Slow resilvering with mirrored ZIL

[Gezeala M. Bacuño II]

On Tue, Jul 16, 2013 at 6:47 PM, Warren Block <wblock at wonkity.com> wrote:

> On Tue, 16 Jul 2013, Gezeala M. Bacuño II wrote:
>
>  On Fri, Jul 5, 2013 at 6:08 PM, Freddie Cash <fjwcash at gmail.com> wrote:
>>
>>
>>> ZFS- on-Linux has added this as "-o ashift=" property for zpool create.
>>>
>>> There's a threat on the illumos list about standardising this s across
>>> all
>>> ZFS- using OSes.
>>>
>>>
>>>
>>>  +1 on this. We tested zfs-on-linux last year and it does automatically
>> handle disk partitioning for correct alignment. What we do is just add
>> ashift=12 option during zpool create. No more gpart/gnop/ashift/import
>> steps.
>>
>> http://zfsonlinux.org/faq.**html#**HowDoesZFSonLinuxHandlesAdvace**
>> dFormatDrives<http://zfsonlinux.org/faq.html#HowDoesZFSonLinuxHandlesAdvacedFormatDrives>
>>
>>
>> Back to FreeBSD ZFS,
>>
>> After reading the thread, I'm still at a loss on this (too much info I
>> guess).. regarding gpart/gnop/ashift tweaks for alignment, do we still
>> need
>> to perform gpart on newly purchased (SSD/SATA/SAS) Advanced Format drives?
>> Or, skip gpart and proceed with gnop/ashift only?
>>
>
> If ZFS goes on a bare drive, it will be aligned by default.  If ZFS is
> going in a partition, yes, align that partition to 4K boundaries or larger
> multiples of 4K, like 1M.
>
>
Your statement is enlightening and concise, exactly what I need. Thanks.


> The gnop/ashift workaround is just to get ZFS to use the right block size.
>  So if you don't take care to get partition alignment right, you might end
> up using the right block size but misaligned.
>
> And yes, it will be nice to be able to just explicitly tell ZFS the block
> size to use.


We do add the entire drive (no partitions) to ZFS, perform gnop/ashift and
other necessary steps and then verify ashift=12 through zdb.

The gpart/gnop/ashift steps, if I understand correctly (do correct me if
I'm stating this incorrectly), is needed for further SSD performance
tuning. Taking into consideration leaving a certain chunk for wear leveling
and also if the SSD has a size that may be too big for L2ARC.