NitroAV FW RAID / sbp0 problems?

[Andrew Edmond]

All --
 
After 2 days of research over here, I am finally writing into the list for
help, because I can't tell what is going on here or how to fix it.  I've
done as much testing, reading, updating of sources I can, so maybe someone
here knows what the !#@ is going on ;)
 
Bottom line is that I'm getting these messages alot, but only when the
filesystem is copying a large file (which this fs has to do a lot):

Apr 29 15:03:36 screamfs kernel: (da0:sbp0:0:0:0): Invalid command operation
code
Apr 29 15:03:37 screamfs kernel: (da0:sbp0:0:0:0): SYNCHRONIZE CACHE. CDB:
35 0 0 0 0 0 0 0 0 0 
Apr 29 15:03:37 screamfs kernel: (da0:sbp0:0:0:0): ABORTED COMMAND asc:20,0
Apr 29 15:03:37 screamfs kernel: (da0:sbp0:0:0:0): Invalid command operation
code
Apr 29 15:03:37 screamfs kernel: (da0:sbp0:0:0:0): SYNCHRONIZE CACHE. CDB:
35 0 0 0 0 0 0 0 0 0 
Apr 29 15:03:37 screamfs kernel: (da0:sbp0:0:0:0): ABORTED COMMAND asc:20,0
Apr 29 15:03:37 screamfs kernel: (da0:sbp0:0:0:0): Invalid command operation
code

I am using a NitroAV system (nice little external RAID firewire units), I
have both the 2 disk and 5 disk versions:
(http://www.nitroav.com/store/customer/home.php?cat=10)
 
I have a new firewire 800 card and am using firewire 800 cables:

Apr 29 05:16:34 screamfs kernel: fwohci0: <Texas Instruments TSB82AA2> mem
0xfbb00000-0xfbb03fff,0xfbc00000-0xfbc007ff irq 19 at dev
ice 7.0 on pci1
Apr 29 05:16:34 screamfs kernel: fwohci0: OHCI version 1.10 (ROM=1)
Apr 29 05:16:34 screamfs kernel: fwohci0: No. of Isochronous channels is 4.
Apr 29 05:16:34 screamfs kernel: fwohci0: EUI64 00:01:56:0e:00:00:2c:9e
Apr 29 05:16:34 screamfs kernel: fwohci0: invalid speed 7 (fixed to 3).
Apr 29 05:16:34 screamfs kernel: fwohci0: Phy 1394a available S800, 3 ports.
Apr 29 05:16:34 screamfs kernel: fwohci0: Link S800, max_rec 4096 bytes.
Apr 29 05:16:34 screamfs kernel: firewire0: <IEEE1394(FireWire) bus> on
fwohci0
Apr 29 05:16:34 screamfs kernel: fwe0: <Ethernet over FireWire> on firewire0
Apr 29 05:16:34 screamfs kernel: if_fwe0: Fake Ethernet address:
02:01:56:00:2c:9e
Apr 29 05:16:34 screamfs kernel: fwe0: Ethernet address: 02:01:56:00:2c:9e
Apr 29 05:16:34 screamfs kernel: fwe0: if_start running deferred for Giant
Apr 29 05:16:34 screamfs kernel: sbp0: <SBP-2/SCSI over FireWire> on
firewire0
Apr 29 05:16:34 screamfs kernel: fwohci0: Initiate bus reset
Apr 29 05:16:34 screamfs kernel: fwohci0: node_id=0xc800ffc0, gen=1,
CYCLEMASTER mode
Apr 29 05:16:34 screamfs kernel: firewire0: 1 nodes, maxhop <= 0, cable IRM
= 0 (me)
Apr 29 05:16:34 screamfs kernel: firewire0: bus manager 0 (me)
Apr 29 05:16:34 screamfs kernel: fwohci0: phy int

You can see the firmware on the remote devices are:

# camcontrol devlist
<Oxford 922G Master 0107>          at scbus0 target 0 lun 0 (da0,pass0)
<Oxford 922G Master 0107>          at scbus0 target 1 lun 0 (da1,pass1)

The tunefs looks good:

screamfs# tunefs -p /dev/da0s1
tunefs: ACLs: (-a)                                         disabled
tunefs: MAC multilabel: (-l)                               disabled
tunefs: soft updates: (-n)                                 disabled
tunefs: maximum blocks per file in a cylinder group: (-e)  2048
tunefs: average file size: (-f)                            16384
tunefs: average number of files in a directory: (-s)       64
tunefs: minimum percentage of free space: (-m)             8%
tunefs: optimization preference: (-o)                      time
tunefs: volume label: (-L)                                 

The disk is formatted UFS2, as seen here:

screamfs# dumpfs /dev/da0s1 | less
magic   19540119 (UFS2) time    Fri Apr 29 14:56:58 2005
superblock location     65536   id      [ 4234b75c 26acb731 ]
ncg     8306    size    781421665       blocks  756835800
bsize   16384   shift   14      mask    0xffffc000
fsize   2048    shift   11      mask    0xfffff800
frag    8       shift   3       fsbtodb 2
minfree 8%      optim   time    symlinklen 120
maxbsize 16384  maxbpg  2048    maxcontig 8     contigsumsize 8
nbfree  80613086        ndir    8       nifree  195622848       nffree  47
bpg     11761   fpg     94088   ipg     23552
nindir  2048    inopb   64      maxfilesize     140806241583103
sbsize  2048    cgsize  16384   csaddr  3000    cssize  133120
sblkno  40      cblkno  48      iblkno  56      dblkno  3000
cgrotor 5351    fmod    0       ronly   0       clean   1
avgfpdir 64     avgfilesize 16384
flags   none
fsmnt   /mnt/screamfs
volname         swuid   0

Lastly:

FreeBSD 5.4-STABLE #0: Fri Apr 29 14:07:36 PDT 2005
CPU: AMD Athlon(tm) XP  2600+ (2079.56-MHz 686-class CPU)
  Origin = "AuthenticAMD"  Id = 0x681  Stepping = 1
Features=0x383fbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,
CMOV,PAT,PSE36,MMX,FXSR,SSE>
  AMD Features=0xc0400000<AMIE,DSP,3DNow!>
real memory  = 939327488 (895 MB)
avail memory = 909545472 (867 MB)

So.... what the !# is causing the Synch cache errors?  I am getting the
kernel to panic I guess because of too many of these errors as well.  Any
help is appreciated.
 
Andrew
 
  <http://www.screamnetworks.com/images/bcard/andrew_edmond_bcard.gif>