windsor@adc.com (Rob Windsor) writes [...]
I don't think I've ever booted from floppy during an upgrade (except for that one time the upgrade triggered a filesystem-related bug -- it went corrupt and I had to run WAFL_check).
More to the point, when I have to perform a WAFL_check under the gun, I'd rather not have to dig around for a fistful of floppies required to boot the filer for maint. Now we are FOUR TIMES more likely to encounter an aborted boot due to a floppy infected with bitrot. It is time to whine.
With DOT 6.x (or maybe the corresponding firmware upgrades are what makes the difference) one can use control/C at the appropriate moment during booting from disc. As in Colossal Cave, "nothing obvious happens". :-) But when it gets to the appropriate point it now offers you the same menu as when booting from floppy, with options (1) to (5) ... and presumably the other magic words also work at this point!
This doesn't seem to be all that well documented.
(How many folks keep a box of fresh, new diskettes around specifically for new revisions of ONTAP? How many just grab the first N unassigned floppies they see laying around their desk? How many use one floppy and re-image it on-the-fly as the filer asks for it? :)
I recycle a set of floppies that I keep specifically for the NetApp filers. I have had to acquire some more because of the expansion in DOT size, which unfortunately is beginning to look like an instance of Moore's Law.
The real pain is testing: I don't want my disaster recovery to depend on a set of floppies that have never actually been booted from. But booting from floppy now takes so long that one has to schedule filer downtime for this: one can't get away with a "brief interruption to service" anouncement.
Chris Thompson University of Cambridge Computing Service, Email: cet1@ucs.cam.ac.uk New Museums Site, Cambridge CB2 3QG, Phone: +44 1223 334715 United Kingdom.