Solved Does FreeBSD block access to hard disk drives, from within a VirtualBox machine?

_martin

Aspiring Daemon

Reaction score: 177
Messages: 842

I was adding all four drives, your instructions had me add just two drives,
I gave you an example how the Linux VirtualBox VM can access physical drives from FreeBSD host. These were not copy-paste instructions. After all, even my drive names are different.
I was assembling raid1, as can be read from the mdadm output.

As for the coredump and related issues ? Hard to say without 0 information on the problem..
 
OP
J

jjthomas

Member

Reaction score: 4
Messages: 83

Your instructions were a guide, and they were perfect. Thank you. FWIW I used my own naming convention. But at the end of the day, I got it resolved.

I'm going to delete the core-dump. I got what I needed.

-JJ
 
OP
J

jjthomas

Member

Reaction score: 4
Messages: 83

I'm gonna guess the Fedora kernel was newer, and made some incompatible changes to the md metadata on the disks. Weird that it only affected one disk, though.
I know, weird. I know Windows 10 will put some kind of signature block on a computer's drives. I have no idea what Fedora did. CentOS uses 3.xx and Fedora uses 5.xx. But I was using the test version of Fedora. Who knows.

Oh well.

-JJ
 

Jose

Well-Known Member

Reaction score: 158
Messages: 302

Might've been a bug in that test kernel. Gawd only knows what kernel patches Fedora tests on their users. In any case, congrats on escaping mdadm/lvm! Welcome to the peace and serenity of ZFS.
 

_martin

Aspiring Daemon

Reaction score: 177
Messages: 842

I think it may have been also due to the partial activation, or attempt to activate that raid with few devices missing. Also older Fedora might have a problem with newer raid devices.
I never used Fedora, but in general Linux/mdadm version 0.9 has metadata in the end of the disk/LUN making LUN extension a pain. Later (1.x?) version puts metadata at the beginning making the LUN extension less of a pain.

When it comes to a VM crash it could have been that VM had insufficient memory to pull this.
 
OP
J

jjthomas

Member

Reaction score: 4
Messages: 83

I gave the VM 8G, the hardware has 32G. It should have been enough.
 
Top