Re: How to check archived log files ?

DB_BLOCK_CHECKSUM is used for *DBWR* to check for data block corruption, not ARCH and redo.

What's more, it shouldn't be touched with a bargepole, for the simple reason that it cripples performance, computing a check sum on both the read and the write. If you want to do much the same thing in 8i without the performance-crippling effects, use DB_BLOCK_CHECKING instead.

The clue is in the parameter name: there is no such thing as a "block" of redo. LGWR flushes variable byte-size redo, not regular sized blocks -basically, whatever is in the log buffer that needs flushing, however big or small. That's why log_checkpoint_interval is measured in *operating system* blocks, not Oracle Blocks.

And as none of this helps our original poster determine whether there is corruption in the *log files*, as opposed to the data files, it's all a bit besides the point.

Regards
HJR

--
Resources for OracleT: www.geocities.com/howardjr2000
=========================================


"godmann" <allanwtham_at_yahoo.com> wrote in message
news:95cd51c.0110042041.61f84faa_at_posting.google.com...


> Hi there,

>

>   I remember from Oracle 7.2 onwards the DB_BLOCK_CHECKSUM can be set to

true.


> You can enable logfile block checksumming. Enabling log block checksumming

> will force Oracle to compute a checksum and place it in the header of each

> redo block. When the block is read by ARCH processs, the checksum is

checked.


> Archiving process will stop if there is checksum error due to corrupted

redo


> block.

>

> Allan W. Tham

> DBA