All of our recent Solaris 10 installs use the
ZFS root feature.  Hence we use luupgrade to
clone off a copy of root, patch that, then activate
and boot into the clone.  So downtime is just the
time needed for the reboot.  This part is similar to
what Linux people have been seeing for years.

What is improved for us though, is the ability to
"boot -L" and "boot -Z" to boot into previous boot
environments if there is some issue with a new
patch.  We do not need to untangle which patch
caused a problem in order to return to production

We actually use a local script called "patchbe"
which wraps some error-checking and other
functions around luupgrade capability.  We check
for pool integrity, run JASS on the clone, etc.

Switching to ZFS root and patchbe has taken out
nearly all of the worries of monthly patching.





Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to