Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
![]() |
![]() |
Home -> Community -> Usenet -> c.d.o.server -> Re: Why a high INSERT activity requires to set a low PCT_USED ?
PCT_USED determines when a block goes *ON* the freelist. Not when it
comes off. If the actual % used in the block falls below the PCT_USED
threshold, it gets added to the freelist. On a table with a lot of
deletes and inserts into the same block(s), setting a high pct_used can
cause freelist managment overhead. That said, I have not personally seen
freelist management cause significant delays. That's doesn't mean it
doesn't happen though.
-- Chuck Hamilton Brian Peasland <oracle_dba_at_remove_spam.peasland.com> wrote in news:3F02EDEF.43C2ADB_at_remove_spam.peasland.com:Received on Wed Jul 02 2003 - 11:23:08 CDT
> Where did you read that?
>
> PCT_USED determines when a block leaves the freelist, or list of blocks
> that can accept a new row of data. If PCT_USED is low, then this block
> leaves the freelist earlier than a table with a higher PCT_USED. If you
> have a lot of inserts, then one would probably want to reduce the
> overhead of moving lots of blocks off the freelist. So a high PCT_USED
> should be warranted. But if there are lots of concurrent transactions
> performing the inserts, then it is possible that multiple transactions
> inserting into the same block can cause more overhead in dealing with
> all of the read consistency that may accompany the operations. So IMO,
> YMMV. I can see both sides of the coin.
>
> HTH,
> Brian
>
> Spendius wrote:
>> >> (that's what I read somewhere) >> Why should a high PCT_USED have a negative impact on >> performances in a table that undergoes lots of INSERTs ?? >> >> Thanks. >> Spendius
>
![]() |
![]() |