/dev/rdsk/c0t2d0s0: no such device

Hi,

on my Sparc20 (Solaris 7) a problem occurred during bootup: After the
"The system is coming up" message the ufs filesystems check starts and
gives the message:
/dev/rdsk/c0t2d0s0: Kein derartiges Gerät oder Adresse (which means
there is no such device or address).
/dev/rdsk/c0t2d0s0: Can't check file system.
/dev/rdsk/c0t2d0s0: Unexpected inconsistency...

Then it waits for control-d or root password.
Pressing the return key just continues the boot process, and the
system comes up just fine, beside there is one harddisk missing.
I did not change anything on the system, and it run perfect before.

Does this mean the hard disk is broken? Even the probe-scsi output
doesn't show the missing disk/target.

(Far from being an expert, I thought that target 2 is related to the
scsi bus itself, not to distinct devices???)

Any hint is very much appreciated!

Regards, Pascal.

Pascal <> wrote:

> /dev/rdsk/c0t2d0s0: Kein derartiges Gerät oder Adresse (which means
> there is no such device or address).
> /dev/rdsk/c0t2d0s0: Can't check file system.
> /dev/rdsk/c0t2d0s0: Unexpected inconsistency...
>
> Then it waits for control-d or root password.
> Pressing the return key just continues the boot process, and the
> system comes up just fine, beside there is one harddisk missing.
> I did not change anything on the system, and it run perfect before.

You have removed that harddisk, but didn't remove its
/etc/vfstab entry. Now Solaris wants to mount a non-
existing disk. That fails of course and so you get into
single user mode. Correct /etc/vfstab.

> (Far from being an expert, I thought that target 2 is related to the
> scsi bus itself, not to distinct devices???)

In short:

c = controller
t = target == SCSI disk ID
d = always 0 for SCSI?
s = slice/partition

mfg
Dennis

P.S:
Unix auf Deutsch ist boese. Ich sag nur "Kein Weltraum links
auf dem Geraet". Ausserdem findet man Englische Fehlermeldungen
viel eher mit einer Suchmaschine Deiner Wahl.

--
from the manpage of cdrecord, section BUGS:

Cdrecord has even more options than ls.

Dennis Grevenstein <> wrote in message news:<chcc1s$ifs$>...
> Pascal <> wrote:
>
> > /dev/rdsk/c0t2d0s0: Kein derartiges Gerät oder Adresse (which means
> > there is no such device or address).
> > /dev/rdsk/c0t2d0s0: Can't check file system.
> > /dev/rdsk/c0t2d0s0: Unexpected inconsistency...
> >
> > Then it waits for control-d or root password.
> > Pressing the return key just continues the boot process, and the
> > system comes up just fine, beside there is one harddisk missing.
> > I did not change anything on the system, and it run perfect before.
>
> You have removed that harddisk,
As I pointed out, I did _not_ change anything on the system, i.e. I
did _not_ remove the disk!

> but didn't remove its
> /etc/vfstab entry. Now Solaris wants to mount a non-
> existing disk. That fails of course and so you get into
> single user mode. Correct /etc/vfstab.
>
That still would not bring the disk back to work, I suppose?!

Is the conclusion of a broken disk correct then?

> > (Far from being an expert, I thought that target 2 is related to the
> > scsi bus itself, not to distinct devices???)
>
> In short:
>
> c = controller
> t = target == SCSI disk ID
> d = always 0 for SCSI?
> s = slice/partition
>
> mfg
> Dennis
>
> P.S:
> Unix auf Deutsch ist boese. Ich sag nur "Kein Weltraum links
> auf dem Geraet". Ausserdem findet man Englische Fehlermeldungen
> viel eher mit einer Suchmaschine Deiner Wahl.
Tja, warum mitten zwischen den englischen Fehlermeldungen was
deutsches auftaucht, ist mir auch schleierhaft. Ich hab das System
nämlich englisch installiert. :-/

Pascal <> wrote:
>> You have removed that harddisk,
> As I pointed out, I did _not_ change anything on the system, i.e. I
> did _not_ remove the disk!
>
> Is the conclusion of a broken disk correct then?

Yes. The disk is dead (or cable/enclosure/etc, but something related to
the disk).

Scott.

Scott Howard wrote:
>
> Pascal <> wrote:
> >> You have removed that harddisk,
> > As I pointed out, I did _not_ change anything on the system, i.e. I
> > did _not_ remove the disk!
> >
> > Is the conclusion of a broken disk correct then?
>
> Yes. The disk is dead (or cable/enclosure/etc, but something related to
> the disk).
>
> Scott.

Cant honestly say "thank you" for that confirmation... ;-)

At least I am sure now.

Do you have any idea how to get a new/used internal disk for a
SparcStation 20 over here in Germany?

Cheers, Pascal.
--------------------------------------
Nicht die Zeit macht uns fertig, sondern das Tempo!

On Mon, 06 Sep 2004 22:43:25 +0200, pascal anschau wrote:
> Scott Howard wrote:
>> Pascal <> wrote:
>> >> You have removed that harddisk,
>> > As I pointed out, I did _not_ change anything on the system, i.e. I
>> > did _not_ remove the disk!
>> >
>> > Is the conclusion of a broken disk correct then?
>>
>> Yes. The disk is dead (or cable/enclosure/etc, but something related to
>> the disk).
>>
>> Scott.
>
> Cant honestly say "thank you" for that confirmation... ;-)
>
> At least I am sure now.
>
> Do you have any idea how to get a new/used internal disk for a
> SparcStation 20 over here in Germany?

If this is a Sparc-20, I thought that it's internal disks were at SCSI IDs
3 and 1? Yeah, (can't trust my flakey memory) I just checked: there's
target 3 and 1 for SCSI disk, and target 6 for CD-ROM. If you're having a
problem with target 2, that must be some external disk? Is there actually
a disk? or are you just assuming? Is that a typo in your /etc/vfstab file?

Dunno what availability is like in Germany or shipping stuff in, but there
are regularly SCA SCSI disks available for cheap on eBay.

--
Juhan Leemet
Logicognosis, Inc.

Juhan Leemet wrote:

>>Do you have any idea how to get a new/used internal disk for a
>>SparcStation 20 over here in Germany?
>
>
> If this is a Sparc-20, I thought that it's internal disks were at SCSI IDs
> 3 and 1? Yeah, (can't trust my flakey memory) I just checked: there's
> target 3 and 1 for SCSI disk, and target 6 for CD-ROM. If you're having a
> problem with target 2, that must be some external disk? Is there actually
> a disk? or are you just assuming? Is that a typo in your /etc/vfstab file?
>
> Dunno what availability is like in Germany or shipping stuff in, but there
> are regularly SCA SCSI disks available for cheap on eBay.

SS20 disks are SCA, 1" high. Just about any disk should work as long
as it fits in and has an SCA connector. AFAIK, the ID's will be
determined automatically by the bus.

/Marcin

On Thu, 09 Sep 2004 14:01:16 +0000, Marcin Dobrucki wrote:
> SS20 disks are SCA, 1" high. Just about any disk should work as long
> as it fits in and has an SCA connector. AFAIK, the ID's will be
> determined automatically by the bus.

Yeah, that's right about the IDs. My (implied) point was related to the OP
originally complaining about:

/dev/rdsk/c0t2d0s0: Can't check file system.
/dev/rdsk/c0t2d0s0: Unexpected inconsistency...

This shows Solaris checking SCSI target 2. The standard backplane wiring
of the Sparc-20 does not set anything to target 2. Hence, his file system
problem is likely either caused by:

1) some non-existant or failing external disk
2) typo in /etc/vfstab (I've done that myself)

I guess I was suggesting (in a round about way?) that he check the
/etc/vfstab file for typos (e.g. putting t2 instead of t3 or t1). I've
done that myself and stared at the screen with a dumbfounded expression...
until the penny dropped... Sometimes we see what we want to see, not what
is actually there. It would be a shame if the OP buys a new disk because
of a typo? OTOH, I have found that more disk(s) are always good to have.

--
Juhan Leemet
Logicognosis, Inc.

Hi folks,

thanks for all the comments.

.... The standard backplane wiring
> of the Sparc-20 does not set anything to target 2. Hence, his file system
> problem is likely either caused by:
>
> 1) some non-existant or failing external disk

That is correct. I was wrong when assuming an internal disk had been
broken; it is (actually, was) an external hard drive.

> 2) typo in /etc/vfstab (I've done that myself)

No typos were involved, as the problem came up after a few weeks of
downtime. The system worked properly before. I will unavoidably have
to accept that the disk is finally broken, I suppose.

> ... It would be a shame if the OP buys a new disk because
> of a typo?
Thanks for the care! ;-)

Cheers, Pascal.