Booting a ZFS root file system via uEFI

Rob Loan

Well-Known Member
Mar 25, 2017
48
10
48
59
Sometimes to get the ZFS root file system the way we want requires an install in legacy mode and conversion to uEFI mode, perhaps because of:
  1. Installer creates an undesirable partition table
  2. Installer can't find cdrom image after grub2 load
  3. hardware requires uEFI bios to initialize (emmc)
  4. black boot screen perhaps because of uEFI grub2 not recognizing newer CPUs
There is a good wiki but sometimes we prefer to keep the boot image on ZFS. So do a legacy boot mode install onto a throw away USB drive. Its partition table will look like:
Code:
root@pve:~# fdisk -l

Disk /dev/sda: 55.9 GiB, 60022480896 bytes, 117231408 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: 10C3E371-D891-423D-B341-FD3BB74DD114

Device         Start       End   Sectors  Size Type
/dev/sda1         34      2047      2014 1007K BIOS boot
/dev/sda2       2048 117214989 117212942 55.9G Solaris /usr & Apple ZFS
/dev/sda9  117214990 117231374     16385    8M Solaris reserved 1
If we zpool attach rpool sda2 nvme0n1 to add the desired root drive there won't be a partition to write the BIOS boot. So lets match the existing partition table:
Code:
root@pve:~# sgdisk -Z -a1 -n1:34:2047 -t1:EF02 -c1:GRUB -a9 -n9:-8M:0 -t9:bf07 -c9:Reserved -a2 -n2:2047:0 -t2:bf01 -c2:ZFS /dev/nvme0n1
Creating new GPT entries.
GPT data structures destroyed! You may now partition the disk using fdisk or
other utilities.
Setting name!
partNum is 0
REALLY setting name!
Information: Moved requested sector from 976756750 to 976756743 in
order to align on 9-sector boundaries.
Setting name!
partNum is 8
REALLY setting name!
Information: Moved requested sector from 2047 to 2048 in
order to align on 2-sector boundaries.
Setting name!
partNum is 1
REALLY setting name!
The operation has completed successfully.
root@pve:~# zpool attach -f rpool sda2 /dev/disk/by-id/nvme-Samsung_SSD_970_EVO_500GB_S466NX0K901726H-part2
root@pve:~# sgdisk -Z -a1 -n1:34:2047 -t1:EF02 -c1:GRUB -a9 -n9:-8M:0 -t9:bf07 -c9:Reserved -a2 -n2:2047:0 -t2:bf01 -c2:ZFS /dev/sdd
***************************************************************
Found invalid GPT and valid MBR; converting MBR to GPT format
in memory. 
***************************************************************
Warning! Secondary partition table overlaps the last partition by
33 blocks!
You will need to delete this partition or resize it in another utility.
GPT data structures destroyed! You may now partition the disk using fdisk or
other utilities.
Setting name!
partNum is 0
REALLY setting name!
Information: Moved requested sector from 976756750 to 976756743 in
order to align on 9-sector boundaries.
Setting name!
partNum is 8
REALLY setting name!
Information: Moved requested sector from 2047 to 2048 in
order to align on 2-sector boundaries.
Setting name!
partNum is 1
REALLY setting name!
The operation has completed successfully.
root@pve:~# zpool attach -f rpool sda2 /dev/disk/by-id/usb-ZALMAN_External_HDD_______XX00000001-0\:0-part2
root@pve:~# zpool status
  pool: rpool
 state: ONLINE
status: One or more devices is currently being resilvered.  The pool will
continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
  scan: resilver in progress since Sat Nov 10 19:31:58 2018
677M scanned out of 870M at 26.0M/s, 0h0m to go
676M resilvered, 77.87% done
config:
NAME                                                      STATE     READ WRITE CKSUM
rpool                                                     ONLINE       0     0     0
  mirror-0                                                ONLINE       0     0     0
    sda2                                                  ONLINE       0     0     0
    nvme-Samsung_SSD_970_EVO_500GB_S466NX0K901726H-part2  ONLINE       0     0     0
    usb-ZALMAN_External_HDD_______XX00000001-0:0-part2    ONLINE       0     0     0  (resilvering)
errors: No known data errors
root@pve:~# zpool iostat -v
                                                            capacity     operations     bandwidth 
pool                                                      alloc   free   read  write   read  write
--------------------------------------------------------  -----  -----  -----  -----  -----  -----
rpool                                                      870M  54.7G     22     34   938K   829K
  mirror                                                   870M  54.7G     55     84  2.24M  1.98M
    sda2                                                      -      -     10      8   639K   168K
    nvme-Samsung_SSD_970_EVO_500GB_S466NX0K901726H-part2      -      -     29     47   733K   862K
    usb-ZALMAN_External_HDD_______XX00000001-0:0-part2        -      -      0    214    260  9.41M
--------------------------------------------------------  -----  -----  -----  -----  -----  -----
root@pve:~# fdisk -l
Disk /dev/nvme0n1: 465.8 GiB, 500107862016 bytes, 976773168 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: 5336CF94-FBD1-411B-A7CC-C359A8638784

Device             Start       End   Sectors   Size Type
/dev/nvme0n1p1        34      2047      2014  1007K BIOS boot
/dev/nvme0n1p2      2048 976756742 976754695 465.8G Solaris /usr & Apple ZFS
/dev/nvme0n1p9 976756743 976773134     16392     8M Solaris reserved 1

Disk /dev/sda: 55.9 GiB, 60022480896 bytes, 117231408 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: 10C3E371-D891-423D-B341-FD3BB74DD114

Device         Start       End   Sectors  Size Type
/dev/sda1         34      2047      2014 1007K BIOS boot
/dev/sda2       2048 117214989 117212942 55.9G Solaris /usr & Apple ZFS
/dev/sda9  117214990 117231374     16385    8M Solaris reserved 1

Disk /dev/sdd: 465.8 GiB, 500107862016 bytes, 976773168 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: gpt
Disk identifier: DC51A57B-EFFE-4957-A4F5-26351FEACD0C

Device         Start       End   Sectors   Size Type
/dev/sdd1         34      2047      2014  1007K BIOS boot
/dev/sdd2       2048 976756742 976754695 465.8G Solaris /usr & Apple ZFS
/dev/sdd9  976756743 976773134     16392     8M Solaris reserved 1

root@pve:~# lsblk
NAME        MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda           8:0    0  55.9G  0 disk 
├─sda1        8:1    0  1007K  0 part 
├─sda2        8:2    0  55.9G  0 part 
└─sda9        8:9    0     8M  0 part 
sdd           8:48   0 465.8G  0 disk 
├─sdd1        8:49   0  1007K  0 part
├─sdd2        8:50   0 465.8G  0 part 
└─sdd9        8:57   0     8M  0 part 
mmcblk0     179:0    0  29.7G  0 disk 
├─mmcblk0p1 179:1    0  1007K  0 part 
├─mmcblk0p2 179:2    0  29.7G  0 part 
└─mmcblk0p9 259:4    0     8M  0 part 
nvme0n1     259:0    0 465.8G  0 disk 
├─nvme0n1p1 259:5    0  1007K  0 part 
├─nvme0n1p2 259:6    0 465.8G  0 part 
└─nvme0n1p9 259:7    0     8M  0 part
normally the efi partition is formatted vfat32, but our partition is too tiny so vfat16 will do:
Code:
root@pve:~# mkdosfs -n BOOT_EFI /dev/nvme0n1p1
mkfs.fat 4.1 (2017-01-24)
root@pve:~# mount /dev/disk/by-label/BOOT_EFI /boot/efi
root@pve:~# df /boot/efi
Filesystem     1K-blocks  Used Available Use% Mounted on
/dev/nvme0n1p1       988   184       804  19% /boot/efi
root@pve:~# grub-install -d /usr/lib/grub/x86_64-efi /dev/nvme0n1 
Installing for x86_64-efi platform.
EFI variables are not supported on this system.
EFI variables are not supported on this system.
grub-install: error: efibootmgr failed to register the boot entry: No such file or directory.
root@pve:~# ls -l /boot/efi/EFI/proxmox/grubx64.efi 
-rwxr-xr-x 1 root root 182784 Nov 10 19:38 /boot/efi/EFI/proxmox/grubx64.efi

root@pve:~# umount /boot/efi
root@pve:~# mkdosfs -n BOOT_EFI /dev/sdd1
mkfs.fat 4.1 (2017-01-24)
root@pve:~# mount /dev/disk/by-label/BOOT_EFI /boot/efi
root@pve:~# df /boot/efi
Filesystem     1K-blocks  Used Available Use% Mounted on
/dev/sdd1            988     0       988   0% /boot/efi
root@pve:~# grub-install -d /usr/lib/grub/x86_64-efi /dev/sdd1
Installing for x86_64-efi platform.
EFI variables are not supported on this system.
EFI variables are not supported on this system.
grub-install: error: efibootmgr failed to register the boot entry: No such file or directory.
root@pve:~# ls -l /boot/efi/EFI/proxmox/grubx64.efi 
-rwxr-xr-x 1 root root 182784 Nov 10 19:55 /boot/efi/EFI/proxmox/grubx64.efi

root@pve:~# update-initramfs -u     # update /etc/zfs/zpool.cache
update-initramfs: Generating /boot/initrd.img-4.15.17-1-pve
root@pve:~# update-grub             # update /boot/grub/grub.cfg
Generating grub configuration file ...
Found linux image: /boot/vmlinuz-4.15.17-1-pve
Found initrd image: /boot/initrd.img-4.15.17-1-pve
Found memtest86+ image: /ROOT/pve-1@/boot/memtest86+.bin
Found memtest86+ multiboot image: /ROOT/pve-1@/boot/memtest86+_multiboot.bin
Adding boot menu entry for EFI firmware configuration
done
So we have grubx64.efi staged, but because we booted legacy, there are no uEFI variables and that's why efibootmgr failed to register the boot entry.
Code:
root@pve:~# efibootmgr -v 
EFI variables are not supported on this system.
root@pve:~# ls /sys/firmware/efi
ls: cannot access '/sys/firmware/efi': No such file or directory
to boot in uEFI let's boot the USB flash drive image file from rEFInd. reboot, turn off bios legacy boot and turn on uEFI boot and select the USB device with the rEFInd image. It will show both the disks with grubx64.efi and will boot either. Now we can add the boot variables to bios:
Code:
root@pve:~# ls /sys/firmware/efi
config_table  efivars  esrt  fw_platform_size  fw_vendor  runtime  runtime-mapsystabvars
root@pve:~# efibootmgr -v
BootCurrent: 0003
Timeout: 1 seconds
BootOrder: 0002,0003,0004,0005,0001,0000
Boot0000* nvme0n1VenHw(99e275e7-75a0-4b37-a2e6-c5385e6c00cb)
Boot0001* mmcblk0VenHw(99e275e7-75a0-4b37-a2e6-c5385e6c00cb)
Boot0002* UEFI : USB : ZALMAN External HDD : PART 0 : OS BootloaderPciRoot(0x0)/Pci(0x14,0x0)/USB(12,0)/HD(1,GPT,f8b891b7-cb78-45e4-9775-97ddf1735a8c,0x22,0x7de)..BO
Boot0003* UEFI : USB : SanDisk : PART 0 : OS BootloaderPciRoot(0x0)/Pci(0x14,0x0)/USB(14,0)/USB(3,0)/HD(1,GPT,30d07d2b-cd7f-44d6-aa6f-db03ee9f3911,0x800,0x2f9f)..BO
Boot0004* UEFI : USB : MX MXUB3SESU-16G 1.00 : PART 1 : OS BootloaderPciRoot(0x0)/Pci(0x14,0x0)/USB(15,0)/HD(2,GPT,eafc5def-6eeb-415e-94d3-48ddd19f63a7,0x214,0x1680)..BO
Boot0005* UEFI : Samsung SSD 970 EVO 500GB : PART 0 : OS BootloaderPciRoot(0x0)/Pci(0x1d,0x0)/Pci(0x0,0x0)/NVMe(0x1,00-25-38-59-81-B0-EF-DE)/HD(1,GPT,fbfd0bd9-face-43a9-bef8-13f145193a06,0x22,0x7de)..BO
Boot0007* USB : SanDisk : PART 0 : Boot DriveBBS(HD,,0x0)..BO
root@pve:~# efibootmgr -qBb 0
root@pve:~# efibootmgr -qBb 1
root@pve:~# efibootmgr -qc -d /dev/disk/by-id/nvme-Samsung_SSD_970_EVO_500GB_S466NX0K901726H -p 1 -L Samsung970EVO -l "\EFI\proxmox\grubx64.efi"
root@pve:~# efibootmgr -qc -d /dev/disk/by-id/usb-ZALMAN_External_HDD_______XX00000001-0\:0 -p 1 -L ZALMAN -l "\EFI\proxmox\grubx64.efi"
root@pve:~# efibootmgr -o 0,1
BootCurrent: 0003
Timeout: 1 seconds
BootOrder: 0000,0001
Boot0000* Samsung970EVO
Boot0001* ZALMAN
Boot0003* UEFI : USB : SanDisk : PART 0 : OS Bootloader
Boot0004* UEFI : USB : MX MXUB3SESU-16G 1.00 : PART 1 : OS Bootloader
Boot0005* UEFI : Samsung SSD 970 EVO 500GB : PART 0 : OS Bootloader
Boot0007* USB : SanDisk : PART 0 : Boot Drive
now it will boot from either drive and we can remove the rEFInd stick and the install drive and expand the volume.
Code:
root@pve:~# zpool iostat
              capacity     operations     bandwidth 
pool        alloc   free   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
rpool        873M  54.6G      1     19  26.4K   224K
root@pve:~# zpool detach rpool sda2
root@pve:~# zpool online -e rpool nvme-Samsung_SSD_970_EVO_500GB_S466NX0K901726H-part2
root@pve:~# zpool iostat
              capacity     operations     bandwidth 
pool        alloc   free   read  write   read  write
----------  -----  -----  -----  -----  -----  -----
rpool        873M   465G      1     19  26.4K   225K
yes I know its not a good idea to mirror a slow USB drive with a fast nvme ssd.

if we wanted to install the legacy grub, overwrite the vfat partition with:
grub-install -d /usr/lib/grub/i386-pc /dev/nvme0n1

What did I miss?
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!