savecore is pointing to wrong disk

[mike worrell]

just rebuilt a b132l+ using a make_recovery from the b132l+ box needing to be replaced.
The old box had disk c0t6d0.... new box has disk c0t0d0... when save core runs at starup it fails saying c0t6d0 can't be found... where do i go to fix this.
thanks,
mike

[Steven E. Protter · Accepted solution]

Shalom Mike,

Ignite if the version has a new enough version has a keyboard interupt. It lets you take the installation and reconfigure the disk after boot.

Watch for it at console, hit the key when prompted and move to the disk configuration tab.

SEP

[Sameer_Nirmal]

Hi,

I presume you have configured dump device during rebuilt. In this case you just need to make "savecore" aware of new device. It is picking up wrong device ( from NVRAM/stable storage).

You need to take a look/examine at the console/syslog messages from the time when the system went down on account of panic etc where core dump occured to determine where the dump header is stored. From the messages, you will know the dump device and the offset used.I guess, if the dump device was pre-configured, then dump header might have been saved on it. Then use "savecore" command with -D & -O argument. Refer savecore(1M) man page.
With this , you could able to save current OS core dump.

To modify NVRAM/stable storage to have new device so that savecore will always use it in future, the system has to be TOC'ed. It has to be done very carefully.You need to sync the FSs and umount all possible file systems before TOC'ing the system.There is TOC button at the back of the server. Once stable storage is updated, that message won't appear again.

I guess it is good idea to do the whole activity in single user mode.

[mike worrell]

HI sameer,

I've never ran TOC... How do I do this, I just looked on the back of the server and didn't see anything to press. THe b132l+ is an older model unix workstation looks like a pc almost.

Thanks,

[Kent Ostby]

Mike -- I'm assuming this is an older box ... 10.20 or before since you are using savecore instead of savecrash.

To force a TOC on a B class, you need to press the recessed TOC button for one minute. This button is located between the monitor port and Mouse/Keyboard ports on the back of the b-class.

BEFORE you do this, you need to check to see if your lvlnboot -v data shows the correct disk.

If it doesn't you need to use lvrmboot and lvlnboot to rebuild it.

[mike worrell]

hi kent,
i just looked and I see the little button. going to try it after i get it to single us state

[mike worrell]

HI Sameer and Kent,

I did the TOC at single user state.
Rebooted afterwards and savecore works fine now. Thanks so much....
PROBLEM IS SOLVED!!