I/O errors on 6TB device

Hi,

recently I asked about problems to expect with >2TB devices and the answer of Douglas made me hope we won't get any problems this time. Unfortunately, we get I/O errors on accessing the device.

The unit is called transtec PV610S, which is actually an Infortrend EonStor A16U-G2421-1 device.

Presently its running in test mode on our failover system with attached to a MPT scsi controller, after thouroughly tested it is supposed to become our main storage device on our primary server with an AIC79XX controller.

Without any knowledge about the scsi protocol and its error numbers it looks like the device is claiming to have more sectors than it actually has, doesn't it?

[17179724.816000] Fusion MPT base driver 3.03.07
[17179724.816000] Copyright (c) 1999-2005 LSI Logic Corporation
[17179724.832000] Fusion MPT SPI Host driver 3.03.07
[17179724.840000] ACPI: PCI Interrupt 0000:02:0a.0[A] -> GSI 24 (level,
low) -> IRQ 19
[17179724.840000] mptbase: Initiating ioc0 bringup
[17179725.312000] ioc0: 53C1030: Capabilities={Initiator}
[17179725.800000] scsi0 : ioc0: LSI53C1030, FwRev=01030600h, Ports=1,
MaxQ=255, IRQ=19
[17179726.064000]   Vendor: Transtec  Model: PV610S16R1B       Rev: 347G
[17179726.064000]   Type:   Direct-Access                      ANSI SCSI
revision: 05
[17179726.084000] sda : very big device. try to use READ CAPACITY(16).
[17179726.084000] SCSI device sda: 12691101696 512-byte hdwr sectors
(6497844 MB)
[17179726.088000] sda: Write Protect is off
[17179726.088000] sda: Mode Sense: 9b 00 00 08
[17179726.088000] SCSI device sda: drive cache: write back
[17179726.100000] sda : very big device. try to use READ CAPACITY(16).
[17179726.100000] SCSI device sda: 12691101696 512-byte hdwr sectors
(6497844 MB)
[17179726.100000] sda: Write Protect is off
[17179726.100000] sda: Mode Sense: 9b 00 00 08
[17179726.100000] SCSI device sda: drive cache: write back
[17179726.100000]  sda:<6>sd 0:0:1:0: SCSI error: return code = 0xb0000
[17179726.132000] end_request: I/O error, dev sda, sector 12691101688
[17179726.132000] Buffer I/O error on device sda, logical block 1586387711
[17179726.136000] sd 0:0:1:0: SCSI error: return code = 0xb0000
[17179726.136000] end_request: I/O error, dev sda, sector 12691101688
[17179726.136000] Buffer I/O error on device sda, logical block 1586387711
[17179726.136000] Alternate GPT is invalid, using primary GPT.
[17179726.136000]
[17179726.136000] sd 0:0:1:0: Attached scsi disk sda
[17179726.160000] sd 0:0:1:0: Attached scsi generic sg0 type 0
[17179729.412000] ACPI: PCI Interrupt 0000:02:0a.1[B] -> GSI 25 (level,
low) -> IRQ 20
[17179729.660000] mptbase: Initiating ioc1 bringup
[17179730.132000] ioc1: 53C1030: Capabilities={Initiator}
[17179730.620000] scsi1 : ioc1: LSI53C1030, FwRev=01030600h, Ports=1,

MaxQ=255, IRQ=20
/10

Now using parted to set a proper partition table, parted will complain about an I/O error and dmesg shows the messages below.

[17179939.464000] sd 0:0:1:0: SCSI error: return code = 0xb0000
[17179939.464000] end_request: I/O error, dev sda, sector 12691101688
[17180107.112000] sd 0:0:1:0: SCSI error: return code = 0xb0000
[17180107.116000] end_request: I/O error, dev sda, sector 12691101688

Any help is appriciated.

Thanks in advance,
Bernd

This is definitely a fusion driver error: it's DID_SOFT_ERROR, which that driver returns for a variety of firmware related conditions or transfer underruns.

I've cc'd the fusion people to see if they can help you diagnose it further.

James

James, many thanks for your help. I will try to do a firmware update on Monday, maybe that helps. Of course, we also appreciate any help further help from LSI/MPT.

Thanks again,
Bernd

PS: Since the controller hardware might be important now, its an onboard controller on a Tyan S2880 mainboard.

0000:02:0a.0 SCSI storage controller: LSI Logic / Symbios Logic 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI (rev 07)
Subsystem: LSI Logic / Symbios Logic: Unknown device 1000
Flags: bus master, 66MHz, medium devsel, latency 72, IRQ 19 I/O ports at a400 [size=256]
Memory at fc9d0000 (64-bit, non-prefetchable) [size=64K] Memory at fc9c0000 (64-bit, non-prefetchable) [size=64K] Expansion ROM at fc700000 [disabled] [size=1M]
Capabilities: [50] Power Management version 2
Capabilities: [58] Message Signalled Interrupts: 64bit+ Queue=0/0
Enable-

        Capabilities: [68]

0000:02:0a.1 SCSI storage controller: LSI Logic / Symbios Logic 53c1030 PCI-X Fusion-MPT Dual Ultra320 SCSI (rev 07)
Subsystem: LSI Logic / Symbios Logic: Unknown device 1000
Flags: bus master, 66MHz, medium devsel, latency 72, IRQ 20 I/O ports at a800 [size=256]
Memory at fc9f0000 (64-bit, non-prefetchable) [size=64K] Memory at fc9e0000 (64-bit, non-prefetchable) [size=64K] Expansion ROM at fc800000 [disabled] [size=1M]
Capabilities: [50] Power Management version 2
Capabilities: [58] Message Signalled Interrupts: 64bit+ Queue=0/0
Enable-

        Capabilities: [68]

Bernd Schubert
PCI / Theoretische Chemie
Universität Heidelberg
INF 229
69120 Heidelberg

-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Eric, thanks for your help. Its 2.6.16.36 on Debian Sarge, using most recent kernel version is not always a good idea on server systems - too many regressions.
Attached is dmesg output from 2.6.19.1.

Sure, its enabled now.

Until Monday I don't have physical access to the machine, so an update from DOS is not possible now. However, I see there is also a linux mptflash, I just can't get it compiled.

gcc -g -O -Wall -I.. -c -o mptflash.o mptflash.c
In file included from ../mptbase.h:58,

                 from mptflash.c:13:

../linux_compat.h:9:30: error: scsi/scsi_device.h: No such file or directory ../linux_compat.h:10:28: error: scsi/scsi_cmnd.h: No such file or directory

The kernel_headers in Debian don't have these files and including the header files directly from the kernel tree doesn't work.
Do you have newer sources (the most recent version I could find are from mptlinux-3.02.60) or a working binary?

Thanks,
Bernd

Bernd Schubert
PCI / Theoretische Chemie
Universität Heidelberg
INF 229
69120 Heidelberg