Re: Does an ALTER SYSTEM CHKPOINT write all dirty buffers to datafiles ?

On 15 Jul 2003 09:44:42 -0700, Joel Garry wrote:

>richard.foote_at_bigpond.com (Richard Foote) wrote in message news:<69f6c1c8.0307141921.45b1c3e_at_posting.google.com>...
>> joel-garry_at_home.com (Joel Garry) wrote in message news:<91884734.0307141442.7b7ff68a_at_posting.google.com>...

>> If the system goes down without finishing the checkpoint it does not
>> mean "you are set up to lose stuff". It simply means Oracle can't
>> guarantee that all changes associated with a previous redo log have
>> been written to disk therefore the previous redo log would be required
>> for instance recovery as well as the current log. *No commited data
>> can be lost* as Oracle guarantees all commited data to be written to
>> the redo log files. If a redo log to be overwritten has not yet had
>> it's corresponding checkpoint complete, Oracle simply waits until the
>> checkpoint is complete to ensure that instance recovery can always be
>> completed. Oracle prefers to hang the system rather than risk loss of
>> data.
>
>So you are saying Oracle waits during a shutdown abort?

When a shutdown abort, Oracle stops all. The recovery will take place at the startup. In case of a shutdown abort, the checkpoint is not finished but the data not in the datafiles are still in the redolog files and they will be read at the startup. What Richard says is that the data in the redolog will never be overwritten if they contain data that are not written in the datafiles (i.e. that are only in the buffer cache). A shutdown abort will not erase any data in the redolog and Oracle is still able to ensure that instance recovery will be completed.

[...]  

>> Redo logs are only the " Achilles' Heel " if you lose all members of
>> the current redo log group, which is quite a different scenario (and
>> one which should never occur if the redo logs are configured
>> appropriately).
>
>A stronger argument can be made about modern storage devices lying to
>Oracle about whether a disk has actually been written to. Which, as
>you say, is a configuration issue.

I don't agree when you say that modern storage devices are lying to Oracle. In fact, they are saying that they ensure that the data will be written on disk later (by using NVRAM or using a autonomous battery). And that what Oracle is waiting. Nothing more. Received on Tue Jul 15 2003 - 12:44:18 CDT