Mark Knecht wrote:
...
> Hopefully these are more of what you are looking for.
Thanks for taking the time. This confirms one thing that I suspected,
i.e. that there are 8 timed-out commands at once whenever such an error
occurs.
(This is the full depth of sbp2's command queue. It means that the error
is not that sbp2 misses a completion notification from the device one
time or another, but rather that the device misses a notification from
sbp2 about the latest new command. But why...? After that, the SCSI
subsystem expectedly stacks more commands on top of the command which is
stuck in sbp2 until sbp2's queue is full and, I suppose, the device
still doesn't get to know of them or doesn't want them for some reason.
They all consequently time out.)
This is also what it looks like with the only affected HDD of mine,
provided I run it with serialization switched off. May or may not have
the same underlying cause as with your disk. My disk is based on a
rarely found TI bridge chip; it worked better with Linux from about a
year ago.
> Let me know what else might help.
...
In case I ever make any headway with this disk here I may be able to ask
for further diagnostics or tests of patches. Thanks,
--
Stefan Richter
-=====-=-==- -=== --===
http://arcgraph.de/sr/
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
|
