I built and ran the program, and it says: "sg driver version: 30124". So
I guess we are using the latest one. At this point, I could try running
no more than two drives on one PCI bus (currently, they're all sharing
the same PCI bus but only two per channel on the dual card), and have
the picker use one of the free channels. I guess it comes down to two
choices since we only have two cards:

A. Have all drives sharing one dual Adaptec card (two drivers per
channel), and picker on its own separate card, in which case picker is
on its own bus and drives are on second bus.

B. Move drives 3 and 4 to other card so drives 1 and 2 are on one bus
and drives 3 and 4 are on second bus. Picker can then share free channel
on one of the two buses.

Not sure which scenario is worse, but we've tried A and it still didn't
fix things, so I guess there's no harm in trying B, and as Ernst
Bokkelkamp pointed out in earlier post, IO could be getting flooded?

George

Byron Servies wrote:
>
> George,
>
> Attached is a short program that will print out the version
> of the sg driver in your system:
>
>         $ cc -o sgvers sgvers.c
>         # sgvers
>         sg driver version: 30124
>
> It needs to be run as root, unless you have changed the permissions
> on your sg devices.  Hopefully, what it does is clear, but in short
> it opens a /dev/sg0 (or other specified device) and tries to send it
> the SG_GET_VERSION_NUM ioctl.
>
> HTH,
>
> Byron
>
>   ------------------------------------------------------------------------
>
>    sgvers.cName: sgvers.c
>            Type: Plain Text (text/plain)

--
Note: To sign off this list, send a "signoff networker" command via email
to listserv AT listmail.temple DOT edu or visit the list's Web site at
http://listmail.temple.edu/archives/networker.html where you can
also view and post messages to the list.
=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=