> Firstly, if you're running logs you really need a pair of devices. ZFS doesn't cope at all well if a log device > fails - you'll potentially loose access to your entire pool.
Yes, we generally have it deployed that way. But, we were in a pinch and hoping for a jump in performance, even if temporarily. I understand about not being able to remove log devices and what not. > And secondly, are you talking here about the entire pool going unusable when this drive started experiencing > problems? If so, that sounds like the issue I've been harping on about for well over half a year now - ZFS > relying on driver timeouts, and pools freezing when any single device has a problem. Yes, the whole pool would stall. I assumed it was because it was the first device in the write path, but based on your thread that may not be the case. _______________________________________________ storage-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/storage-discuss
