On Fri, Mar 26 at 7:29, Edward Ned Harvey wrote:
> Using fewer than 4 disks in a raidz2 defeats the purpose of raidz2, as > you will always be in a degraded mode.
Freddie, are you nuts? This is false. Sure you can use raidz2 with 3 disks in it. But it does seem pointless to do that instead of a 3-way mirror.
One thing about mirrors is you can put each side of your mirror on a different controller, so that any single controller failure doesn't cause your pool to go down. While controller failure rates are very low, using 16/24 or 14/21 drives for parity on a dataset seems crazy to me. I know disks can be unreliable, but they shouldn't be THAT unreliable. I'd think that spending fewer drives for "hot" redundancy and then spending some of the balance on an isolated warm/cold backup solution would be more cost effective. http://blog.richardelling.com/2010/02/zfs-data-protection-comparison.html Quoting from the summary, "at some point, the system design will be dominated by common failures and not the failure of independent disks." Another thought is that if heavy seeking is more likely to lead to high temperature and/or drive failure, then reserving one or two slots for an SSD L2ARC might be a good idea. It'll take a lot of load off of your spindles if your data set fits or mostly fits within the L2ARC. You'd need a lot of RAM to make use of a large L2ARC though, just something to keep in mind. We have a 32GB X25-E as L2ARC and though it's never more than ~5GB full with our workloads, most every file access saturates the wire (1.0 Gb/s ethernet) once the cache has warmed up, resulting in very little IO to our spindles. --eric -- Eric D. Mudama edmud...@mail.bounceswoosh.org _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
