On 06/23/10 10:49 PM, James Carlson wrote:
Richard L. Hamilton wrote:
On 6/18/10 9:52 PM, Cynthia McGuire wrote:
Template Version: @(#)sac_nextcase 1.70 03/30/10
SMI
This information is Copyright (c) 2010, Oracle
and/or its affiliates. All rights reserved.
Seem to be missing some content here ... did you
perhaps forget to
forward something?
For what it's worth, there's already a UUID in use in
Zones to identify
Solaris instances. Is this another UUID?
To save people the trouble, I tracked down a link to the
full text:
http://arc.opensolaris.org/caselog/PSARC/2010/226/fasttrack.txt
Upon reading it, I was quite curious what the benefit would be,
although I could imagine that esp. in a testing environment (and
even otherwise), being able to match up a crash dump with a
syslog message might be handy.
It might be handy, but it shouldn't be Committed. If it were Committed,
then the empty "doc impact" section in that document is a syntax error,
as a Committed interface _implies_ the creation of a man page entry for
this.
>
I don't believe that anything in syslog should ever be considered as
anything better than "not an interface." It's historically a dumping
ground for free-form text, with the assumption that either humans or
system administrators will be reading it. If you do need a stable
interface, then something more formal (like ereports or kstats) will be
needed.
OK, agreed we should not consider committing any string in syslog.
(If all the author meant by saying "Committed" was "I have no intent of
changing this message," then make it "Not An Interface" and just never
bother changing it.)
Yep, that's my intention so I'd be happy to list this as "Not An
Interface". I'll update the materials and ask Cindi to include
those in the case directory.
Thanks
Gavin
_______________________________________________
opensolaris-arc mailing list
[email protected]
