From: Weiping He on
One SUN V880 server got some error message like
"panic[CPU7]/thread=140a000 Could not find major number for 'rootnex'"
during applying latest Solaris patches.

The machine keeps reboot then.
I run hardware diagnosis and could not find any error.

Can somebody help me with this?

Remove CPU7 will solve the problem?

Thanks a lot!



From: Andrew Gabriel on
In article <d0hocd$h4h$1(a)solaris.cc.vt.edu>,
"Weiping He" <whe(a)microstrategy.com> writes:
> One SUN V880 server got some error message like
> "panic[CPU7]/thread=140a000 Could not find major number for 'rootnex'"
> during applying latest Solaris patches.
>
> The machine keeps reboot then.
> I run hardware diagnosis and could not find any error.
>
> Can somebody help me with this?

Sounds like /etc/name_to_major is corrupt or missing.
Boot from CDROM (or network), mount the root filesystem,
and see what's in the file. (The act of just mounting the
root filesystem might fix it in any case, if the problem
is due to an unrolled transaction log.)

> Remove CPU7 will solve the problem?

No. The thread reporting the problem happens to be running
on CPU7, but CPU7 is most unlikely to be the cause of the
problem.

--
Andrew Gabriel
From: Weiping He on
Thanks for you reply!
Yes!
That's the answer!
/etc/name_to_major is blank.
How can I restore it then?

Thanks

"Andrew Gabriel" <andrew(a)cucumber.demon.co.uk> wrote in message
news:d0hqn0$dk1$1(a)sparta.btinternet.com...
> In article <d0hocd$h4h$1(a)solaris.cc.vt.edu>,
> "Weiping He" <whe(a)microstrategy.com> writes:
> > One SUN V880 server got some error message like
> > "panic[CPU7]/thread=140a000 Could not find major number for 'rootnex'"
> > during applying latest Solaris patches.
> >
> > The machine keeps reboot then.
> > I run hardware diagnosis and could not find any error.
> >
> > Can somebody help me with this?
>
> Sounds like /etc/name_to_major is corrupt or missing.
> Boot from CDROM (or network), mount the root filesystem,
> and see what's in the file. (The act of just mounting the
> root filesystem might fix it in any case, if the problem
> is due to an unrolled transaction log.)
>
> > Remove CPU7 will solve the problem?
>
> No. The thread reporting the problem happens to be running
> on CPU7, but CPU7 is most unlikely to be the cause of the
> problem.
>
> --
> Andrew Gabriel


From: Mark Round on
Backup ?

Or you could specify what version of Solaris you're running, and
someone here could post a copy of their file, assuming that's the only
thing that got wiped. Do you know how it got deleted or overwritten ?

If it's more serious, and the whole SUNWcsr package (which includes
/etc/name_to_major on my Solaris 9 box) got removed, you could try
re-adding the package. Assuming you've booted using the Saolrsi CD-ROM,
and have mounted the root slice under /a, you could do something along
the lines of :

# pkgadd -d . -R /a SUNWcsr

Although after doing this, you'd have to make sure you re-applied any
patches that touched SUNWcsr.

-Mark

From: Mark Round on
>Assuming you've booted using the Saolrsi CD-ROM,

Duh. I obviously meant "Solaris CD-ROM"... Time to put the keyboard
down, step away from the root prompt and get some caffeine :)

-Mark

 |  Next  |  Last
Pages: 1 2
Prev: oracle solaris inst tmp problem
Next: ssh on Solaris 10