Advantage of FORCE LOGGING over NOLOGGING

Question

Hi,
Can you please help me on the advantages of using force logging mode with a standby database and the effect of it in indexes etc. Also, it may help if you could also share ideas on difference between the two modes?
Thanks,
Jennah

Answer 1

<i>>>  Can you help me what factors would be sacrificed</i>
This really depends on your system, in most cases you will not be able to see a difference. However i did a small test:
- drop index, restart db
- create index with logging (measure time/redo size)
- drop index, restart db
- create index with logging (measure time/redo size)
Result:
logging - Elapsed: 00:02:40.68  / Redo size: 800mb
nologging - Elapsed: 00:02:20.29 / Redo size: 1.5mb
Here the full test:
[code]SQL> select a.name, b.value from v$statname a, v$mystat b where a.statistic# = b.statistic# and a.name = 'redo size';
NAME                                                                  VALUE
redo size                                                             28304
SQL> CREATE UNIQUE INDEX "SAPR3"."CDCLS~0" ON "SAPR3"."CDCLS"
("MANDANT", "OBJECTCLAS", "OBJECTID", "CHANGENR", "PAGENO")
  PCTFREE 10 INITRANS 2 MAXTRANS 255
  STORAGE(INITIAL 65536 NEXT 65536 MINEXTENTS 1 MAXEXTENTS 2147483645
  PCTINCREASE 0 FREELISTS 1 FREELIST GROUPS 1 BUFFER_POOL DEFAULT)
  TABLESPACE "PSAPCLUI" LOGGING;
Index created.
Elapsed: 00:02:40.68
SQL> select a.name, b.value from v$statname a, v$mystat b where a.statistic# = b.statistic# and a.name = 'redo size';
NAME                                                                  VALUE
redo size                                                         834714816
SQL> select segment_name, bytes/1024/1024 "Size_MB" from dba_segments where segment_name = 'CDCLS~0'
SEGMENT_NAME            Size_MB
CDCLS~0                     800
drop index / db restart here
SQL> select a.name, b.value from v$statname a, v$mystat b where a.statistic# = b.statistic# and a.name = 'redo size';
NAME                                                                  VALUE
redo size                                                             28992
SQL> CREATE UNIQUE INDEX "SAPR3"."CDCLS~0" ON "SAPR3"."CDCLS"
("MANDANT", "OBJECTCLAS", "OBJECTID", "CHANGENR", "PAGENO")
  PCTFREE 10 INITRANS 2 MAXTRANS 255
  STORAGE(INITIAL 65536 NEXT 65536 MINEXTENTS 1 MAXEXTENTS 2147483645
  PCTINCREASE 0 FREELISTS 1 FREELIST GROUPS 1 BUFFER_POOL DEFAULT)
  TABLESPACE "PSAPCLUI" NOLOGGING; 
Index created.
Elapsed: 00:02:20.29
SQL> select a.name, b.value from v$statname a, v$mystat b where a.statistic# = b.statistic# and a.name = 'redo size';
NAME                                                                  VALUE
redo size                                                           1520824
SQL> select segment_name, bytes/1024/1024 "Size_MB" from dba_segments where segment_name = 'CDCLS~0';
SEGMENT_NAME            Size_MB
CDCLS~0                     800[/code]