RE: Some Latch Questions

Here is what I found in MetaLink:

Cache buffers LRU chain latch:..............................
        Increasing the parameters DB_BLOCK_BUFFERS and DB_BLOCK_WRITE_BATCH
        (undocumented in Oracle7) will reduce waits on this kind of latches.

From: Oracle, Helen Schoone <mailto:hschoone_at_us.oracle.com>23-Dec-99 14:54 Subject: Re : PMON failed to acquire latch, see PMON dump

Hi. Actually, this latch is not acquired based on an alter session command, we get this latch during user session initialization. We need to get the latch at this time because we are copying the parameter values which are stored in the SGA to private memory.

However, as stated by Bumby, this has usually occurred as a side effect of other contention which was resolved after tuning to resolve that issue. Regards,
Helen Schoone
Oracle Server EE Analyst

> -----Original Message-----
> From: Diego Cutrone [SMTP:dcutrone_at_afip.gov.ar]
> Sent: Monday, August 14, 2000 3:13 PM
> To: Multiple recipients of list ORACLE-L
> Subject: Some Latch Questions
>
> Hi,
> Forgive my poor english.
> An Oracle 7.3.4 (AIX 4.3.3) database is suffering of some "latch
> waits".
> Especially on "cache buffers lru chain " and some on "modify parameter
> values"
>
> What can I do about them?
> I think that the first latch is needed for the server processes in order
> to
> gain access to the LRU list in the buffer cache. Contention there means
> that there
> are a lot of processes that want free buffers.
> I think it would help if I increase db_block_buffers parameter. am I
> right?
> if not, What about increasing DBWR thoughtput ?
>
>
> What does it mean the second latch?
Received on Mon Aug 14 2000 - 15:46:56 CDT