[Orca-users] Re: disk run doesn't show all disks

[legrand at speakeasy.net]

Mike, Blair


I just got to reading the orca-discuss list last night.

Having c0t15d0 show up as sd14 is the correct behavior for
my hardware raid controllers.  Sun has a bug (feature) that prevents
the virtual disks on my hardware raid controllers , SUN A1000 EMC
Clarion FC4500 and maybe others from  showing up at all with the SE
toolkit that Orca uses.

With the RAWDISK extention All of my raid disks show up as sdnn
instead of as cxtyd0. I am a little lazy.  To map them to the long
name would have required parsing the /etc/inst_to_path file
for any disks that failed to be resolved by the OS.

The OS is suppose to do this but it apparently does not for raid
controllers. When I talk to Sun about iostat -x -n core dumping
(the -n maps the short device name (s14) to the long device name
(c0t0d0) they told me to  apply a patch (I forget which one but it
replaced the iostat binary so that it no longer core dumped and (the
raid disks to show up as  sdnn with the -n option.)

They also told me that Solaris 2.6 and 2.7 would not be fixed
to work properly.  But that Solaris 2.8 would be (this was for
a Sun A1000 raid controller.)  I made sure to report the error
for a hardware/software combination that sun supported.   Instead
of it not working with the Orca/SE toolkit.)

Mike, exactly what combination of software and Hardware are you using?
Solaris?
Se Toolkit?
Orcallator script?

and is your disk drive real or is is a virtual disk on a raid
controller.

Alan
















--- In orca-discuss at y..., Mike Jackson <misterj at s...> wrote:
> Blair Zajac wrote:
> >
> > Mike,
> >
> > Could you try the attached code to help track down the problem you
> > we're seeing?
> >
> > From the author of the new RAWDISK routines:
> >
> > "I am enclosing the code that I used to debug the RAWDISK
routines.  It
> >  emulates either isostat or sar -d (I am not sure which.)
> >
> >  If you could get the person to run it we could isolate if the
problem
> >  is with my code."
> >
> > Regards,
> > Blair
>
> Hi Blair,
>  I will see if I can test it tomorrow. FYI, the cot15d0 did appear
as a
> graph the beginning of today after I changed the RAWDISK option as
you
> advised. I suppose the sd14 graph will just disappear once it
expires,
> in another 12 hours or so.
>
> Thanks,
> Mike