Perhaps someone on this mailing list can shed some light onto some odd zfs circumstances I encountered this weekend. I have an array of 5 400GB drives in a raidz, running on Nexenta. One of these drives showed a SMART error (HARDWARE IMPENDING FAILURE GENERAL HARD DRIVE FAILURE [asc=5d, ascq=10]). I preemptively replaced it, using "zpool replace tank c3t0d0 c3t5d0". The resilver started, but quickly hung at "scrub: resilver in progress, 0.37% done, 132h14m to go". NFS stopped working and I think the system had some responsiveness issues. I did have automatic hourly/daily/weekly snapshots running on the filesystem at the time. I rebooted it, but it would not come up in any sane state, sometimes becoming pingable, but never becoming ssh-able or consolable over serial (as it is configured to do). I tried using various live cds, to no avail. I eventually got it to boot after much gnashing of teeth in Nexenta''s single user mode into a login prompt, but only after both of the drives affected by the replacement were physically removed. Having both drives removed allowed the system to give a login prompt. The resilver proceeded normally, and I watched the resilver complete. I physically reattached the drives and rebooted the system, at which time the pool was online and no longer in a degraded state. The system now boots normally. So, after all that, my primary question is how did the resilvering (which I liken to a rebuilding of the 5 drive array) take place with only 4 drives online? Shouldn''t it have been writing data/parity to the replacement drive? Is this normal and the expected behavior? Thanks for any insight! Thomas