Tuesday, July 3, 2012

Error when init 6 is executed

Hi Everyone 

I reimaged a intel box with solaris 10 x86 10/09 update 8 
with ZFS as filesystem 
Then i enabled security with user hardening. 

When I am rebooting a system with init 6 command im getting an error. 

bash-3.00# init 6 
propagating updated GRUB menu 
/bin/sh: /usr/lib/lu/lulib: not found 
bootadm: error propagating updated GRUB menu 

But system is rebooting fine 

I need to skip this error when system is rebooting. Where i have to modify 
to avoid this error. 
Please help me to resolve it quickly. 
You're right about the disk write comment, but we *never* use the 
shutdown command (pain in the neck to remember the options). Our command 
is 
sync;sync;init 6 (or 5) 
Works just fine - the sync commands flush the caches and it's less 
typing...
Its a bug. Make sure your OS is patched
I installed a latest cluster patch also... 

Is there any way to modify that?? 
Hi Everyone, 

I added SUNWluu and SUNWlur packages to avoid error but showing some 
output 
when init 6 is executed 

bash-3.00# init 6 
propagating updated GRUB menu 
File </boot/grub/menu.lst> propagation successful 
File </etc/lu/GRUB_backup_menu> propagation successful 
File </etc/lu/menu.cksum> propagation successful 
File </sbin/bootadm> propagation successful 

I want to disable this output. Please please help me to resolve this 
init 6 or reboot should be avoided in Solaris, Unix, or Linux 

On shutdown is the command you should use, it will still do the equivalent 
of a reboot, but ensures all disk writes are complete. It will ensure the 
system is brought down as gracefully as possible. I have seen many disk 
failures that could have been avoided had the user used shutdown instead of 
reboot. 

For Solaris the shutdown syntax for a reboot is: 

shutdown -i 6 -y -g 0 

You should do man on shutdown and look at the options, because you can even 
delay the command and send messages to users to ask them to get off the 
server. 

For Linux shutdown syntax is much different: 

shutdown -r now 

In this case is also has additional options, but this will

If you have trouble remembering the options, (the fact that they change 
between versions doesn't help) simply create your own alias or script. 
Still have had so many problems with users who rebooted that I wish the 
command did not exist. And I feel the same way about init 6. I have spent 
way too many hours trying to fix disks because someone did a reboot or an 
init 6 and a disk did not come back. If it was the root disk it is obvious, 
on others it may not be immediately obvious.
You will find that the Sun approved method of rebooting a machine is to 
use the init command. That said, I am a firm fan of shutdown command in 
preference to init command, or even worse, the reboot command, which is 
commonly cited for reboots when performing works in single-user mode. 

Either way, whether you use shutdown or init to reboot a box, before 
issuing the command you do need to either check that the system is in a 
safe state to enter a reboot, or ensure that the correct scripts are 
implemented in /etc/rcX.d (or SMF on Sol10) to ensure a safe shutdown of 
the machine. init 6 will run the shutdown commands, if they exist and 
are set to run at the correct runlevels. 
Older versions (Solaris <= 9) did not attempt to shutdown cleanly when running reboot, leaving your database or other application in an inconsistent state. 

Ancient and now unsupported versions (Solaris <= 8) did not sync filesystems when doing an init 6 or reboot, leaving your filesystems in an inconsistent state as well as your applications. 

It has been about 6 years since Solaris 10 was introduced, and it does not matter whether you use reboot, init 6 or shutdown -i6 to reboot, as they all do the correct things in the correct order now.

0 comments:

Post a Comment

 
Design by BABU | Dedicated to grandfather | welcome to BABU-UNIX-FORUM