I had this problem once, it was caused by a faulty SCSI controller
configuration. What brand of controller do you have?
In my case, with an Adaptec 2940-UW controller, I was explicitly setting
the speed and not letting the controller auto-negotiate with the disk.
Once I set it correctly (auto-negotiation), the problem went away. A
call to Adaptec confirmed that the explicit setting should not be used
unless there is a specific problem. It is something to do with the SCSI
spec and the way the controller works. Not giving an explicit setting
forces the controller to negotiate with each drive. Unfortunately, there
is no way to query status from the controller after that so you have no
way to tell what setting it wound up using.
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]]On Behalf Of Matthew Coene
> Sent: Thursday, June 24, 1999 7:18 AM
> To: [EMAIL PROTECTED]
> Subject: SCSI Timeouts.
> Importance: High
> Sensitivity: Confidential
>
>
>
> Anybody have an educated guess as to why I might be getting
> huge amounts of SCSI bus timeouts?...
>
> I tried to install oracle onto a RAID-5 partition from a IDE CD-ROM.
> Everything is fine for a minute or two, then the errors start
> comming, and
> they keep coming for a another minute or two, then it locks
> my machine up
> solid.
>
> the errors look like:
>
> scsi : aborting command due to timeout : pid 5526, scsi 0,
> channel 0, id 8,
> lun 0
> Write (6) 06 b7 bf 80 00
> SCSI host 0 abort (pid 5582) timed out - resetting
> SCSI bus is being reset for host 0 channel 0
>
> Of course there are tonnes of these to reflect each of the 4
> devices on the
> scsi chain, the hex output is also different.
>
> Could this be because of the latency of trying to move stuff
> from an ide
> chain to the raid-5, or is this just another bug to be worked
> out of the
> code?
>
> Again...
>
> Any help would be appreciated...
>
> Regards,
>
> Matthew Coene
> -Systems Administrator
>
>
