Re: lacie BIGGEST FW800 - SUCCESS!!

[Stefan R. <st...@s5...>]

Hugh Dixon wrote:
> Swtiched the Lacie Biggest box off (much bleeping ensued), switched it 
> back on and let it settle.
> Result:  Detected successfully....!

So the Biggest Disk's bridge was the culprit, not the host adapter.

> ieee1394 disable_irm=1 appears to be an absolute requirement.

I will get back to this eventually. We have to tweak nodemgr to
work without that parameter.

> noteably it's showing both the nonsense GUID and a 'real' one.

The real one is of course from the disk. The nonsense GUID won't
matter as long as you don't attach another LaCie card to the same
bus which features the same (then not exactly globally unique)
GUID.

> this is in a 33Mhz/32bit PCI slot - any suggestions on what to use to 
> test the data transfer rates in the current setup?

E.g. "hdparm -tT /dev/sdXY" for raw read spead of the block
device, bonnie++ for tests with filesystem. Something more
sophisticated (or hdparm or bonnie++ together with specifically
created partitions) to take physical layout on the disk(s) into
account.

> Stefan would you like me to send you the more verbose output?

If you kept logs from a session without disable_irm and one
with, I'd like to have a look at them.

> Question to ask is whether Lacie should really be able to fix whatever 
> freeze-up is happening I suppose?

It may be difficult for them to track down how the Biggest Disk
got into this state in the first place. We only know how the
state looked (disk answering with ack_busy_X all the time).
-- 
Stefan Richter
-=====-=-=-= -=== ==-=-
http://arcgraph.de/sr/