Rumor has it that on Fri, Jul 21, 2006 at 01:41:28PM +0200 Hannes Reinecke said: > IIRC this is > > 'DGC' 'DISK' > 'DGC' 'RAID 10' > 'DGC' 'RAID 5' > > Hrmph. There is one bit which doesn't quite work out. > While the hardware handler knows how to handle error codes and how to > switch > paths for a specific device, it doesn't know _when_ to switch it. > I don't think it's a clever idea to switch paths whenever you encounter > an > passive path. Seems like you could do a nice ping-pong that way ... Yes. I agree that having the handlers in SCSI is a better place for them. However, the architectural decision to have multipath knowledge at a higher level makes that harder. Some sort of interface into the scsi layer handlers from the dm/block layer would be more needed. That way dm or some piece that has the knowledge of the multipath situation can trigger the failover. There will still be a need to bubble up information from the decoded error codes so higher layers can make the correct decision. Such an interface would need to be in both directions. Cheers, Phil > > Cheers, > > Hannes > > - > : send the line "unsubscribe linux-scsi" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- Philip R. Auld, Ph.D. Egenera, Inc. Software Architect 165 Forest St. (508) 858-2628 Marlboro, MA 01752 - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
