Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> v$buffer_pool_statistics: How to interpret data?
Hi,
I'm observing something which I cannot explain (Oracle 10.2.0.2 on 8-CPU HP/UX): I thought that v$buffer_pool_statistics contains upto date information about how many physical reads have been done since instance startup and how many reads in total have been done.
Now I select from the view and store the values in Excel. Then I do a FTS on a quite large table (about 12 GB in size, buffer cache only 2 GB), thus I would assume that after this query the physical reads and total reads increased by about 12 GB of data.
I assumed that the v$ view does not contain real time data so I waited about 5 minutes after the query to give it the time to reflect the new block reads.
But I don't see the data, and I see only a small increase in the physical reads, although the hard disk was used to 100% (using glance utility), so the query really used the hard disk, it really used a FTS and it really did not fit into the buffer cache.
Here is a table which contains the data from v$buffer_pool_statistics before (KEEP, RECYCLE, DEFAULT) and 5 minutes after the FTS (KEEP*, RECYCLE*, DEFAULT*):
NAME KEEP KEEP* BUF_GOT 3928800 3931367 FREE_BUF_INS 3642027 3644697 DIRTY_BUF_INS 2986 2986 DB_BLCK_CHG 3459789 3459789
PHYS_READS 3916261 3918828 PHYS_WRITES 86056 86056 NAME RECYCLE RECYCLE* BUF_GOT 12138557 12138590
DB_BLCK_CHG 47761840 47761840 DB_BLCK_GETS 47826166 47826166 CONSIST_GETS 10026356 10027706 PHYS_READS 11449803 11449836
NAME DEFAULT DEFAULT* BUF_GOT 53661160 53662236 FREE_BUF_INS 39830003 39830787 DIRTY_BUF_INS 744520 744520 DB_BLCK_CHG 246043831 246051058
PHYS_READS 47240215 47241120 PHYS_WRITES 6538609 6540360
Where is my mistake? I used to determine the buffer cache hit ratio in
a given time interval by querying this view and determine the
difference between the number of physical reads and total reads (i.e.
db_block_gets+consistent_gets).
But using the information above that seems to be wrong...(I know I
should not concentrate too much on buffer cache hit ratio)
Regards,
Alex
Received on Tue Aug 01 2006 - 07:33:17 CDT
![]() |
![]() |