Solaris

Hi Everyone,

Thanks in advance for reading my thread. I am not very familiar with Unix so please bear with me if I say anything stupid.

I have a SunFire V100 server with Solaris 8, which I made into a Boot/Install/Profile server. Recently, I re-install the V100 box with Solaris 10.

The Solaris 10 installation had no errors. Same with the boot server installation. I retained the jumpstart directory ,/etc/hosts, and /etc/ethers from before the new installation (when it was still using Solaris 8).

I tried to jumpstart another V100 server with a Solaris 8 image. There were no errors during the entire process. However, when the jumpstarted server tried to boot after the jumpstart, the file system seems to be missing! It asks me to check file system manually, but it shows this error.

"Can't stat /dev/rdsk/c0t2d0s0".

The configuration for my network interfaces were also errored out.

"ifconfig: plumb: dmfe0: Bad file number."

If anyone have ANY idea as to how to fix this problem. Please tell me.

Thanks!
CityShrimp

Solaris 8 is fairly old, while a V100 is a fairly new server. It is possible your jumpstart server has a older version of Sol 8 which doesn't have hardware support for a V100. I've run into that situation in the past. The fact it can't find your hard drive at all and doesn't recognize the network interface address makes me think that might be your situation too.

Sun releases HW support updates for their operating systems. You'll see them listed like "Solaris 8 HW 02/02" for the Feb. 2002 release. Make sure you have the most recent one and see if that fixes your problem. The newest we have for Solaris 8 is 02/04, but I'm not sure if there is a newer one than that available.

Thanks for the reply rhfrommn!

I will try to get an image with a newer version of Solaris, but I am not sure if it's actually a hardware support problem since I could perform the jumpstart when my jumpstart server is using Sol 8.

But when I upgraded the jumpstart server to Sol 10, it fails. Even though everything else stays constant. I am still jumpstarting the same box, and using the same image as before.

Thanks!
Cityshrimp

Oh, I misunderstood your problem. Solaris 10 should definately have hardware support for a V100 in it so it isn't that.

One other idea I have - I see it isn't finding c0t2d0s0. Is it possible that device address is incorrect? Most solaris boxes I've adminned have c0t0d0s0 as the root partition. I know some of the newest sun servers have their internal controllers for fibre-channel disks laid out differently though. I think I've seen one machine where it was c1t0d0s0 instead. So maybe that is the error here - your jumpstart server is getting the device address different from what the server thinks it is.

That stuff should be in your jumpstart directory. The class (a.k.a. profile) and rules files are where you would look to see if any of that stuff is wrong now that you upgraded to Sol 10 instead of Sol 8.

The other possibility is the hard drive has failed. But that wouldn't explain not being able to find the network interface. That is why I think the devices under /dev being messed up on your newly jumpstarted server is more likely.

That looks correct to me, on the V100:

Code:

AVAILABLE DISK SELECTIONS:
       0. c0t2d0 <WDC WD400BB-22DEA0 cyl 19156 alt 2 hd 16 sec 255>
          /pci@1f,0/ide@d/dad@2,0

Try this:

boot in single user mode from the network ( boot net -s )
when the system is booted, mount the root disk on /mnt.
cd /mnt/etc and have a look at what you have in /mnt/etc/vfstab and compare that with you disk layout.
also ifconfig -a and check if your /mnt/etc/host and /mnt/etc/hostname.* corresponds.

Thanks for your reply, Reborg!

I tried what you asked, but I don't really see any problems. I'll post the output just to be sure.

Here's my disk layout in the profile.
# profile keywords profile values
# ----------------- -----------------
install_type flash_install
archive_location nfs://10.15.36.32/export/hom
ra9205_64bit_dmfe.archive
partitioning explicit
filesys c0t2d0s0 1400 /
filesys c0t2d0s1 1500 swap
filesys c0t2d0s5 2000 /var
filesys c0t2d0s6 10000 /opt
filesys c0t2d0s7 free /export/home

This is what's in /mnt/etc/vfstab after I mounted the root disck (/dev/dsk/c0t2d0s0)

#device device mount FS fsck mount mount
#to mount to fsck point type pass at boot options
#
#/dev/dsk/c1d0s2 /dev/rdsk/c1d0s2 /usr ufs 1 yes -
fd - /dev/fd fd - no -
/proc - /proc proc - no -
/dev/dsk/c0t2d0s1 - - swap - no -
/dev/dsk/c0t2d0s0 /dev/rdsk/c0t2d0s0 / ufs 1 no
-
/dev/dsk/c0t2d0s5 /dev/rdsk/c0t2d0s5 /var ufs 1 no
-
/dev/dsk/c0t2d0s7 /dev/rdsk/c0t2d0s7 /export/home ufs 2
yes -
/dev/dsk/c0t2d0s6 /dev/rdsk/c0t2d0s6 /opt ufs 2 yes
-
swap - /tmp tmpfs - yes -

The IP and hostnames under ifconfig, /mnt/etc/hosts, and /mnt/etc/hostnames.* corresponds with each other.

Thanks for all your help!
CityShrimp