DragonFlyBSD

• Matthias Schmidt wrote:

Crash dump is on its way to leaf in ~matthias/crash/crash_vn.7z (7zip
archive). The box runs master from Sep 27.

Quick note, one can unzip the 7zip archive w/ archivers/p7zip:

% 7x vn_crash.7z

I know, another compression program, but it compresses the 250MB crash
dump to 37MB :)

Cheers

Matthias

Hrm. I don't think you can create an 8 TB UFS1 filesystem. The
limitation for UFS is 2TB, and even that might be a bit dangerous
because ufs uses 32 bit filesystem block numbers. Theoretically
it is based on the UFS frag which is usually 1K or 2K, and negative
block numbers are used to index meta-data, so (2^31) * 1K = 2TB.

I don't quite know why it crashed... the VCHR is missing its cdev
    pointer (it is NULL for some reason), but it could be related to the
    size of the backing store you tried to create the UFS filesystem with.

-Matt
                    Matthew Dillon 
                    <dillon@backplane.com>

• Matthew Dillon wrote:

Hrm. I don't think you can create an 8 TB UFS1 filesystem. The
limitation for UFS is 2TB, and even that might be a bit dangerous
because ufs uses 32 bit filesystem block numbers. Theoretically
it is based on the UFS frag which is usually 1K or 2K, and negative
block numbers are used to index meta-data, so (2^31) * 1K = 2TB.

I didn't try to do that :) I just run "mkinitrd" to create a new
initrd, that's all. According to the script it creates a 15 MB vn(4)
disk.

I don't quite know why it crashed... the VCHR is missing its cdev
pointer (it is NULL for some reason), but it could be related to the
size of the backing store you tried to create the UFS filesystem with.

Do not have the exact numbers here, but IIRC about 40GB.

Cheers

Matthias

Can you please try to reproduce the issue with the recent fixes to vn(4) ?

Cheers,
Alex Hornung