sanity check on scsi terminator settings, please

sanity check on scsi terminator settings, please

Post by Corey Wiru » Sat, 15 Dec 2001 08:21:30



Hi All,

I've got a Compaq DeskPro with a integrated AM73C974 scsi controller.  The
problem is that I keep getting i/o errors with the disks while installing
RH7.2.  I heard once that wrong scsi terminator settings (or drive parity
settings) on the scsi chain can cause inconsistant reads and writes.  Is
that true?

I have on the cable (in order from the controller on out):

controller at id:12
scsi hard disk a at id:4
scsi CDROM at id:2
scsi hard disk b at id:0.

I have the terminator on disk b since it's on the end of the chain and it's
set to get terminator power from the drive.  There's no way I can tell if
the scsi controller itself has to be terminated - I doubt it.

I ran Seagate OnTrack Diagnostics on the two drives and there were no errors
on the surface scan.  My root file system is 1048Meg and this doesn't
include the various mounts on other partitions for /usr/local, /var, /tmp,
etc.  So, I should have LOTS of disk.

Any ideas would be appreciated!
Corey.

 
 
 

sanity check on scsi terminator settings, please

Post by Trevor Hemsl » Sat, 15 Dec 2001 17:10:49


On Thu, 13 Dec 2001 23:21:30 UTC, "Corey Wirun"


> Hi All,

> I've got a Compaq DeskPro with a integrated AM73C974 scsi controller.  The
> problem is that I keep getting i/o errors with the disks while installing
> RH7.2.  I heard once that wrong scsi terminator settings (or drive parity
> settings) on the scsi chain can cause inconsistant reads and writes.  Is
> that true?

> I have on the cable (in order from the controller on out):

> controller at id:12

Only wide devices can talk to ID's > 7. If all of your devices are
wide then you can leave it on 12, if some are narrow then you need to
change this to a number less than or equal to 7. The standard for
controllers is that they are set to id 7 since this is the highest
priority for bus negotiation. 7 is best even if all devices _are_ wide
(and I doubt they are).

Quote:> scsi hard disk a at id:4
> scsi CDROM at id:2
> scsi hard disk b at id:0.

> I have the terminator on disk b since it's on the end of the chain and it's
> set to get terminator power from the drive.  There's no way I can tell if
> the scsi controller itself has to be terminated - I doubt it.

Given that you can set the controller's id to 12, it must be a wide
controller? Is the last device on the bus also wide? If it isn't and
you have that one terminated then the extra 18 lines (difference
between 50 and 68) will be unterminated. If the controller is wide
then the termination will also need to be wide!

--
Trevor Hemsley, Brighton, UK.


 
 
 

1. RH9 FTP Install - Sanity Check Please!

All,

Has *anyone* completed a Red Hat 9 install via FTP?  I've done several
FTP installs of Red Hat prior to 9 but now I'm getting a ftp error.
It appears to be looking for a updates.img file which doesn't exist.
Please reply to this post if you have successfully done a Red Hat 9
install via FTP and if you're feeling extra kind a ftp site +
directory would be appreciated.

Thanks in advance.

Regards,
Dolpho37

2. RealTek 8029AS

3. Sanity check please: tunneling over TELNET

4. 2.2-970302-GAMMA install problems

5. Sanity Check on natd/ipfw please

6. root partition has already been mounted read-write. Cannot Check !!

7. : Sanity checking for drives that claim to be LBA-48, but aren't!

8. Help with EPIPE - broken pipe error ...

9. 5.0 sanity checks list

10. sanity CHeck

11. Booting from Compact Flash- sanity check

12. 32bit applications, 64bit kernel, dev_t issue(s) + sanity check (mine)

13. Mail Server on Local Network - Sanity Check