lru latch

From: <Babu.Nagarajan_at_Cummins.com>
Date: Fri, 19 Jul 2002 10:13:21 -0800
Message-ID: <F001.0049CD46.20020719101321@fatcity.com>

All

We are testing out the scalability of our new GS160 server. We are trying to run multiple processes concurrently to see how the server/db is able to handle load.

Most of these jobs are very similar - sqlplus sessions selecting count(*) from different tables with a where clause. This is pulled directly of the cursors used by the application.

When run serially these jobs run in less than 1 minute - which is pretty good considering the size of the able and the where clause. If run simultaneously ieach degrades to around 4 min each.

I looked at the wait events. It is to be waiting on db_file_scattered_read, db_file_sequential_read and cache_buffer_lru_chain latch. the proportion of thse waits is almost equal - approx 30% of the wait time is in each of these.

I can understant the db_file waits - it has to get data from the disk and probably I need to tune the IO more like IO balancing etc. But I cant understand why Oracle is trying to get a LRU latch here.

I use the following query to tell me how much of the db buffers is still free

select decode(state, 0, 'Free',1,decode(lrba_seq,0,'Available','Being used'), 3,'Being Used',state) "Block Status", count(*) from x$bh group by
decode(state, 0, 'Free',1,decode(lrba_seq,0,'Available','Being used'), 3,'Being Used',state)
/

And this tells me that more than 50% of the buffers are "Available". Why do I see LRU latch waits?

Also the physical reads do not decrease - even if I run the same load test multiple times. First time I can understand it has to fetch the data from the disk
but subsequent runs should get it from the memory. I dont do full table scans so the data has no reason to go to the end of the LRU list.

We are running NUMA architecture with direct_io disabled on Oracle 8.1.7.2

Any thoughts/suggestions?

Babu

--

Please see the official ORACLE-L FAQ: http://www.orafaq.com
--

Author:
  INET: Babu.Nagarajan_at_Cummins.com

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
--------------------------------------------------------------------

To REMOVE yourself from this mailing list, send an E-Mail message to: ListGuru_at_fatcity.com (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing). Received on Fri Jul 19 2002 - 13:13:21 CDT