Conference ssag::ask_ssag

6651.0. " mismathc between actual and restored configuration KZPSC " by EVTAI1::POUSSARD () Mon May 05 1997 11:00

	Hi,


	We have an important customer in FRANCE, which bought 40 KZPSC, but he
asked me for the following question


	We have a RAID 5 configuration and a disk which is a HSP disk



	Let's assume that a disk in the RAID 5 set becomes failed. So, the HSP
disk becomes a member of the RAID 5.

	Now, let's assume, the customer didin't still has time to change this
disk, and we have a controller failure.

	So, we put another controller and we restore the original configuration


	So, there is a mismatch between the original configuration, and the 
last up_to_date configuration after the RAID 5 member failure.

	So, the controller thinks always that the HSP is the original one and
declares this disk as a HSP despite it is, now, an RAID 5 member.

	Effectively, when we boot the DIGITAL UNIX OS and we wants to mount 
the RAID5 logical device, the mount is hanging.....
	The only way is to know what was happening before, and retire physically
the original failed disk...., so that the RAID 5 starts in degraded mode.


	The customer is very annoying about that and asks me if it exists a way 
to restart properly the system, ( even in degraded mode ) when there was a 
controller failure after a disk failure which changed the configuration.
	I know we can find what was happening before the controller failure by 
looking at the swxcrmon.log files, but is ther a more skilful method ??







	Thanks for your reply.....


				Gilles.

    	The answer is to replace the EEPROM. SEE 8.4.6 in user's guide