Pb between an UltraTrak SX8000 and an Adaptec 29160

Pb between an UltraTrak SX8000 and an Adaptec 29160

[didier]

I have just connected a brand new UltraTrak SX8000
(from Promise with 8 x 200GB IDE drives) to an
adaptec 29160. The UltraTrak SX8000 is connected
via a brand new LVD cable (1meter).

While everything seems fine, I am getting a bunch of

(scsi0:A:0:0): parity error detected in Data-out phase. SEQADDR(0x56)
SCSIRATE(0xc2)
        No terminal CRC packet recevied
(scsi0:A:0:0): parity error detected in Data-out phase. SEQADDR(0x54)
SCSIRATE(0xc2)
        No terminal CRC packet recevied
(scsi0:A:0:0): parity error detected in Data-out phase. SEQADDR(0x54)
SCSIRATE(0xc2)
        No terminal CRC packet recevied
(scsi0:A:0:0): parity error detected in Data-out phase. SEQADDR(0x55)
SCSIRATE(0xc2)
        No terminal CRC packet recevied
(scsi0:A:0:0): parity error detected in Data-out phase. SEQADDR(0x90)
SCSIRATE(0xc2)
        No terminal CRC packet recevied

in my /var/log/messages.

I was wondering if someone could give me some insight.

The UltraTrak SX8000 has its internal termination enabled.

Here is what I get during my boot:

Nov 14 11:04:32 dagdhasrv kernel: Linux version 2.4.18-17.7.xsmp
(bhcompile@daffy.perf.redhat.com) (gcc version 2.96 20000731 (Red Hat
Linux 7.3 2.96-112)) #1 SMP Tue Oct 8 12:37:04 EDT 2002
[..........]
Nov 14 11:04:35 dagdhasrv kernel: SCSI subsystem driver Revision: 1.00
Nov 14 11:04:35 dagdhasrv kernel: kmod: failed to exec /sbin/modprobe -s
-k scsi_hostadapter, errno = 2
Nov 14 11:04:35 dagdhasrv kernel: scsi0 : Adaptec AIC7XXX EISA/VLB/PCI
SCSI HBA DRIVER, Rev 6.2.8
Nov 14 11:04:35 dagdhasrv kernel:         <Adaptec 29160 Ultra160 SCSI
adapter>
Nov 14 11:04:35 dagdhasrv kernel:         aic7892: Ultra160 Wide Channel
A, SCSI Id=7, 32/253 SCBs
Nov 14 11:04:35 dagdhasrv kernel: 
[........]
Nov 14 11:04:35 dagdhasrv kernel: blk: queue c35ef618, I/O limit 4095Mb
(mask 0xffffffff)
Nov 14 11:04:35 dagdhasrv kernel:   Vendor: Promise   Model: 7 Disk
RAID5      Rev: 1.10
Nov 14 11:04:35 dagdhasrv kernel:   Type:   Direct-Access
ANSI SCSI revision: 03
Nov 14 11:04:35 dagdhasrv kernel: blk: queue c35ef418, I/O limit 4095Mb
(mask 0xffffffff)
Nov 14 11:04:35 dagdhasrv kernel: scsi0:A:0:0: Tagged Queuing enabled.
Depth 253
[.......]
Nov 14 11:04:35 dagdhasrv kernel: (scsi0:A:0): 160.000MB/s transfers
(80.000MHz DT, offset 62, 16bit)
Nov 14 11:04:35 dagdhasrv kernel: SCSI device sda: -1951217920 512-byte
hdwr sectors (100488 MB)
Nov 14 11:04:35 dagdhasrv kernel: Partition check:
Nov 14 11:04:35 dagdhasrv kernel:  sda: sda1

Any idea ?

Thanks in advance for any suggestion

Re: Pb between an UltraTrak SX8000 and an Adaptec 29160

[Justin T. Gibbs]

> I have just connected a brand new UltraTrak SX8000
> (from Promise with 8 x 200GB IDE drives) to an
> adaptec 29160. The UltraTrak SX8000 is connected
> via a brand new LVD cable (1meter).
>
> While everything seems fine, I am getting a bunch of
>
> (scsi0:A:0:0): parity error detected in Data-out phase. SEQADDR(0x56)
> SCSIRATE(0xc2)
>         No terminal CRC packet recevied

I've never seen false positives from this diagnostic.  It looks like
your Promise device is not properly implementing DT transfers.  Is
there a firmware update for the device?

--
Justin

RE: Pb between an UltraTrak SX8000 and an Adaptec 29160

[didier]

The UltraTrak was updated to the latest available firmware
before connected to the server.

I e-mailed promise tech-support last week but did not get
any answer so far.

On a different note, who should I contact regarding the display
bug:

Nov 14 11:04:35 dagdhasrv kernel: (scsi0:A:0): 160.000MB/s transfers
(80.000MHz DT, offset 62, 16bit)
Nov 14 11:04:35 dagdhasrv kernel: SCSI device sda: -1951217920 512-byte
hdwr sectors (100488 MB)
                                                   ^^^^^^^^^^^
Nov 14 11:04:35 dagdhasrv kernel: Partition check: 
Nov 14 11:04:35 dagdhasrv kernel:  sda: sda1

Thanks again, Didier.

> > I have just connected a brand new UltraTrak SX8000
> > (from Promise with 8 x 200GB IDE drives) to an
> > adaptec 29160. The UltraTrak SX8000 is connected
> > via a brand new LVD cable (1meter).
> >
> > While everything seems fine, I am getting a bunch of
> >
> > (scsi0:A:0:0): parity error detected in Data-out phase. 
> SEQADDR(0x56)
> > SCSIRATE(0xc2)
> >         No terminal CRC packet recevied
> 
> I've never seen false positives from this diagnostic.  It 
> looks like your Promise device is not properly implementing 
> DT transfers.  Is there a firmware update for the device?

RE: Pb between an UltraTrak SX8000 and an Adaptec 29160

[Alan Cox]

On Mon, 2002-11-18 at 13:01, didier wrote:
> Nov 14 11:04:35 dagdhasrv kernel: (scsi0:A:0): 160.000MB/s transfers
> (80.000MHz DT, offset 62, 16bit)
> Nov 14 11:04:35 dagdhasrv kernel: SCSI device sda: -1951217920 512-byte
> hdwr sectors (100488 MB)
>                                                    ^^^^^^^^^^^
> Nov 14 11:04:35 dagdhasrv kernel: Partition check: 
> Nov 14 11:04:35 dagdhasrv kernel:  sda: sda1

Linux 2.4.x doesnt support single volumes over 1Tb. Thats fixed in 2.5

RE: Pb between an UltraTrak SX8000 and an Adaptec 29160

[didier]

> Linux 2.4.x doesnt support single volumes over 1Tb. Thats fixed in 2.5

Doh. My mistake.

However I am still experiencing some weird behaviour even with small
Raid5 Array.

I have recreated a new Raid 5 array with 4 drives (size = 558GB,
CHS map = 72944/254/63).

Nov 20 08:38:49 dagdhasrv kernel: SCSI device sda: 1171874688 
                          512-byte hdwr sectors (-499511 MB)
                                                 ^^^^^^^^^^^
                                
Nov 20 08:38:49 dagdhasrv kernel:  sda: sda1

Nov 20 08:43:45 dagdhasrv kernel: ^INo terminal CRC packet recevied
Nov 20 08:43:45 dagdhasrv kernel: (scsi0:A:0:0): parity error detected
in Data-out phase. SEQADDR(0x8a) SCSIRATE(0xc2) 
Nov 20 08:43:45 dagdhasrv kernel: ^INo terminal CRC packet recevied
Nov 20 08:43:45 dagdhasrv kernel: (scsi0:A:0:0): parity error detected
in Data-out phase. SEQADDR(0x55) SCSIRATE(0xc2) 

Any idea ?

Thanks - Didier

RE: Pb between an UltraTrak SX8000 and an Adaptec 29160

[Alan Cox]

On Wed, 2002-11-20 at 13:53, didier wrote:
> Nov 20 08:38:49 dagdhasrv kernel: SCSI device sda: 1171874688 
>                           512-byte hdwr sectors (-499511 MB)
>                                                  ^^^^^^^^^^^

That does look like a math bug in the driver but it would only mess up
printing.

> Nov 20 08:43:45 dagdhasrv kernel: ^INo terminal CRC packet recevied
> Nov 20 08:43:45 dagdhasrv kernel: (scsi0:A:0:0): parity error detected
> in Data-out phase. SEQADDR(0x8a) SCSIRATE(0xc2) 
> Nov 20 08:43:45 dagdhasrv kernel: ^INo terminal CRC packet recevied
> Nov 20 08:43:45 dagdhasrv kernel: (scsi0:A:0:0): parity error detected
> in Data-out phase. SEQADDR(0x55) SCSIRATE(0xc2) 

and this still appears to be broken hardware

Re: Pb between an UltraTrak SX8000 and an Adaptec 29160

[Simon Burley]

Alan Cox wrote:

<snip>

> Linux 2.4.x doesnt support single volumes over 1Tb. Thats fixed in 2.5

Can I ask what the symptoms of this failure to support >1Tb volumes
would be? Failure to identify the device? Random read/write failures?
Completely corrupted disk? It's the first time I've heard this!

The reason I ask is because we have a Storcase IDE to SCSI raid
enclosure with RAID controller here containing 8x Maxtor 160Gb disks
(configured as one RAID 3 volume), and it has been 100% reliable since
day one. No strangeness, failures of any sort, errors in syslog etc.

Vendor: DS570     Model:                   Rev: 0001
Type:   Direct-Access                      ANSI SCSI revision: 03
scsi2:A:6:0: Tagged Queuing enabled.  Depth 253

(scsi2:A:6): 160.000MB/s transfers (80.000MHz DT, offset 31, 16bit)
SCSI device sdb: -2053816320 512-byte hdwr sectors (47958 MB)

So I see the same problem as Didier, but fdisk version 2.11f got the
size right under 2.4.19-ac4:

df -kl
Filesystem           1k-blocks      Used Available Use% Mounted on
/dev/sdb1            1102989912 847397312 199563908  81% /local


Or am I missing something?

Simon

Re: Pb between an UltraTrak SX8000 and an Adaptec 29160

[Mike Dresser]

On Wed, 20 Nov 2002, Simon Burley wrote:

> > Linux 2.4.x doesnt support single volumes over 1Tb. Thats fixed in 2.5
>
> Can I ask what the symptoms of this failure to support >1Tb volumes
> would be? Failure to identify the device? Random read/write failures?
> Completely corrupted disk? It's the first time I've heard this!
>

What about the software raid /dev/md0?

I remember vaguely about the group that was doing raid5 on 75 gig IBM
Deathstars(can't remember url) having issues with terrabyte volumes, and
doing a pair of half terrabytes instead.

The article is so old and out of date though

www.nobell.org/~gjm/linux/ide-raid/

Speaking of deathstars:

www-cdf.final.gov/offline/warnings.html

Note about deathstars, and about 3ware and WD drives. =)

Mike

Re: Pb between an UltraTrak SX8000 and an Adaptec 29160

[Alan Cox]

On Wed, 2002-11-20 at 15:22, Simon Burley wrote:
> Can I ask what the symptoms of this failure to support >1Tb volumes
> would be? Failure to identify the device? Random read/write failures?
> Completely corrupted disk? It's the first time I've heard this!

There is a fundamental 2Gb minus a bit limit (2^32 sectors). One or two
bits of the kernel (LVM1 I believe is one) use signed maths here and
there. So 1-2Gb generally works but its not something to guarantee.

2.5.x supports wy more

RE: Pb between an UltraTrak SX8000 and an Adaptec 29160

[Andre Hedrick]

Didier,

I am waiting to close a contract to fix the problems w/ Promise products,
and yes I have the contract but have not signed the agreement to do the
work.  The issue is how they do their i2o v/s scsi.

Cheers,

Andre Hedrick
LAD Storage Consulting Group

On Mon, 18 Nov 2002, didier wrote:

> 
> 
> The UltraTrak was updated to the latest available firmware
> before connected to the server.
> 
> I e-mailed promise tech-support last week but did not get
> any answer so far.
> 
> On a different note, who should I contact regarding the display
> bug:
> 
> Nov 14 11:04:35 dagdhasrv kernel: (scsi0:A:0): 160.000MB/s transfers
> (80.000MHz DT, offset 62, 16bit)
> Nov 14 11:04:35 dagdhasrv kernel: SCSI device sda: -1951217920 512-byte
> hdwr sectors (100488 MB)
>                                                    ^^^^^^^^^^^
> Nov 14 11:04:35 dagdhasrv kernel: Partition check: 
> Nov 14 11:04:35 dagdhasrv kernel:  sda: sda1
> 
> Thanks again, Didier.
> 
> > > I have just connected a brand new UltraTrak SX8000
> > > (from Promise with 8 x 200GB IDE drives) to an
> > > adaptec 29160. The UltraTrak SX8000 is connected
> > > via a brand new LVD cable (1meter).
> > >
> > > While everything seems fine, I am getting a bunch of
> > >
> > > (scsi0:A:0:0): parity error detected in Data-out phase. 
> > SEQADDR(0x56)
> > > SCSIRATE(0xc2)
> > >         No terminal CRC packet recevied
> > 
> > I've never seen false positives from this diagnostic.  It 
> > looks like your Promise device is not properly implementing 
> > DT transfers.  Is there a firmware update for the device?
> 
> -
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

RE: Pb between an UltraTrak SX8000 and an Adaptec 29160

[Alan Cox]

On Mon, 2002-11-18 at 20:07, Andre Hedrick wrote:
> 
> Didier,
> 
> I am waiting to close a contract to fix the problems w/ Promise products,
> and yes I have the contract but have not signed the agreement to do the
> work.  The issue is how they do their i2o v/s scsi.

The i2o stuff works. Not that the cards I've seen perform very well.
SX8000 is an external raid box. SX6000 is the i2o scsi card which is
supported for some random subset of firmware versions by me and in part
with their help and assistance. 

Alan