Murphy


More information about the test mailing list