As far as disk ownership changes go, you should be able to just
do a disk reassign –s oldsysid –d newsysid for each head individually.
This will reduce the risk of missing disks as anything with the old sysid will
be reassigned.
All the rest looks alright (hopefully someone else will confirm),
though you will probably want to netboot the 3040 onto the correct version (7.2.2)
if it doesn’t already have it on the cf card.
Thanks
Oliver Bassett
Senior Systems Engineer
Infinity Solutions Ltd
P O Box 3323, Auckland
Ph: +64 9 921 8123, Mob: +64 21 574 994
Fax: +64 9 309 4142
www.infinitysolutions.co.nz
From: owner-toasters@mathworks.com
[mailto:owner-toasters@mathworks.com] On Behalf Of Jeff DiNisco
Sent: Wednesday, 9 May 2007 9:13 a.m.
To: toasters@mathworks.com
Subject: upgrading from 270c to 3040
I'm
considering upgrading from a 270c to a 3040 (all aggrs, no traditional
vols). I'm wondering about the proper procedure to swap the heads.
Here's what I'm thinking the steps will be, but I'd like someone to confirm or
let me know if there's a better way to go about it...
on 270c...
change root
vol size to accommodate 3040 (both)
upgrade ontap
to 7.2.2 (both)
modify /etc/rc
for different interface names (both)
record
ownership info, which filer owns which disks (both)
disable cluster
shutdown both
heads
replace EFH
mods w/ ESH2 mods
on 3040...
attach
shelves to both 3040 heads
boot 1 head
into maintenance mode
remove
ownership on all disks
take ownership
of only the disks that were previously owned by 1st filer as recorded in above
step
run
"storage release disks"
run
"mailbox destroy local"
shutdown head
boot other head
into maintenance mode
take ownership
of only the disks that were previously owned by 2nd filer as recorded in above
step
shutdown head
boot 1 head,
confirm things look good
boot other
head, confirm things look good
install 7.2.2
for 3040 architecture
apply new
licenses
enable/test
failover
I think that
might work, not sure. As I see it, the biggest risk is missing a disk or
2 when assigning ownership. My guess is the filer would attempt to
recover the aggregate by rebuilding on a spare. Is there an autorebuild
option I can turn off that would allow me to recognize that a have missing
disks, boot back into maint mode, and fix manually? Just trying to
uncover my biggest areas of risk in the procedure and mitigate them if
possible. Thanks.
JeffD