You and Ron are right. I was in a mindset of a totally clean environment
where DB2 would not have to go through its cleanup. But I can now see
where such a strategy might make business and technical sense.  

Still, I would rather hold DB2's ability to recover in reserve as
opposed to having to play my trump card every time. Besides, there might
be the rare situation were DB2 will not be able to recover. More, I
would not want folks to think that other DBMS's or applications could be
expected to recover the same way.

So, in a larger context, DB2, IMS, and such could be said to be
anomalies. 

I am wrong about DB2. But I still argue that recovery begins and ends
with the application.    

Hal

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On
Behalf Of R.S.
Sent: Thursday, July 14, 2005 3:50 AM
To: [email protected]
Subject: Re: DB2 and flashcopy2

Hal Merritt wrote:

> What about the work in progress in DB2 buffers that is not yet
committed
> (flushed) to DASD? The I/O might be consistent, but how would the
data?

Data written to DASD will be consistent. DB2 is *transactional* 
database. All uncommitted operations (transactions) will be rolled off.
It prevents DB2 database from data errors in case of hardware failure.
Consistent point-in-time copy will create similar image of DB2 datasets.

Data is consistent, but "consistency point in time" is a little bit 
earlier than time of copy. You cannot predict the delta. You can cause 
the delta to be zero: set log suspend.


-- 
Radoslaw Skorupka
Lodz, Poland

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html

Reply via email to