Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Repairing the boot_archive

2 views
Skip to first unread message

Giorgos Keramidas

unread,
Apr 13, 2006, 7:15:31 AM4/13/06
to
After an unclean shutdown (due to power loss) one of our amd64 systems,
which runs Solaris 10 Update 1, can only boot in failsafe mode. If I
try to start it the normal way, it prints constantly an error about:

Short read
Short read
[...]

and hangs there. If there any way to detect why this happens (and fix
it, without a reinstall)?

- Giorgos

Casper H.S. Dik

unread,
Apr 13, 2006, 7:16:31 AM4/13/06
to
Giorgos Keramidas <kera...@ceid.upatras.gr> writes:

Boot the failsave archive; mount the root disk under /a,
remote /a/platform/i86pc/boot_archive and rebuild it using
bootadm update-archive -R /a

Why this happened is hard to tell.

Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

Giorgos Keramidas

unread,
Apr 13, 2006, 7:36:16 AM4/13/06
to
On 13 Apr 2006 11:16:31 GMT, Casper H.S. Dik <Caspe...@Sun.COM> wrote:
> Giorgos Keramidas <kera...@ceid.upatras.gr> writes:
>>After an unclean shutdown (due to power loss) one of our amd64 systems,
>>which runs Solaris 10 Update 1, can only boot in failsafe mode. If I
>>try to start it the normal way, it prints constantly an error about:
>
>> Short read
>> Short read
>> [...]
>
>>and hangs there. If there any way to detect why this happens (and fix
>>it, without a reinstall)?
>
> Boot the failsave archive; mount the root disk under /a,
> remote /a/platform/i86pc/boot_archive and rebuild it using
> bootadm update-archive -R /a

Fantastic, thanks :)

> Why this happened is hard to tell.

No problem. I'll look around to see what fsck changed, and now that the
system is back up, I can check most of the important files with the
checksums I kept last afternoon.

Thanks again :)

Casper H.S. Dik

unread,
Apr 13, 2006, 9:07:18 AM4/13/06
to
Giorgos Keramidas <kera...@ceid.upatras.gr> writes:

>Fantastic, thanks :)

Still, it's one of those "should not happen" scenarios.

(And if it happens more often, we need to make it easier for you
to recover automatically)

0 new messages