[mythtv-users] Backend with storage group on raid array

[Gary Buhrmaster]

On Sat, Oct 15, 2011 at 10:12, Raymond Wagner <raymond at wagnerrp.com> wrote:

> ... Second, because of the use of parity, it suffers what is called the
> 'write hole'.  Any write that is not exactly on stripe boundaries will
> require all the existing data to be read, and new parity to be
> calculated, before it can be written out.

Pedantically, "write hole" is usually specified as the vulnerability
where the need to read/write the entire block in may result in
writing only some blocks back (in the case of a failure) resulting
in inconsistent data and parity.

> ...  ZFS can do the same with a
> non-volatile flash drive used as a level 2 cache (L2ARC).

ZFS does not suffer from the classically defined "write hole"
because it does not write in place (all data is COW).
The ZIL log (yes, redundant log) on SSD can indeed
substantially improve ZFS commit performance.

To another post, the ZIL log is actually quite small (often
less than 100MB), which means SSD wear leveling can
actually be quite effective in extended the lifetime of the
SSD (and it is small enough so that using expensive
SLC flash is also viable).

Gary