ZFS bug: was creating ZIL ignores vfs.zfs.min_auto_ashift, should be ZIL sets improper ashift with AHCI controllers

Steven Hartland killing at multiplay.co.uk
Fri Nov 7 13:31:20 UTC 2014

On 07/11/2014 11:59, Andriy Gapon wrote:
> On 07/11/2014 11:48, Steven Hartland wrote:
>> On 07/11/2014 07:31, Borja Marcos wrote:
>>> snip...
>>> You are right!
>>> This time I had not altered min_auto_ashift but I had added a quirk to
>>> ata_da.c. So it
>>> correctly used the quirk sector size to adjust ashift.
>>> Seems the bug should be "improper cache file generation for a pool used for
>>> boot". I'll
>>> see if I can run some more tests.
>> I believe the attached is the correct fix, but I've asked upstream as this test
>> was explicitly added ages ago.
> It could be that this problem is of no interest / relevance to the illumos
> upstream as they have more restrictions on how a root pool can be configured.  I
> am not sure if they allow slog devices for root pools.
illumos does indeed have the following restrictions for root pools 
according to zfs_ioc_vdev_add
* Concatenated devices are not supported
* Intent log device is not supported

So your right but I'm still curious if there are any reasons why we 
shouldn't sync the zpool.cache file for the root pool as that seems 
independent to this? The only thing I can think of is that it was an 
optimisation but this would still break things like a reguid.


More information about the freebsd-fs mailing list