one hour checkpoints

From: amonte <ax.mount_at_gmail.com>
Date: Thu, 18 Jan 2007 12:17:37 +0100
Message-ID: <85c1fb130701180317rf5692d9j1fd3ff84f0aa1384@mail.gmail.com>

Hello all

I have a quite busy Oracle APPS database running in HPUX. Today during peak hour I saw this in alert

Thu Jan 18 10:19:50 2007
Beginning log switch checkpoint up to RBA [0x17d1.2.10], SCN: 0x0784.b14a52c1
Thu Jan 18 10:22:09 2007
Beginning log switch checkpoint up to RBA [0x17d2.2.10], SCN: 0x0784.b14a8673
Thu Jan 18 10:26:18 2007
Beginning log switch checkpoint up to RBA [0x17d3.2.10], SCN: 0x0784.b14aee22
Thu Jan 18 10:51:32 2007
Beginning log switch checkpoint up to RBA [0x17d4.2.10], SCN: 0x0784.b14d0ddb
Thu Jan 18 11:09:24 2007
Beginning log switch checkpoint up to RBA [0x17d5.2.10], SCN: 0x0784.b14ea5dd

.........................

Thu Jan 18 11:12:04 2007
Completed checkpoint up to RBA [0x17d1.2.10], SCN: 0x0784.b14a52c1 Thu Jan 18 11:16:19 2007
Completed checkpoint up to RBA [0x17d2.2.10], SCN: 0x0784.b14a8673 Thu Jan 18 11:16:45 2007
Completed checkpoint up to RBA [0x17d3.2.10], SCN: 0x0784.b14aee22 Thu Jan 18 11:18:22 2007
Completed checkpoint up to RBA [0x17d4.2.10], SCN: 0x0784.b14d0ddb Thu Jan 18 11:23:56 2007
Completed checkpoint up to RBA [0x17d5.2.10], SCN: 0x0784.b14ea5dd

From v$log_histroty I see the log switch frequency

select * from v$log_history
where FIRST_TIME > ('20070118 090000');

     RECID      STAMP    THREAD#  SEQUENCE#          FIRST_CHANGE#
FIRST_TIME          NEXT_CHANGE#


---------- ---------- ---------- ---------- ----------------------

--------------- ----------------

      6096  612181190          1       6096          8266491016370 20070118
090018    8266491515585
      6097  612181328          1       6097          8266491515585 20070118
101949    8266491528819
      6098  612181578          1       6098          8266491528819 20070118
102208    8266491555362
      6099  612183092          1       6099          8266491555362 20070118
102618    8266491694555
      6100  612184164          1       6100          8266491694555 20070118
105132    8266491799005
      6101  612185404          1       6101          8266491799005 20070118

110923 8266491936643

Redo log are 600MB each.

How can a checkpoint take so long?

Cheers

Alex

--

http://www.freelists.org/webpage/oracle-l Received on Thu Jan 18 2007 - 05:17:37 CST