[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
| 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 |
Hi,
Some days ago I had a problem in a data warehouse machine with the following
configuration:
-AXP 8400 5/350 (4 cpu, 2 Gb memory....)
-8 HSZ40 with Write Back Cache in dual redundancy configuration
attached to 4 KZPSA
-Each set of two HSZ40 has 3 Raid 5 of RZ29B-VW
-Operating System:Digital Unix v3.2C
-LSM configured in most of the Raid 5 disk
-Oracle v7.2.3
-Firmware version HSZ40:v27Z (no patches installed)
HSZ40 Configuration:
2 Controllers (id 7 & 6)
SCSI id�s: 0,1 preferred to HSZ40 (id 7)
SCSI id�s: 2,3 preferred to HSZ40 (id 6)
3 units: D100,D200,D300 with Raid 5 (5 disks)
The problem was that after an Oracle�s user can�t access to a volume
associated to unit D100, DECevent(v2.2)reported a battery failure of
HSZ40 (id 7).The Oracle�s user tried to access to unit D100 and can�t
do it.The access to units D200 and D300 was Ok.
After the batteries were changed, following the procedure,the HSZ40
console gave a message about the cache had data not cleared previously.
The cache data was cleared using: CLEAR_ERRORS INVALID_CACHE NONDESTROY .
After this, all the units were available via HSZ40, not under O.S..
A message "Unit has lost data" was showed in the HSZ40 console for all
the units.
I made CLEAR_ERRORS LOST_DATA D100/D200/D300 and after that only the unit
D100 started the reconstruction.When the reconstruction finished, the units
were available by the O.S. but then the Oracle�s user detected that some
data had been corrupted.
So, my question is:
Why there was data corruption?
What is about WBC???
Regards
Fernando Roldan
[email protected]
[Posted by WWW Notes gateway]
| T.R | Title | User | Personal
Name | Date | Lines
|
|---|