>Casper.Dik at Sun.COM writes:
>> (With Windows, Ubuntu and one Solaris partition, typical people can't spare
>> another 15GB for an alternate boot environment on their systems)
>
>What are you talking about?
>
>My ABEs are rarely more than 5GB -- and that's being generous, as I
>expect the OS to grow quite a bit over time.
All systems I have use 10GB because that's the minimum I found
upgrades to work in; I am just allowing for some future growth.
I've had upgrades fail and require considerable cleanup with 10GB
root partitions. So I consider 15GB "ample, but not excessive"
and 10GB "the absolute minimum".
>You don't need to copy all of your data as well. Just the OS.
>
>> And, as far as I can tell, in-place upgrade is nothing more than:
>>
>> - boot from network/DVD media
>> - establish existing root as "alternate liveupgrade root"
>> - perform liveupgrade
>
>The "perform liveupgrade" step is pfinstall. Yes, and that's
>basically what it does.
>
>Minus, of course, the miniroot construction, testing, and the support
>to make sure that it works from release to release.
But we must do the miniroot construction and testing anyway.
(And as of yet I have not dared throwing away the partition which contains
the grub menu even though it seems that ludelete would allow it; does
that work?)
The other issue I have with abandoning "jumpstart upgrade" for liveupgrade
is the fact that liveupgrade will be much more operator intensive.
For SPARC, upgrade is now as easy as:
ssh host "reboot -- net:dhcp - install"
for x86 it's slighly more difficult (but the systems with proper ILOMs
allow for "ipmitool chassis bootdev pxe; ssh host reboot")
Liveupgrade requires the following steps for each upgrade:
- ludelete
- lumake (requires figuring out which boot device to use)
- luupgrade (what where those paths again)
- reboot
Specifically step 2 and step 3 require considerably more thought than
before.
(Ignoring the fact that for many of the systems I use converting to
liveupgrade is not really an option) to begin with.)
Perhaps we can have a poll on OpenSolaris, see which percentage of installs
are liveupgradable.
My guess is that we'll do away with at least 80% of free testing we get
from Solaris Express/Solaris Express Community Edition. And I fear that
that is an optimistic guess.
I can live with "unsupported" because inter express upgrades are
unsupported now anyway. But just as now "unsupported, while accepting
and fixing reported bugs". Or perhaps community supported.
Casper
