[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference ssdevo::hsz40_product

Title:HSZ40 Product Conference
Moderator:SSDEVO::EDMONDS
Created:Mon Apr 11 1994
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:902
Total number of notes:3319

885.0. "Urgent question!" by TKTVFS::NAKAMURA_Y () Thu May 22 1997 03:57

	Hello,

	We recieved following information from ORACLE-Japan. And they will be
	notify the following information to the our customers.

	I know blitz TD# 2254. I bilieve its problem occurred only at mirrorset.
	But ORACLE said its problem occur at all, not only mirrorset.

	Is this following informatin correct?

	Best regards.
	Yasuhide Nakamura

ORACLE Information

>  Digital HSZ controllers cause record level corruption
>  
>  Recently there have been problems at two customer sites where Digital's HSZ 
>  controllers have caused record level corruption in Oracle 7 databases.  This 
>  condition is triggered by block sizes larger than 16 Kb being written.  The 
>  corruption does not jeopardize file structure integrity and does not log a 
>  hardware error via the operating system, thereby making it look like Oracle 
>  has written a corrupt record into the database. 
> 
>  Digital Field Service has a test which is run from the disk controller (HSZ) 
>  which will verify this condition is occuring.  There is also a patch available 
>  for the HSZ to correct the problem.  Customers were notified of the 
>  availability of this patch.  However, because the subject of the patch letter 
>  was "Technical Issue with Mirroring(Raid 1) on Array Controllers Running HSOF 
>  3.0 and 5.0", some customers may not have recognized the need to apply it to 
>  their systems.  This problem is not unique to mirrored HSZs.   
>                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>  If the customer has Digital HSZs, they should be asked to contact Digital 
>  Field Service and request that they run a test to verify the integrity of the 
>  HSZ cache and that they request the appropriate patch, HSOF V3.0Z-3 or HSOF 
>  V5.0Z-2 or HSOF V5.0J-3 or any other applicable patches, be applied to their 
>  hardware. 
T.RTitleUserPersonal
Name
DateLines
885.1SSDEVO::T_GONZALESThu May 22 1997 13:264
    This patch fixes a problem seen whan BBR takes place on a member of
    an HSZ mirrored set. It IS unique to HSZ mirrored sets.  If a customer
    has experienced database problems and mirrored sets are not
    invovled,then the problem is due to some other issue.
885.2Thanks!TKTVFS::NAKAMURA_YThu May 22 1997 22:017
	Thank you for your quick reply.

	We can explain the TD# 2254 to ORACLE engineer with confidence.

	Thanks.
	Yasuhide Nakamura