On Mon, 05 Sep 2005 01:45:22 +0000, Mark Bole wrote:
> Mladen Gogala wrote:
>
>> On Fri, 02 Sep 2005 09:43:57 +0200, Sybrand Bakker wrote:
>>
>>
>>>I frequently observed log file sync problems when increasing
>>>log_buffer.
>>>This is logical as the log buffer is locked by LGWR when LGWR flushes
>>>the buffer to the online redo log file.
>>>IMO, you should decrease log_buffer to anything below 1M and you are
>>>shooting yourself in your foot now.
>>
>>
> [...]
>> In version 8, transactions would wait for the LGWR to flush the entire
>> buffer, but that is no longer the case. Transactions now wait until
>> only their data has been written to disk. So, large buffer is no longer an
>> obstacle. [...]
>
> I don't follow your logic here. What if the transaction's data is at
> the end of a large buffer?
>
> -Mark Bole
Actually, it was a word of mouth. My understanding is that log_buffer is
written in the several I/O requests and that transaction is notified as
soon as its commit record is written to the disk, not necessarily waiting
for the whole buffer to be written down. When I come to think of it, 10.2
event doesn't prove it as it would need a waiting address, too.
--
http://www.mgogala.com
Received on Mon Sep 05 2005 - 11:40:01 CDT
