bugGNU GRUB - Bugs: bug #54192, Fails to assemble RAID6 on...

 
 

You are not allowed to post comments on this tracker with your current authentication level.

bug #54192: Fails to assemble RAID6 on sata/scsi devices, but works with virtio (qemu)

Submitted by:  Lars Kruse <sumpfralle>
Submitted on:  Wed 27 Jun 2018 01:24:54 AM UTC  
 
Category:  Disk & Partition Severity:  Major
Priority:  5 - Normal Item Group:  Software Error
Status:  None Privacy:  Public
Assigned to:  None Originator Name: 
Open/Closed:  Open Release: 
Release:  2.02 Reproducibility:  Every Time
Planned Release:  None

Wed 27 Jun 2018 02:47:56 PM UTC, comment #1: 

Correcting a detail of my original submission:

wrong
btw.: Booting from a RAID1 with metadata version 0.9 (instead of RAID6 with v0.9) works also with the "scsi" interface via qemu.

correct
btw.: Booting from a RAID1 with metadata version 0.9 (instead of RAID6 with v1.2) works also with the "scsi" interface via qemu.

(RAID6 was used with metadata v1.2)

Cheers,
Lars

Lars Kruse <sumpfralle>
Wed 27 Jun 2018 01:24:54 AM UTC, original submission:  

Hello,

I configured RAID6 consisting of four drives.
The disks use the GPT partition scheme.
The following partitions are defined on each disk:

  • 1M - BIOS Boot
  • the rest - LVM RAID

The RAID is complete (non-degraded) and uses the metadata version 1.2.
The RAID contains an LVM (which contains the root filesystem, incl. /boot/). But I never saw a sign, that the RAID is assembled by Grub - thus I guess, the LVM is not relevant.
The system runs Debain Stretch.

When booting the server, the rescue console is visible and an error message regarding the missing boot blockdevice is presented.
"ls" shows only the partitions of the disks (e.g. "(hd0,gpt1)"). The RAID is not visible.

The same error is shown (and "ls" outputs the same) when trying to boot these disks via qemu as "scsi" disks:

qemu-system-x86_64 \
    -drive format=raw,if=scsi,readonly,file=/dev/sda \
    -drive format=raw,if=scsi,readonly,file=/dev/sdb \
    -drive format=raw,if=scsi,readonly,file=/dev/sdc \
    -drive format=raw,if=scsi,readonly,file=/dev/sdd

When using the "virtio" blockdevice driver of qemu it magically works:

qemu-system-x86_64 \
    -drive format=raw,if=virtio,readonly,file=/dev/sda \
    -drive format=raw,if=virtio,readonly,file=/dev/sdb \
    -drive format=raw,if=virtio,readonly,file=/dev/sdc \
    -drive format=raw,if=virtio,readonly,file=/dev/sdd

I am surprised by the different behaviour of both blockdevice virtualization backends. Everything else remained unchanged. The disks (and their partitions) can obviously be read while using the "scsi" interface (as demonstrated by the "ls" command).
Thus it feels a bit like a subtile interaction between the sata/scsi module and the RAID module (crazy wild guessing).

btw.: Booting from a RAID1 with metadata version 0.9 (instead of RAID6 with v0.9) works also with the "scsi" interface via qemu.

Do you have any idea, how I could debug the situation?

Thank you for your time!

Cheers,
Lars

Lars Kruse <sumpfralle>

 

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by sumpfralle (Submitted the item)
  •  

    Do you think this task is very important?
    If so, you can add your encouragement to it.
    This task has 0 encouragements so far.

    Only logged-in users can vote.

     

     

     

    No changes have been made to this item

    Back to the top


    Powered by Savane 3.7