So I've taken the plunge to upgrade to RC3 to assist in trying to debug upgrades and I've run into a snag. I've posted this in storage instead of the upgrade section due to it being specifically about storage drives. My system is running on an Intel i7 with 24GB RAM (I'm at work and will post hardware details if needed later) so I have lots of horsepower. My storage consists of 2 ZFS array's and a standalone drive for OS.
Disks ada0-3 are in storage0 zpool.
Disks ada4-6 are in volume0 zpool.
Disk ada7 is my standalone drive for FreeBSD OS etc.
My problem is when I fist installed my FreeBSD9 release, I tried working with the softraid on my Intel and SiI controller so the metadata is on each drive but was unhappy with the tweaking required especially if I needed to recover (how ironic). I took the drives out of RAID mode and am running just in AHCI mode and continued with individual drives in the zpool. When I did my update via command line the new kernel got installed and it appears that graid(8) starts automatically via the generic kernel now and it won't boot from disk ada7 indicating it's part of a raid configuration (detected from the metadata) and tries to mount the raid pool. I get dropped to the bootloader helper and I can't force a mount on my drive because graid thinks the volume is broken and refuses to let me proceed.
I see via this post someone else had a similar problem:
http://forums.freebsd.org/showthread.php?t=34691
However when I boot my 9.1-RC3 to a live CD and try to do a graid status, there is nothing present and a geom list shows all drives available.
My question is: is there a quick way to clear the metadata on my boot drive (ada7) just so I can continue with the upgrade or should I tear the drive out and try clearing it from a procedure like this?:
http://kezhong.wordpress.com/2011/06/14/how-to-remove-bios-raid-metadata-from-disk-on-fedora/
Additionally, am I going to have problems mounting my zpools after I clear the boot drive's metadata and have to clear metadata from each of the other 6 drives potentially causing problems with the zpool?
Or am I better off just reloading 9.0-RELEASE and get the system back operational and work from there?
Advice or solutions are appreciated.
Cheers,
Disks ada0-3 are in storage0 zpool.
Disks ada4-6 are in volume0 zpool.
Disk ada7 is my standalone drive for FreeBSD OS etc.
My problem is when I fist installed my FreeBSD9 release, I tried working with the softraid on my Intel and SiI controller so the metadata is on each drive but was unhappy with the tweaking required especially if I needed to recover (how ironic). I took the drives out of RAID mode and am running just in AHCI mode and continued with individual drives in the zpool. When I did my update via command line the new kernel got installed and it appears that graid(8) starts automatically via the generic kernel now and it won't boot from disk ada7 indicating it's part of a raid configuration (detected from the metadata) and tries to mount the raid pool. I get dropped to the bootloader helper and I can't force a mount on my drive because graid thinks the volume is broken and refuses to let me proceed.
I see via this post someone else had a similar problem:
http://forums.freebsd.org/showthread.php?t=34691
However when I boot my 9.1-RC3 to a live CD and try to do a graid status, there is nothing present and a geom list shows all drives available.
My question is: is there a quick way to clear the metadata on my boot drive (ada7) just so I can continue with the upgrade or should I tear the drive out and try clearing it from a procedure like this?:
http://kezhong.wordpress.com/2011/06/14/how-to-remove-bios-raid-metadata-from-disk-on-fedora/
Additionally, am I going to have problems mounting my zpools after I clear the boot drive's metadata and have to clear metadata from each of the other 6 drives potentially causing problems with the zpool?
Or am I better off just reloading 9.0-RELEASE and get the system back operational and work from there?
Advice or solutions are appreciated.
Cheers,