What's the output of metastat and metadb?

 *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
Octave J. Orgeron
Solaris Virtualization Architect and Consultant
Web: http://unixconsole.blogspot.com
E-Mail: unixconsole at yahoo.com
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*



----- Original Message ----
From: Christine Tran <[email protected]>
To: Liane Praza <liane.praza at sun.com>
Cc: Stephen Hahn <sch at sun.com>; smf-discuss at opensolaris.org
Sent: Friday, March 6, 2009 4:40:06 PM
Subject: Re: [smf-discuss] /lib/svc/method not mounted in time for SMF manifest?

> Could it be you're booting from a half of a root mirror that's not synced
> properly.  Have you checked that?

I've thought of that but there's no way we can tell.  The thing boots
up as a mirror, with the submirrors in maintenance.  There' s no way
to tell which side of the mirror we booted from, and even if so, after
several manual invocation of "metasync", it should be good.  The
problem, I think, is that metsync -r (called by mdmonitor) doesn't get
run (because mdmonitor can't find its method.)  It's really not a
"real" problem (in that the disks are out of sync), it's the the
process that's the problem.  Reboots have been done gracefully, as
init6, so there's no administrative reason the disks shoulld be in
some corrupt state.  -CT

> If it "shows up later" and you don't have
> /lib as a separate filesystem (and the file isn't a symlink), I'm having a
> hard time coming up with an alternate hypothesis but it's usually a
> transient condition that gets better once you've booted and the mirrors have
> synced.
>
> (No, I haven't used SDS recently enough to remember how to check that.)
>
> liane
>
_______________________________________________
smf-discuss mailing list
smf-discuss at opensolaris.org



      

Reply via email to