Re: Coalescing of Free Space

From: Badri Narasimhan <badri_at_concentric.net>
Date: 1996/12/04
Message-ID: <5840jb$b3g_at_herald.concentric.net>#1/1


Hi:

        Note also that even if u set pctincrease to a positive number, PMON will only coalesce 8 extents every time it wakes up. If the first tablespace it attacks (which is usually file # 0 - that is the system tablespace) has more than 8 uncoalesced extents, that will be the only one it coalesces! Thus, you will have to set PCTINCREASE for an extended period of time and let it sit there for it to work through all tablespaces.

-Badri

In article <32A442BD.4879_at_bulmer.com>, Peter Wales <paw_at_bulmer.com> wrote:
>Paul Osborn wrote:
>>
>> Hi:
>>
>> A friend of mine running Oracle 7.? on a Sequent box told me that
>> their site had a problem with Oracle coalescing free space.
>> Specifically, they found that Oracle did not coalesce free extents if
>> the associated tablespace had its PCTINCREASE parameter set to 0.
>> Once the parameter was set to 1%, Oracle began to coalesce the free
>> space.
>>
>> My question is: has anyone else had this experience?
>>
>> Regards,
>> Paul
>>
>> +--------------------------------------------------------------+
>> | Paul Osborn Menlo Software |
>> | posborn_at_menlosoftware.com http://www.menlosoftware.com/ |
>> | V: 415.324.1286 744 College Avenue |
>> | F: 415.324.1287 Menlo Park, CA 94025 USA |

 +--------------------------------------------------------------+

>> |
>
>This is certainly the case in ORACLE 7.1.6 under unix.
>PCTINCREASE = 0 means no coalesance
>PCTINCREASE > 0 means that PMON will attempt to coalesce free extents
>(that are next to each other physically) every 5 minutes. This is
>documented on ORACLE UK's OCIS system.
>
  Received on Wed Dec 04 1996 - 00:00:00 CET

Original text of this message