[OpenIndiana-discuss] What happens when a ZIL drive dies?

[Matt Clark]

Yes, it's interesting to consider the possible real world performance.  ZIL writes, even by DDRDrive's stats, are very closely clustered around the tail of the ZIL, with most seeks being to near locations - it's certainly not going to seek more than 8 gig away.  So probably just leaving it at one big partition would be fine too.  In my home (i.e. toy) implementation the SLOG is just a partition on an 80GB drive shared with the root pool and L2ARC, and frankly the gigabit network runs out of puff long before I can tax the server IO with sync writes.  If anyone has one doing more than toy stuff it would be good to hear.

On 6 Jun 2012, at 16:14, Richard Elling wrote:

> NB, the 10k IOPS is for a write span limited to 8GB. For full-size random writes, the
> IOPS dives to an astonishing 300. Still 2x a HDD, but only 2x a HDD. For slog use,
> consider a partition size 8GB or less.
> -- richard