Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Chris Miller on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

initrd/ scsi/ aic7xxx drivres

Status
Not open for further replies.

sonun

IS-IT--Management
Dec 26, 2001
384
US
I was geting parity errors with scsi 29160 controller which was attached to a autolaoder from a 2.4.6 redhat.

Vendor: IBM Model: ULT3581-TA Rev: I105
Type: Medium Changer ANSI SCSI revision: 02
Vendor: IBM Model: ULT3580-TD1 Rev: 16E0
Type: Sequential-Access ANSI SCSI revision: 03
(scsi0:A:6): 80.000MB/s transfers (40.000MHz, offset 23, 16bit)
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x1a3) SCSIRATE(0x93)
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x1a3) SCSIRATE(0x93)
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x74) SCSIRATE(0x93)
(scsi0:A:6:1): data overrun detected in Data-in phase. Tag == 0x3.
(scsi0:A:6:1): Have seen Data Phase. Length = 256. NumSGs = 1.
sg[0] - Addr 0x37d6e040 : Length 32
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x1a3) SCSIRATE(0x93)
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x1a3) SCSIRATE(0x93)
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x74) SCSIRATE(0x93)
(scsi0:A:6:1): data overrun detected in Data-in phase. Tag == 0x2.
(scsi0:A:6:1): Have seen Data Phase. Length = 256. NumSGs = 1.
sg[0] - Addr 0x37d6e040 : Length 32
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x1a3) SCSIRATE(0x93)
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x77) SCSIRATE(0x93)
(scsi0:A:6:1): parity error detected in Data-in phase. SEQADDR(0x74) SCSIRATE(0x93)
(scsi0:A:6:1): data overrun detected in Data-in phase. Tag == 0x3.
(scsi0:A:6:1): Have seen Data Phase. Length = 256. NumSGs = 1.
sg[0] - Addr 0x37d6e040 : Length 32
Vendor: SEAGATE Model: ST39102LW Rev: 0004
Type: Direct-Access ANSI SCSI revision: 02
(scsi1:A:0): 80.000MB/s transfers (40.000MHz, offset 15, 16bit)
scsi1:0:0:0: Tagged Queuing enabled. Depth 253
Detected scsi disk sda at scsi1, channel 0, id 0, lun 0
SCSI device sda: 17783240 512-byte hdwr sectors (9105 MB)


Hence I thought I would create an 2.4.6 image as the /etc/lilo for the 2.4.6 kernel read,
image=/boot/vmlinuz-2.4.6
label=linux
read-only
root=/dev/hda2

And so, I did a,
/sbin/mkinitrd linux-2.4.6.img 2.4.6
No module aic7xxx found for kernel 2.4.6

I was not sure what it meant, so I rebuilt the krenel and turned off all support for the low level aic7xxx scsi drivers. But that was of no help either.
So what do I do.
Please advise.
Thanks a bunch.
 
Funny u should ask...I just encountered the same thing (almost) with a new Dell Poweredge 1650 server. Easiest fix: recompile your kernel and put the AIC7xxx drivers in the kernel.
 
I did that, but when I recomplie with that driver, then for some reason the ethernet interface refuses to come up.
But, why doe sit happen that whenever i try,
mt -f /dev/st0 status,
my computer just freezes.
Though the dmesg says it sees the st0 device.
Is it because it lacks the proper driver for it. If so, how should go about fixing the problem. Get a new kernel and then try ??
Please advise.
Thanks.
 
I am reposting some stuff from another forum "arkeia" which is a backup manger software. If you read through the bits, you will see my problem. Would gretly apprecuiet if anyone has the patience go thorugh this and provide any suggestion.

I am using the arkeia prof trial version for my IBM3581 LTO hosted on Redhat7.2(2.4.7-10).

My server sees /dev/st0 but does not communicte to it. My server does not see sg0 but arkeia doe and communicated to it. The trial license I get from arkeia inculdes the driver for the changer (sg0) only and not for st0 (tape drive). So arkeia leaves that the task of the actual backup is taken care of by the server's device driver (st0) while all arkeia does through its device driver (sg0) is to manage backups, keep logs, et al.
The server sees all the devices needed, but I suspect that it lacks the necessary driver which shold actually do the backup and this what even arkeia utilizes when it actually comes to writing data onto the tape ???

When I try to actually do the backup, it says it is loading this tape from this library and it loads the tape. But there is where all the action ends. I guess that is so as now arkeia is asking the servre to take over. Arkeia says to the server, ok now this is the tape I want teh abckup to be done, this is the ihformation I want to be backed up, this is the compresion required and so now you do the backup. The server then tries to, by whatever command is uses it actually uses to do the backup (tar, et al) write info to the tape and fails and my guess is that its is so, since the server doesnot have the apropriate driver to talk to the tape drive. I think the device is /dev/st0, but the driver for it is where I am running against a wall. I have fouhgt with different kernels, pathces and rebuilds but everywhere I fail at the same palce.
For now just tell me this I had talked to Sam Brown before when he was testing 3581 ( ijust want to know is if was using any redhat kernel and if so then what version of the kernel he was using)
The manual that came with the autolaoder says under supported OS and drivers tat currently autoloader is supported on redhat2.4.2 only !! and I wonder if this might be a problem !!!

The firmware when installed on the app kernel, creates its own special drives called /dev/IBMtape0. Does this mean that any new device (special,, not so ordinary like a IBm 3581 autoloader) that is attached a linux machine must have a new deive created for it to be able to comuunicate to it.

PLEASE ADVISE !!!!
.....................................
 
I always have better luck with patching a vanilla kernel, with the aic7xxx driver from here:


When you make the initrd image, if it complains that it cannot find "no module for aic7xxx found for kernel 2.4.x", you can do one of two things:

1. Check /etc/modules.conf for what 'alias scsi_hostadaptor' is, and change it to the module that is in: /lib/modules/2.4.x/kernel/drivers/scsi/aic7xxx

Which may be aic7xxx_old or aic7xxx_mod. If you change this in /etc/modules.conf to the proper file name, next time you make the initrd image, it should not complain.

2. You could rename the aic7xxx_mod file in that directory to aic7xxx and re-make your initrd image.

The aic7xxx driver from the above link is always the newest, and most updated driver which usually fixes a lot of issues. I have found it better than the older RH driver in my machine, but YMMV. Bruce Garlock
bruceg@tiac.net
 
you first do a
1. # modprobe st
2. # modprobe aic7xxx
3. # lsmod
To check that whether the modules are loaded or not.

 
Jodha, where do I try these coms,
on a normal m/c, I get the following o/p,
Module Size Used by
st 26464 0 (unused)
aic7xxx 102624 0 (autoclean)
sr_mod 11888 0 (autoclean)
scsi_mod 83584 3 (autoclean) [st aic7xxx sr_mod]
emu10k1 49872 0 (autoclean)
soundcore 3824 4 (autoclean) [emu10k1]
raid0 3616 1 (autoclean)
md 41824 1 [raid0]
usb-uhci 21504 0 (unused)
usbcore 28144 0 [usb-uhci]

But he thing is that if I want to tst this on a kernel I am trying to build, when and how do I do it ??

Also, I would greatly appreciate great links (in your opinion) to kernel building, pathcing, et al.

Thanks a bunch.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top