major pita w/LSI Tyan TA26 combo

[Ken Gunderson]

Greets Folks:

Been there and done that w/this before but I'm exasperated and don't
know where to go from here. 

Anyways, system details:

FreeBSD 5.4 and 6.0
Tyan TA26 <http://www.tyan.com/products/html/ta26b2882.html>
2 x Opteron 252
4 x 1GB DDR400 ECC Registered Ram
LSI 320-2x and LSI 320-1
Fujitsu U320 SCSI drives, both 10K and 15K rpm

Problem:

Create an array, subject it to some moderate I/O and the logical drive
becomes "degraded" w/a "failed" drive.  The logical drive cannot be
rebuilt.  The physical drive is in fact fine, e.g. can format and
consistency check. Same problem if swap in a different drive, etc.,
e.g.:

megarc -ldInfo -a0 -L2


        **********************************************************************
              MEGARC MegaRAID Configuration Utility(FreeBSD)-1.04
(03-02-2005) By LSI Logic Corp.,USA
        **********************************************************************
          [Note: For SATA-2, 4 and 6 channel controllers, please specify
          Ch=0 Id=0..15 for specifying physical drive(Ch=channel,
Id=Target)]

        Type ? as command line arg for help


        Finding Devices On Each MegaRAID Adapter...
        Scanning Ha 0, Chnl 1 Target 15  

                *******Information Of Logical Drive 2*******

          Logical Drive : 2( Adapter: 0 ):  Status: DEGRADED
        ---------------------------------------------------
        SpanDepth :03     RaidLevel: 1  RdAhead : No  Cache: DirectIo
        StripSz   :128KB   Stripes  : 2  WrPolicy: WriteThru

        Logical Drive 2 : SpanLevel_0 Disks
        Chnl  Target  StartBlock   Blocks      Physical Target Status
        ----  ------  ----------   ------      ----------------------
        0      01    0x00000000   0x0447c000   ONLINE
        0      02    0x00000000   0x0447c000   ONLINE

        Logical Drive 2 : SpanLevel_1 Disks
        Chnl  Target  StartBlock   Blocks      Physical Target Status
        ----  ------  ----------   ------      ----------------------
        0      03    0x00000000   0x0447c000   ONLINE
        1      04    0x00000000   0x0447c000   FAILED

        Logical Drive 2 : SpanLevel_2 Disks
        Chnl  Target  StartBlock   Blocks      Physical Target Status
        ----  ------  ----------   ------      ----------------------
        1      14    0x00000000   0x0447c000   ONLINE
        1      15    0x00000000   0x0447c000   ONLINE


Autorebuild is on, but not able to rebuild, even from w/in LSI BIOS.

Tyan ran tests and concluded that it works fine on Win32 so the
problem is therefore not their problem...

LSI has been a LOT more cooperative but apparently not able to reproduce
the problem.  That's funny because I've been able to reproduce the
problem on 3 DIFFERENT TA26's!!

The issue w/one involved a 320-1 and a simple 2 drive RAID1
mirror that puked under moderate I/O load, e.g. build/install
world.  Upgrading to a 320-2x seems to have solved the problem.  That
particular machine is also using 10K drives and RAID5. Transfers of
approx. 100-500MB from a  RAID1 volume to the RAID5 volume don't seem
to cause problem but I'm afraid to stress test it more than that.  The
320-2x LSI sent me sports the latest "Tundra" chips.  The system
detailed above has older chips, fwiw.

To summarize, Tyan's position is that FreeBSD is an unsupported OS and
that it's a driver issue.  Well, I've already tested w/Scott's latest
and greatest amr (at least as of a couple months ago).  I've also been
using FBSD for many years and have great confidenc in it so I'm
inclined not to swallow Tyan's driver line.  I also see other's having
issues w/Adaptec cards, etc.

Does anybody have ideas and/or receommend anything that actually
works?!?! 

TIA--

-- 
Best regards,

Ken Gunderson

Q: Because it reverses the logical flow of conversation.
A: Why is putting a reply at the top of the message frowned upon?