Ashton Fagg <[email protected]> writes: > Ooh, yeah actually you are probably right. > > I also misread your initial email - you said "iscsictl should wait and > fail", I read it as "iscsid should wait and fail". ;-) > > Sorry for the confusion. I'll experiment a bit and see if I can come up > with a patch. I should not code before coffee. :-)
Yep, so just checked this out. Adding "sleep 5" to the rc.d script does indeed bring about a clean boot. I re-did my experiments and yes, it appears that the second start of iscsid fails with the device busy error - as expected, my diff does nothing for that except make it fail 15 seconds later than without. :-) The "sleep 5" thing is good enough for now, at least. I'd prefer to *not* have to remember to fix that after every snapshot upgrade, so I'll see if I can figure out the iscsictl fix. - Ash
