> Thanks for pointing this out. It seems like there is
> some overlap with 
> fcinfo. Particularly around the -l, -D, -d and -O
> just in terms of 
> discovery. Since fcadm will be along shortly with the
> addition of NPIV 
> (http://www.opensolaris.org/os/project/npiv/) perhaps
> it would be better 
> to add ping support to fcadm. fcadm will be a
> superset of fcinfo and 
> would seem to be a natural fit for this type of
> functionality.
> 
> I also recall some standards work in T11 that was
> pushing towards a more 
> comprehensive storage network ping (snping) standard
> CLI but I haven't 
> seen anything concrete yet.
> 
> - John
> 
> Richard L. Hamilton wrote:
> > "ping" for Fibre Channel; might be nice to
> have...works on Solaris.
> > Been awhile, but don't recall any major problems
> compiling or anything.
> > 
> > http://www.estorian.com/fc_ping.php
> >  
> >  
> > This message posted from opensolaris.org
> > _______________________________________________
> > storage-discuss mailing list
> > [email protected]
> >
> http://mail.opensolaris.org/mailman/listinfo/storage-d
> iscuss

What had originally interested me in fcping was situations where
there's a power hit and the system comes up before the SAN does;
that tends to leave a bunch of systems in single-user mode because
of failed fsck, requiring manual intervention (usually just run them
manually or reboot again, since by the time manual intervention
happens, the SAN would probably already have been up again).

What enables that is similar to certain usages of ping: being able to
detect success or failure by the return code; thus, one might write
a script that runs very early in boot (before all but the first few
filesystems (typically on internal disks) get mounted) which could
use something like fcping to test access to the SAN drives, and
sleep and retry for some reasonable amount of time before giving
up.  That could avoid a lot of unneeded manual intervention in
such situations, getting systems up faster, and sometimes
preventing someone from getting called in.

Note: that scenario was pre-Solaris 10; I haven't looked or tested
to determine whether SMF would have similar problems in that
situation.
 
 
This message posted from opensolaris.org
_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to