On 05/11/10 11:49 AM, Dave Miner wrote:
On 05/ 6/10 06:31 PM, John Fischer wrote:
All,
Sorry for the duplicate email. The file that was attached
is also located at:
http://hub.opensolaris.org/bin/download/Project+caiman/System+Configuration+Project/scsmfdesignv0.2%2Dnodename.odt
A couple more comments now that I've read the spec.
section 4: s/cvsd/cvcd/
Oye. How did I miss that one. Probably from my days of using cvs.
8.1: This isn't specific to first boot, but occurs on every boot.
Yes. Of course.
8.3: Would it be better to include the default value ("unknown") in
the manifest, rather than hard-coding it in the method?
That makes sense.
8.4.1: Currently this service never fails, so it's probably good that
you're proposing to actually check what "uname -S" returns since
failure to set the nodename could cause all sorts of interesting
results. However, SYS_NMLN is the value it actually checks against,
not MAXHOSTNAMELEN (which should be logically equivalent, honestly,
but for which you can find multiple inconsistent definitions in the ON
source). Fortunately, it's unlikely that anyone will provide a 257+
character hostname so the failure case is largely theoretical.
Yeap. I caught that one when I was working on the next rev. of the
document.
While I'm wary of scope creep, anyone who reads this and looks at the
manifest in the appendix might be prompted to ask about the domain
name, identity:node, and /etc/defaultdomain. Would it make sense to
just bundle these both up now?
Probably.
Dave
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
