Search and replace GUID in initramfs script

[ghfields]

Motivation and Context

Fix a mailing list reported regression (http://list.zfsonlinux.org/pipermail/zfs-discuss/2019-January/033049.html) following #8052 for a user who kept root pool data in the root dataset.

Description

After #8052, ZFS_BOOTFS and ZFS_RPOOL values are reconstructed following a sed command that ensures $pool is the pool name, not a GUID. This broke a user's root pool setup, who used the zfs root dataset as their root. After #8052, initramfs attempted to mount rpool/rpool, instead of just rpool. This is corrected by doing a search and replace for only the GUID value within ZFS_BOOTFS and ZFS_RPOOL instead, leaving everything else alone. As mailing responses indicate, this should not be considered a best practice for root pool layout, however it is valid and did work in the past.

How Has This Been Tested?

Installed Ubuntu 18.04.1 to root of dataset
-- Note: Ubuntu's update-grub added a trailing "/" that needed to be manually removed from kernel line
System booted as expected.
Edited the grub.cfg and replaced kernel line pool name with pool GUID
System booted as expected.
Checked for regressions by moving data from root dataset to "rpool/ROOT/ubuntu" and set the kernel line appropriately.
System booted as expected.
Edited the grub.cfg kernel line to "<GUID>/ROOT/ubuntu"
System booted as expected.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the ZFS on Linux code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• All new and existing tests passed.
• All commit messages are properly formatted and contain Signed-off-by.

[Conan-Kudo]

@ghfields Do we need this change for dracut too? People can and do use dracut on Debian/Ubuntu, and it's the initramfs system used on all other major distributions...

[codecov]

Codecov Report

Merging #8356 into master will decrease coverage by 0.12%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #8356      +/-   ##
==========================================
- Coverage   78.77%   78.65%   -0.13%     
==========================================
  Files         382      380       -2     
  Lines      117770   115731    -2039     
==========================================
- Hits        92775    91028    -1747     
+ Misses      24995    24703     -292

Flag	Coverage Δ
#kernel	79.06% <ø> (-0.28%)	⬇️
#user	67.53% <ø> (+0.43%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2531ce3...e446497. Read the comment docs.

[ghfields]

@ghfields Do we need this change for dracut too? People can and do use dracut on Debian/Ubuntu, and it's the initramfs system used on all other major distributions...

I'm unsure of the Dracut status. Not sure if it imports by GUID or if a root dataset root pool confuses its boot. I've seen initramfs and dracut changes happening in separate pull requests. I'll make sure dracut maintainers are informed about this so they can maintain parity.

[gmelikov]

@Conan-Kudo the initial change with regression was only applied to initramfs script, so I don't think we need to change dracut (but I didn't test it).

[behlendorf]

Can we get this PR refreshed to address the review feedback.

[gmelikov]

LGTM

[ghfields]

@rlaager Could you review this one liner bugfix?

[rlaager]

LGTM