Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 23 Jun 2010 17:28:01 -0600
From:      Mark Costlow <cheeks@swcp.com>
To:        freebsd-questions@freebsd.org
Subject:   Problem running fdisk via sysinstall
Message-ID:  <20100623232801.GA5925@risu.swcp.com>

next in thread | raw e-mail | index | archive | help
I hope this question isn't too stupid.

I have a machine with a 3Ware RAID card, with 4 SATA drives attached.

2 drives are 250GB in a RAID1 volume, and act as the boot disk with
a standard freebsd partiction map (/, /var, /usr, and swap on this disk).

The other 2 drives are 1TB in a RAID1 volume, intended to be mounted
as a separate data partition.  At boot both volumes are recognized:

Jun 22 18:38:51 ebi7 kernel: da0 at twa0 bus 0 target 0 lun 0
Jun 22 18:38:51 ebi7 kernel: da0: <AMCC 9550SX-4LP DISK 3.08> Fixed Direct Access SCSI-5 device 
Jun 22 18:38:51 ebi7 kernel: da0: 100.000MB/s transfers
Jun 22 18:38:51 ebi7 kernel: da0: 238408MB (488259584 512 byte sectors: 255H 63S/T 30392C)
Jun 22 18:38:51 ebi7 kernel: da1 at twa0 bus 0 target 1 lun 0
Jun 22 18:38:51 ebi7 kernel: da1: <AMCC 9550SX-4LP DISK 3.08> Fixed Direct Access SCSI-5 device 
Jun 22 18:38:51 ebi7 kernel: da1: 100.000MB/s transfers
Jun 22 18:38:51 ebi7 kernel: da1: 953664MB (1953103872 512 byte sectors: 255H 63S/T 121575C)

da0 is fine, and the system boots off of it with no problem.

When I try to add da1 to the system, I get the following:

* Run systinstall, Configure, Fdisk, select da1
* Get the friendly warning about the large geometry, click "Yes"
* Hit "A" to use entire disk.  Hit "W" to save, click "Yes",
select "None" for boot record.
* Fdisk says: "Wrote FDISK partition information out successfully."
* Per handbook, get out of sysinstall and re-run it, then
try to Label.  In the label editor, it knows nothing about
da1 (the device can be selected when going into the label
editor, but I can't create any partitions).

At the time when Fdisk says "Wrote FDISK partition information out,
successfully." this gets logged to /var/log/messages:

Jun 23 17:11:18 ebi7 kernel: GEOM: da1: corrupt or invalid GPT detected.
Jun 23 17:11:18 ebi7 kernel: GEOM: da1: GPT rejected -- may not be recoverable.


I've tried several variations, including running the command-line
equivalents, but keep hitting this same error (fdisk thinks everything
is good, but the GPT error is logged).  I've also noticed that
/dev/da1 exists, but there is no /dev/da1s1 or /dev/da1s1e ... I'm
not sure when those should get created.

And the final possibly-relevant tidbit: these drives used to be
part of a different RAID on a linux system.  They've been re-initialized
into the RAID card on this system, and I've zero'd the first 1k of
the volume with dd, so I don't *think* that should be a factor.
I've worked with about a dozen systems with the same hardware in
the configuration outlined above and haven't seen this problem
before.  But I'm usually using fresh new disks so maybe it matters.

I've googled this issue and found several people reporting similar
symptoms over the years, but haven't found any posted solutions
aside from telling people to read geom(8).

Any hints or clue-by-fours?

Mark
-- 
Mark Costlow    | Southwest Cyberport | Fax:   +1-505-232-7975
cheeks@swcp.com | Web:   www.swcp.com | Voice: +1-505-232-7992

abq-strange.com -- Interesting photos taken in Albuquerque, NM
                   Last post: Shoe Pole - 2009-07-07 20:18:22



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100623232801.GA5925>