timeouts

["Julian Satran" <Julian_Satran@il.ibm.com>]

The timeout in 9.3 is for missing things
not connection teardown. I thought the text is explicit enough but obviously
it is not.

How about:

iSCSI recovery actions are often dependent
on iSCSI time-outs being recognized and acted upon before SCSI time-outs.
Determining the right time-outs to use for various iSCSI actions (command
acknowledgements expected, status acknowledgements, etc.) is very much
dependent on infrastructure (hardware, links, TCP/IP stack, iSCSI driver).
As a guide, the implementer may use an average Nop-Out/Nop-In turnaround
delay multiplied by a "safety factor" (e.g., 4) as a good estimate
for the basic delay of the iSCSI stack for a given connection. The safety
factor should account for the network load variability.  For connection
teardown the implementer may want to consider also TCP common practice
for the given infrastructure.

Julo