solaris 10 installation hangs at filesystem preparation stage

Dear all,

I am facing problems trying to configure Solaris 10 which was preinstalled on a Sun Ultra 45 workstation.

Passing the network configuration stage (some errors included) the installation simply hangs while trying to prepare the filesystem, here are the messages I got:

runCmd error in /usr/sbin/sysidput with output

Unable to run cmd /usr/sbin/sysidput

System identification complete.

mv: cannot access /a/lib/svc/method/fs-root.orig

Please wait while filesystems are prepared...

The installation hangs here. I left the process running overnight, but nothing happened, I had to open a new console and shutdown the system. This behaviour is 100% reproducible. Now I will try getting the CDs for installation.

How can I find out what is happening?

Thanks in advance

[839 byte] By [David.Guzman] at [2007-11-26 12:16:53]
# 1

I also recently purchased a new Sun Ultra 45 workstation that came pre-installed and am also having the same issue with the installation.

I booted the system up and completed the Installation menus as expected.

Once that was done a small window pops open and just sits there

mv: cannot access /a/lib/svc/method/fs-root.orig

Please wait while filesystems are prepared...

there are no files in /a which would cause the above error

I am now going to have to install using CD images which I downloaded.

This has not been a very pleasent experience

RobertCO at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 2

Let me add my name to the list.

My customer has purchased 5 Ultra 45 workstations, with Solaris 10 01/06 pre-installed.

They *all* exhibit this behaviour, as described. He asked me if this is a known issue, and there is nothing I could find on SunSolve about it.

I wonder if a Sun engineer can respond to this.

Michael Abboud

Alphawest Services Pty Ltd

mabboud123 at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 3

These forums are NOT Sun techsupport.

They are hosted on the Internet for user community discussions.

Although some company employees may occasionally join in,

they do so as individuals and not as company representatives.

If you want a response from Sun and guidance from their techsupport,

open a service case.Your warranty includes that privilege

http://www.sun.com/secure/contact/

rukbat at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 4

Thanks rukbat,

Will do - although if the same issue is appearing at three different sites, and is being discussed here at Forums, perhaps a response from Sun to the wider community wouldn't be a bad thing after all.

Michael.

PS: Not sure what your angle is on the company employee bit (are you a Sun employee?)

mabboud123 at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 5

The request was made in this thread, for a Sun engineer to respond to the issue.

I was simply trying to guide you that such a thing was not going to happen,

and why it wasn't going to occur. Such a response could be mistaken

as a company response instead of just a thought from a single person.

A support case is the best method to get accurate and accountable help.

Click my username, to glance at my forum profile.

( former employee )

rukbat at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 6
...Ah, Ok I understand now, thanks.I will go ahead and log a support call, and will see what the outcome is.Michael.
mabboud123 at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 7

Here is what Sun Support had to say:

===

Hi Michael,

Yes the customer has to rebuild the OS on these systems.

There is no bug id for this and as per our old cases notes this is a known issue on these systems.

Regards,

Hema

Michael Abboud wrote On 01/18/07 08:50,:

Thanks Hema,

So you are saying the customer has to rebuild his Ultra-45 himself?

Can you please pass any known bug ID about this issue, or further evidence that this is a known issue within Sun?

That would provide at least some consolation...

Thanks and regards,

Michael.

===

> ...Ah, Ok I understand now, thanks.

>

> I will go ahead and log a support call, and will see

> what the outcome is.

>

> Michael.

mabboud123 at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 8

Mike:

Take your support case and go back to them.

Request that it be escalated to a backline support engineer,

for a more understandable explanation.

Ask that it be given to a backline support person that is also customer-facing

(not all of them deal directly with customers).

Ask for a "Sun-badged engineer, not someone from a partner".

You need a better answer than that crap !

That original answer is just an attempt to tell you to go away.

rukbat at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 9

Hey all,

rukbat was right - I put the pressure on the Sun frontline support, and got a Sun backline response, below.

Thanks for the encouragement rukbat! - it paid off.

===

Engineering provided us with a workaround for this problem.

And our backline team have confirmed this works.

The workaround is to restore /lib/svc/method/fs-root to slice 0. Here's how:

Boot from slice 1 into single-user mode:

ok boot disk:b - shell

Mount slice 0 on /a. Replace c1t0d0s0 with the appropriate disk address on your machine.

# mount /dev/dsk/c1t0d0s0 /a

Copy fs-root from slice 1 to slice 0:

# cp -p /lib/svc/method/fs-root /a/lib/svc/method/

Optionally, you can check to see if a filesystem was created on slice 3. Again, replace c1t0d0s3 with the correct address on your system.

# df -k /dev/dsk/c1t0d0s3

If you don't see a valid usage report such as

Filesystemkbytesusedavail capacity Mounted on

/dev/dsk/c1t0d0s32173699299 2151961920%

go ahead and use newfs to install a filesystem:

# newfs /dev/rdsk/c1t0d0s3

newfs: construct a new file system /dev/rdsk/c1t0d0s3: (y/n)? y

You can now reboot from slice 0:

# reboot -- disk:a

===

Michael.

mabboud123 at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...
# 10
Happy to help.Thanks for including the workaround here, so that others need not go through what you had to experience.
rukbat at 2007-7-7 14:53:37 > top of Java-index,Archived Forums,Socket Programming...