Troubleshoot Solaris JumpStart

(Thanks google for putting this page as top 3 searching results!)


Solaris 10 Jumpstart Problems are typically in the following areas:
1. Boot Problems.
2. Identification Problems.
3. Configuration and Installation Problems.
4. Misc problems during the JumpStart.

Details:
1. Boot Problems
Boot process problems are normally associated with Layer 3 protocols RARP, TFTP or BOOTPARAMS configurations.

a. Resolving RARP Problems

If a jumpstart repeatedly issues the following message and failed during the boot process:

"Timeout waiting for ARP/RARP packet"
"Timeout waiting for ARP/RARP packet"
"Timeout waiting for ARP/RARP packet"
"Timeout waiting for ARP/RARP packet"
"Timeout waiting for ARP/RARP packet"
"Timeout waiting for ARP/RARP packet"
"Timeout waiting for ARP/RARP packet"
"Timeout waiting for ARP/RARP packet"
"Timeout waiting for ARP/RARP packet"
....

In general, this is because the JumpStart client can't obtain RARP service from a boot server.

Check the following:
1. cable or other physical connection problem? does the client use the right interface? (Check OBP NIC aliase, etc.)
2. is in.rarpd daemon running on the boot server?
3. does /etc/ethers and /etc/hosts or NIS have the correct entry for client's MAC,IP,hostname?


b. Resolving TFTP problems
If the client shows the "Timeout waiting for ARP/RARP packet" ONLY once and stops,
it indicates the JumpStart server may not be able to obtain TFTP services from a boot server.

Usually, this errory message indicates that there is no entry for the client in
/tftpboot/ directory (inetboot.hex_ip).

Solution: Remove the client and re-add it.
#/export/install/Solaris_10/Tools/rm_install_client client1
#/export/install/Solaris_10/Tools/add_install_client -c server1:/export/config \
-p server1:/export/config client1 sun4u

other possible causes:
incorrect platform group argument to the add_install_client script was used.
the boot server is not configured to allow the in.tftpd daemon to run on demand.


#inetadm | grep tftp
make sure tftp is enabled in inetd, tftp uses udp/69 and bootp uses udp/67

#inetconv
#inetadm | grep ftp

c. Resolving BOOTPARAMS problems

If the client passes RARP and TFTP, but stops booting after displaying numeric value,
such as 23e00, the client may not be able to obtain BOOTPARAMS information from boot server.

Solution:
c1. Check if the boot server's /etc/bootparams file has an entry for the client
c2. Check if the daemon rpc.bootparamd is running on the server.
If not,
Re-run add_install_client script to add client in /etc/bootparams file and restart
rpc.bootparams daemon.


To manually start bootparams (optional)
#svcadm enable network/rpc/bootparams:default
#svcs | grep bootparams

2. Troubleshoot Identification Problems

Problems in the JumpStart client identification process usually relate to errors in
sysidcfg file (e.g. no id info) or from a name service. If client cannot obtain a
response from a server for any identification item, the client interrupts the auto
identification process and asks from information.

check sysidcfg file if client stops and asks for any of the following info:
IPv6 networking
netmask
IP address of the default router
secuirty policy
name service
timezone
system locale
time server info
root password


If the JumpStart server uses NIS for some of above info, please check NIS.

Resolving Name Service Problems

If you use a name service, and the JumpStart client interrupts the identification process to obtain identification items other than the following, check the corresponding map or table information in the name service, and correct the problem you find:

* Will the client implement IPv6 protocols?
* What is the IP address of the default router?
* What security policy will the client implement?
* What is the root log in password?


The previous items can only be provided using the sysidcfg file.

You can use the sysidcfg file to provide information that a name service could otherwise provide. You must verify the content of the sysidcfg file or any information that it provides. Information provided in the sysidcfg file overrides information in name services.


3. Troubleshoot Configuration Problems
Configuration problems usually relate to rules and profile files.

If admin updated rules file and didn't to run check script to update rules.ok, client
will returns error:

"checking rules.ok file...
Warning: Could not find matching rule in rules.ok "

To fix, just verify rules file has an entry matches the client and re-run check script
to recreate rules.ok file.

(rules.ok file contains rules file and version info plus checksum.)

If profile contains configuration error (e.g. wrong slice in filesys field), clients
will show the error.

For example,

"ERROR: Field 2 - Disk is not valid on this sytem (c0t4d0s0)"

Verify the profile and enter the correct value will fix the problem.


4. Troubleshoot Installation Problems.

Installation problems usually relate to NFS issue.

If client passes RARP and TFTP, but panics

"panic - boot : Could not mount filesystem
Program terminated
ok"

this shows client (inetboot program) cannot mount the root (/) file system defined in /etc/bootparams file.

Solution:
verify /etc/dfs/dfstab on the boot has entries to share required directories;
check /etc/bootparams file to ensure entry is correct for the client, path for root is right.
check nfs is running on the installation server, if not manually start it.


#svcadm enable network/nfs/server:default


For begin and finish scripts, during installation,

check log files /tmp/begin.log, finish.log, install_log, and /var/sadm/system/sysidtool.log.

After installation, check
/var/sadm/system/logs/begin.log, finish.log, install.log and sysidtool.log



/tftpboot: Contains the boot image for the JumpStart software client
/etc/ethers : Matches client Ethernet address
/etc/dfs/dfstab : Provides configuration for network shared directories
/etc/hosts : Associates names with IP addresses
/etc/bootparams : Contains entries that network clients use for booting
/export/config/client/sysidcfg : Contains IPV6, defaultrouter, security, rootpw
rules : associates client host with profile