> For high availability, yes. But for performance, I was *under the > impression* that Linux needs to be fooled into using the multiple > paths (haven't been able to confirm this with end-to-end performance > tests). This is done by LVM or raid-tools striping (RAID 0). *As I > understand it* the "fooling" works as follows - when a striped > volume is detected, the Linux kernel will continue with data > transfers before the previous one finishes. Then the multiple > I/O paths to the DASD will be utilized.
Hmm. This sounds kind of like the tricks you used to have to do under early 390 releases with defining multiple addresses for the same device with different control units to fake the system into initiating more than one I/O to the same device at the same time (sounds maybe/kinda/sorta what PAV is supposed to accomplish). I would suspect that this probably won't be too effective on the 390 because the details of the I/O system aren't directly visible to the Linux driver -- you'd have to do some strange IOCP things to get it to present the same device with multiple devnums so that this gadget could try to do it's stuff in the kernel. Do your disks have PAV capability? If so, try turning on this widget against a PAV volume and see what happens. -- db
