VClone with Oracle

[Mr. Basem Abdulla]

When we do a Vclone of a disk that contains oracle database 9i and we leave the disk for sometimes we can not start the oracle instance, although this happens in random times. many times it works but sometimes we find block corruption on oracle system table spaces, we noticed that if there is any kind of leveling on our EVA4000 after the clone or on other process there is a big possiblity of oracle data corruption , Can anyone support ?
[Attached Doc files with detials]

[Ivan Ferreira]

Hello Basem, welcome to the forum.

The attached file shows the problem when you try to shutdown the database, but it's not clear the "vclone" process involved.

Can you describe what you are doing and trying to do? What is the purpose of the clone? How do you clone the disks?


Please also ensure to read:

http://forums1.itrc.hp.com/service/forums/helptips.do?#28

[Amar_Joshi]

Basem,

It's almost next to impossible that a releveling will corrupt the data. It's not that I am passing on a warranty statement but trying to drive you in the right direction.

In a nutshell, I understand the following Database backup process:

1. DB should be in hot backup mode or shutdown.
2. Create a VClone of Vdisks.
3. Bring back the DB from hot-backup mode to normal.

In hot-backup mode, database actually sync the tables and put them in state where they have consistent data for backup purposes. In my discussions with DBAs, they commented that if the DB is not in backup mode and DB is not performing any activity and no pending transactions are held in buffer, a VClone backup of it will be readable. But if in most of the cases DB keeps tables open and performing transaction, any VClone created at this time will be a "Fuzzy" backup and can't be used.

Please confirm, if you are putting the DB in hot-backup mode or not?

my 2 cents worth..

[Mr. Basem Abdulla]

Dears,
Thanks for your replies ,, Well let me explain the process as follow :

1. We shutdown the database on Production A
2. We make a VClone, and we present the clone desk to another machine Test B and mount the drive and start the oracle instance form the Vclone disk.
3. we perform the backup from the Test B machine because it takes a long time. while Productio A machine is back online.
[The above sequence is done every week]

The incidint is sometimes when we try to try to startup the database on the Vclone disk, we find the corruption that i sent earlier, in other cases the database startup ok but after days we try to shut it down it gets that corruption. There are some scripts that shuts down the database, and sometimes we try to shutdown mannually.
[I will send ZIP file contains the 2 cases with Solaris and Oracle logs]
We verify oracle datafiles by running the verification tools and it reports no errors but when we shut it down we recieve that error.
REPLY ..

[whiz_kid]

SQL> shutdown immediate [Normal shutdown command without script files]
ORA-01122: database file 3 failed verification check
ORA-01110: data file 3: '/u01/oracle/demodata/system03.dbf'
ORA-01251: Unknown File Header Version read for file number 3
SQL


i guess ur system03.dbf is already corrupted and ur doing a vclone for a corrupted datafile that would obviously get cascaded to vclone.