I've Googled to try to find others who noticed a similar symptom as what I am about to describe, but found none...
I recently installed osol 2008.10, on a system with 4 1TB drives. By way of background, on each of the drives, I created one 64GB slice, and then one slice with the rest of the disk. I have "rpool" as a mirror using the 64GB slices on each of the first two drives, "spool" as a stripe using the 64GB slices on the last two drives (which I used mostly for swap), then "apool" as a RAIDZ1 using the large slices on each of the drives. Everything works well. Before I deploy the system, I wanted to test the redundancy. I first disconnected the second drive and rebooted. Everything worked great -- "rpool" and "spool" showed a degraded status with "zpool status", so I reconnected the second drive and rebooted, and the pool showed "normal". Then, I tried the same experiment, but with the second drive disconnected, rather than the first. I had previously installed the boot blocks, and so was able to boot. However, the boot sequence hung (for overnight). I restarted and booted the kernel with "-v", and the last message was something about finding "sd4". Interestingly, if I reconnect the first drive and reboot successfully in verbose mode, I seem to see only messages up to "sd3" (though I might miss the extra, as the successful boot results in a lot of messages). If I boot using the Live CD, and bring up a terminal (or boot into text mode), a "zfs import" just hangs in an unkillable state. Any ideas of what else I could try? Jason -- This message posted from opensolaris.org
