Hi questioners, I wanted to test ZFS support for a 10-disk RAID-Z2 pool I have under FreeBSD 8, as I found Solaris 11 too unstable for my needs.
One disk in the pool is faulted, and is physically not connected to the machine at the moment. The remaining 9 (plus one slog device) are attached, and the entire disk is used on each. I installed FreeBSD 8.0-BETA2 AMD64 on a clean disk, and after rebooting into my new install, ran zpool import. "No pools available" greeted me. (Well, a message about the CDDL module being loaded and about prefetch being disabled as the system has 2GB RAM greeted me, followed by "No pools available".) The arcmsr CLI utility reports all disks connected (in JBOD). (The card is an ARC-1280ML, with disks attached via SAS-4i -> 4x SATA2 cables.) I'm using arcmsr in JBOD mode, so I asked zdb -l /dev/da0 ... zdb -l /dev/da10 what they thought. All four said they saw no meaningful labels on the disks. fdisk /dev/da0 ... da10 reported that the GPT labels are corrupt or otherwise unreadable uniformly across all disks. I'm presuming this has something to do with feeding ZFS the entire disk and not just a slice of it. Fascinatingly, under SXCE b117, all of the disk labels (all four for each disk) in question are quite readable with zdb -l [dev], leading me to believe something's rotten in the state of Denmark. Is this some known caveat with FreeBSD+ZFS? I can't seem to find any documentation for ZFS on FreeBSD 8. Thanks for anything you can offer, - Rich Ercolani -- The sooner you fall behind, the more time you have to catch up. _______________________________________________ email@example.com mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"