Alright we have a strange problem that i never encountered.

We have a V445 Sun Fire we are trying to install it via a tadpole laptop (jumpstart laptop) the v445 takes the flash just fine, but we get an error upon reboot saying unable to locate boot device. Well our SUN guru said we need to install patch 121690-04 to fix this problem well one way to install is copy the file to / the at the ok prompt type boot disk /flash-update-SunFire445

well we can not do this since the system can not locate the boot device so that we can load it. So what i did next was boot net -s so i can boot from the Jumpstart laptop then i did a mount /dev/dsk/c0t0d0s0 /var/tmp I mounted it here since i needed /mnt for the Jumpstart laptop directory where the patch files are located. so i now have the following

mount -F nfs -o rw 10.200.xxx.xxx:/jumpstart /mnt
mount /dev/dsk/c0t0d0s0 /var/tmp
I copied over all the patch information to /var/tmp
then did a chmod 755 flash-update* files in that directory of /var/tmp

since it stated i needed to have those files on the local disk in root.

well the .sh script included in that patch refers to directories that in a normal operating state. ie. /kernel/drv but on a mounted system for me it is /tmp/root/var/tmp/kernel/drv when i do a pwd in that directory

do i need to change all the variables and the normal areas to have /tmp/root/var/tmp in front of them for the system to accept the system firmware patch correctly? every time i run the .sh i get "Could not find a writable driver location" i believe it is because it is trying to write to the local temp OS and not the real os hard drive locations.

Have you tried booting to Solaris CD and just loading an OS on the server, then updating firmware?
Sounds like your hypothesis is correct on the path for $XDRV, though.
I would copy the .sh to a backup file then edit it and change that variable to your pwd results for /kernel/drv.
Be careful though - this file contains the microcode to download to the prom, and you could lose your prom if this file gets corrupted.
Good Luck & please post back your findings.

Instead i opted for the safer route. I went in and loaded solaris 10 on the v445 since it would take even longer for me to load solaris 9 with all the drivers it would need. then i loaded the firmware then i jumpstarted the unit and it worked just fine.
now my question is if i wanted to test the changes i made to that script and if it did mess up the firmware could i recover the system myself?