Jumpstart Enterprise Server - First Jump
Tuesday Jun 19, 2007
It has been on my to-do list for a long time to figure out how to build a jumpstart server. Well I just got a server to set up for a customer, went to slap the Solaris DVD into it, and low and behold, there is no DVD. Well guess I have just find my motivation to figure out!
My current enviroment is my trusty MacBook Pro. I am running VMWare Fusion Beta 3, and Nevada 64a.
Installing JET 4.3.3 to see if I can get a Jump Server up and running.
The directions say to add the package, so I have. The directions then say that everything will be in a directory called /opt/jet. The package add creates a directory called /opt/SUNWjet. This is the proper naming convention and the docs just me be off.
Directions now saying that the images need to be in /export/install. There are 3 sub directories that need to be created in that directory pkgs, patches, and media.
Next need to add these dirs to NFS mount. Directions say vi /etc/dfs/dfstab. When I open it, it says don't edit the file manually, but use the command sharemgr. MAN for sharemgr, leads me to the following game plan.
Create a group:
sharmgr create jet
Add a mount to the group:
sharemgr add-share -d "JET Packages" -s /export/install/pkgs jet
Wound up with an entry in dfstab added the other 2:
sharemgr add-share -d "JET Patches" -s /export/install/patches jet
sharemgr add-share -d "JET Media" -s /export/install/media jet
Now with everything happily in the dfstab, directions say to restart NFS. I am assuming that since there is a new command, no restart necessary. Low and behold I can get to the first share with my mac.
Just tried to mount the second share and got the spinning beachball of death. Waited over a minute and nothing. Can't do anything. Hitting the power button. So I guess I just restarted NFS anyways <grin>.
Back up and running. I have successfully tested all 3 mount points. When I went back into test, my last attempt was still there. I had misspelled patches. Looks like a mispelled mount . Who knows with beta, and I am not about to try it again.
Moving on, the docs say that the root password in the new server will be newroot. Making a note of now, because now way I am going to remember hours from now when the box is actually up and running, hopefully.
Next we need to get the installation media on the server. I am starting with Solaris 10 11/06 Sparc. Make sure you have the correct disk that you want. That mistake can cause hours of time trying to figure out why it does not work, and it can happen to the best of us. Mounting the Sparc disk and then executed the following command:
/opt/SUNWjet/bin/copy_solaris_media /cdrom/cdrom0
so the directions do say that this is a long process, but was not expecting 2 1/2 hours. I now have a directory under media called Solaris 10. It does not differentiate that it is Sparc. Will cross that bridge later.
Now we need to make a template for the server:
/opt/SUNWjet/bin/make_template <target server name>
Client template created in /opt/SUNWjet/Templates
Now we need to edit the template, at a minimum we need to set the following three variables. This includes my settings.
base_config_ClientArch=sun4u
base_config_ClientEther=0:3:ba:3f:bb:a7
base_config_ClientOS=10
I also edited the network settings section:
base_config_sysidcfg_ip_address=192.168.163.169
base_config_sysidcfg_netmask=255.255.255.0
Next we need to set up the Jumpstart server
/opt/SUNWjet/bin/make_client v240
A whole lot of stuff just happened. Looks like the tfpboot directory was created and the tftp service was started. Will have to see what effect this had on my Sun Ray server, later.
Now the tricky part is to get to the open boot prompt, through the ssh connection to the ALOM. The trick is to connect to the SC, boot the machine, drop back to the sc (#.), type the word break, and then console to get back to the console. This will send a break to the openboot and get you the ok prompt.
Once at the openboot prompt we can do the network boot:
boot net - install - w
aarrggh after all that getting an error:
message overflow on /dev/log minor #5 -- is syslogd(1M) running?
uncle google says to update the firmware.
Okay firmware updated making a stab at it again.
<sigh> same error. Calling it a night on the upgrade.
Working on this again. I read through the sharemgr docs http://docs.sun.com/app/docs/doc/819-1634/6n3vrg294?a=view I also noticed that the share that was created by the jet install added anon=0. This sets the user id unknown users.
I ran the following to set the following
sharemgr set -P nfs -p anon=0 jet
sharemgr show -p jet
jet nfs=(anon="0")
/export/install/pkgs nfs=(anon="0")
/export/install/patches nfs=(anon="0")
/export/install/media nfs=(anon="0")
This change allowed the jumpstart to start. I googled around and came up with this description of anon=0.
All basically said the anon=0 allows the client to read root owned files on
the mounted file system which in this case is the OS image and jumpstart needs
this.
At this point Solaris 10 is installed on my server and I can get to the task at hand with it!
|
|
|
|
