[CentOS-virt] iscsi target rescan means connections hang for five seconds

Tue Jan 3 15:03:22 UTC 2012
Dennis Jacobfeuerborn <dennisml at conversis.de>

I'm starting to experiment with iscsi based volumes and while this works 
fine so far I ran into a problem. When I add a LUN to a target and refresh 
tgtd all i/o to that target seems to hang for five seconds.
This is obviously less than ideal if the plan is to use the target for lots 
of guests. There is also the possibility to create one target with one LUN 
for each guest disk but that sounds like a management nightmare and the 
libvirt pool functions become useless.

Is there a way to make this less intrusive? How do you guys cope with this