3 Dead Blade 100 CDROMs

I have a stack of 3 dead Blade 100s dredged from some back room somewhere that for some reason people want installed in their desktops instead of a new PC running Linux or BSD which will cost 1/3 as much and run 3X faster... :-(

Two of them don't boot and one boots but we don't know the root password.

On all three, when I try to boot a Solaris 8 CDROM with "boot cdrom" at the ok prompt, I get

Boot device: /<a href="mailto:pci&#64;1" target="_blank">pci@1</a>,f/<a href="mailto:ide&#64;d" target="_blank">ide@d</a>/<a href="mailto:cdrom&#64;1" target="_blank">cdrom@1</a>,0

Can't read disk label

Can't open disk label package

Can't open boot device

Is there a different command I need to use on the Blade 100 for a cdrom boot? Or do all three have a dead CDROM? Thanks in advance..

[916 byte] By [wsanders1] at [2007-11-25 23:06:44]
# 1
Well, probe-ide doesn't show anything except the system disk so I guess all 3 CDROMs are dead.
wsanders1 at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 2

Wiley,

Get the optical drive replaced with a generic 32x or 48x drive,

and install the OS that way (if you're not already running over to your jumpstart server).

Move the functional drive to the next system, and so on...

An end-user usually doesn't need a CD drive, once they have NFS connectivity.I survived for three years with an inoperable CD drive in my SB100 at work.

Lastly, use a newer distribution of Solaris 8.

The very first distro's, such as HW 10/00 or HW 01/01were <i>really</i> not suitable for those desktops.

Sol 8 HW 2/02 would be an excellent choice, or any Sol 9 or Sol 10.

--

and as for the root password issue, search these hardware forums,

keyword "root password"

The issue's been discussed here and on Google many, many times.

Bill at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 3
Bill,<i>An end-user usually doesn't need a CD drive, once they have NFS connectivity.</i>Maybe the SA has disconnected the data cable from the cd-rom drive. Opening the tray by pushing the button without power might be impossible ...Michael
maal at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 4

Michael,

The CD drives in early SB100's were very prone to just sponteously go bad.

One day, the drive is good.

The next day, the drive is bad, no warning.

Best proof would be to replace one of them with a known, good drive.No sense spending any deeper mental effort on them.

Once a system has an OS installed, and if there is NFS connectivity, they could mount a CD drive from a nearby system.We used to do that all the time, at work.

I'm sensing that Wiley has the experience and the resources to get some replacement drives, if he chooses to.

Bill at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 5

Unfortunately I ASSUMED it would take less time to install via a CDROM (the sol 8 7/01 distro should work) then to set up a new jumpstart server from scratch. Wrong.

Anyway now that I'm 1/2 way through debugging this seems to be a combination of pinched IDE cables (a Sun Blade 100 favorite mode of failure) and dead hardware. I can see both disks and the CDROM with probe-ide but still can't boot.

Also, the Sun IDE cables are CS (cable select) capable and the sun blade 100 doc says drives should have the cable select jumper installed. However, the CDROM seems to be hardwired as IDE master, so with a non-CS-aware bog standard IDE cable (all I have lying around) you need to remove the CS jumper on any slave drive which lives on the same bus as the CDROM.

Now I can see all 3 devices (2 disks and CDROM). The CDROM happens to be primary IDE (0) master, but when I try to 'boot cdrom0' the prom fails over to diag-boot from the net, meaning it now sees the drive but can't boot from it. So I really have a dead CDROM. Or else these boxes won't really boot from Sol 8 7/01. I guess I'll be setting up a jumpstart server.

wsanders1 at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 6

As I kind of expected to hear ...CDROM == failed.

Edit:and yes, the July 2001 distro would be fine.

It is newer than what is mention in the SSH link for that system

<a href="http://sunsolve.sun.com/handbook_pub/Systems/SunBlade100/SunBlade100.html " target="_blank"> http://sunsolve.sun.com/handbook_pub/Systems/SunBlade100/Sun Blade100.html</a>

Bill at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 7

Yep, I cannot get the CDROM to boot the box at any IDE address.

Unfortunately, the Jumpstart config I am working with bombs off the Blade 100 when it tries to start openwindows. "Invalid argument", The jumpstart is allegely built on Solaris 8 7/01 but who knows.

Time to download Solaris 10.

wsanders1 at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 8
Ah yes, 'boot net - install w' keeps openwindows from starting and bombing off on an incorrectly patched jumpstart server.
wsanders1 at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 9
Happy to hear there is some small successes.
Bill at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...
# 10

I now have quite a pile of dead Sunblade 100s and V100s that have died in exactly this manner:

- Secondary IDE disk starts getting flaky, CD still OK

- Replace disk

- New dIsk not visible, previously working CDROM not visible, even when each device is the only device present in the secondary IDE bus.

This points squarely at the IDE cabling. Nice to see my Sun boxes failing due to the same $2 parts that fail in my $200 PCs.

I am never putting an IDE-based Sun box into service again, and the ones I have are going to the trash pile.

wsanders1 at 2007-7-5 17:57:54 > top of Java-index,Sun Hardware,Workstations - General Discussion...