[Stgt-devel] [Scst-devel] Integration of SCST in the mainstream Linux kernel

[Pete Wyckoff]

mangoo at wpkg.org wrote on Tue, 05 Feb 2008 18:04 +0100:
> Anyway - if tgtd can be only killed with KILL signal - isn't it risky to 
> do so? For example, something may be still cached, not full transferred? 
> Will SCSI stack take care of it properly?

I too would like a clean shutdown mechanism.  For testing and system
scripts.  Target should go away cleanly:  drop incoming tasks, wait
for pending ops to finish and send responses, drop existing
connections, close 3260, flush logs, exit.  It may be somewhat
complex to get right.  Recently iscsid grew a SIGTERM handler too.
See abd1595d358b1cfe6c059aeea74f6ecdc748f461.

		-- Pete