I decided recently I'd better start using OpenSolaris to get more familiar with the new advances they are driving into regular Solaris. I really like the new ZFS file system and I had an old 6 pack SCSI JBOD ... so, I attached it to my system and created my home directory as a RAIDz. The disks were only 16GB a piece so I only ended up with about 70GB - still pretty healthy. Even cooler - I can create snapshots at will. Fun!
I thought I had set up a hot stand by - guess not (see image). Knowing the disks were pretty old, I figured I could experience some failure and practice recovery. Today is the lucky day! I heard the array make a wheezing screech and then a CLUNK! nothing else. I went about adding Windows XP to Sun's Virtualbox without a hitch. When my shiny new Windows XP was all installed and patched, I started looking around. That when I noticed this (again, see image). Weird? "resilvered"? Never heard that term before. Fortunately, Jeff Bonwick's blog had the answers. Well, he had the definition of "resilvered" anyway. It worries me that it wasn't notified other than an entry in /var/adm/messages. Luckily, I routinely look at zpool status and the logs (but an email would be nice). I guess it is not yet completely broken. The link they provided indicated if it were really bad, zfs would fail the drive. It looks like it only remapped 6.5k. So there's nothing to do (except blog?). So, I ran "zpool clear junk c2t13d0" and it cleared the error. That was it.