If You break Sun Enterprise M3000/M4000/M5000/M8000/M9000 Servers Domain


Here I want share information about how to recover Sun Enterprise M3000/M4000/M5000/M8000/M9000 Servers Domain if you break it from Oracle/Sun Solaris 10 OS environment.

“Any word at all is a candidate.

Use words to see the entire dictionary”

 

This messages you can usually meet whet you trying in panic way to bring system back online ………

Now I will explain what exactly going on in similar scenarios.

This is example of log from XSCF of domain console:

Jan 17 12:07:35 WAT 2010      root@server1 # telnet server2 21
Jan 17 12:07:35 WAT 2010 Trying 10.10.10.10...
Jan 17 12:07:35 WAT 2010 Connected to server2.
Jan 17 12:07:35 WAT 2010 Escape character is '^]'.
Jan 17 12:07:36 WAT 2010 220 server2 FTP server ready.
Jan 17 12:07:41 WAT 2010 Type 'go' to resume             --------- “CTRL + \
Jan 17 12:07:49 WAT 2010 {ba} ok
Jan 17 12:07:49 WAT 2010 Any word at all is a candidate.  --------- “CTRL + ]
Jan 17 12:07:49 WAT 2010 Use words to see the entire dictionary
Jan 17 12:07:49 WAT 2010
Jan 17 12:07:49 WAT 2010 Any word at all is a candidate.  --------- “CTRL + ]
Jan 17 12:07:49 WAT 2010 Use words to see the entire dictionary
Jan 17 12:07:49 WAT 2010
Jan 17 12:07:49 WAT 2010 Any word at all is a candidate.  --------- “CTRL + ]
Jan 17 12:07:49 WAT 2010 Use words to see the entire dictionary
Jan 17 12:08:02 WAT 2010
Jan 17 12:08:02 WAT 2010 Any word at all is a candidate.  --------- “CTRL + ]
Jan 17 12:08:02 WAT 2010 Use words to see the entire dictionary
Jan 17 12:08:05 WAT 2010
Jan 17 12:08:05 WAT 2010 {ba} ok
Jan 17 12:08:06 WAT 2010 {ba} ok
Jan 17 12:08:24 WAT 2010 {ba} ok
Jan 17 12:08:25 WAT 2010 {ba} ok
Jan 17 12:10:43 WAT 2010 {ba} ok
Jan 17 12:10:43 WAT 2010 {ba} ok
Jan 17 12:10:43 WAT 2010 {ba} ok
Jan 17 12:10:46 WAT 2010 {ba} ok boot   ---------- necessary to type “go” but not “boot
Jan 17 12:10:46 WAT 2010 Resetting...
Jan 17 12:13:02 WAT 2010 .POST Sequence 01 CPU Check
Jan 17 12:13:04 WAT 2010 LSB#05 (XSB#05-0): POST 2.15.0 (2010/10/06 14:24)

 

1) To close connected TELNET session you must type “CTRL+]” to receive telnet> prompt.

But in really you typing “CTRL+\”  – and it generate break signal from Solaris OS and bring system to OpenBoot ok prompt.

 

2) in panic mode you even doesn’t reading console output still trying to close telnet session just typing “CTRL +]” several times ……

 

3) instead for type “go”, you type “boot” and rebooting the system ……

 

Why it can be happen ???????

 

Answer: it because the Mx000 domain security mode is switched off. If you will check from XSCF console:

XSCF> showdomainmode -d 01 
Host-ID :847c73cd
Diagnostic Level :min
Secure Mode :off
Autoboot :on
CPU Mode :auto

 

To solving this issue you must enable secure mode of Mx000 domain by command:

 

XSCF> setdomainmode -y -d 01 –m secure=on

 

Then you need to reboot a domain. You can do it from Solaris OS by “reboot” command or from XSCF controller by “reset” command.

After apply this change – the combination of keys “CTRL + \” will no more effecting to Solaris OS and will not break system to OpenBoot “ok” prompt.

 

Eldar Aydayev ©

UNIX Systems Professional Consultant | Aydayev’s Investment Business Group


1676. 23rd Ave, Noriega St. San Francisco, CA 94122


E-mail: eldar@aydayev.com


URL: http://eldar.aydayev.com


LinkedIn: http://www.linkedin.com/in/eldar


Phone: +1 (650) 2062624