Introduction

Recently came across the following issue that a database is throwing  this message : Private strand flush not complete  during the switch of the redo logfile group. It is interesting to see the various opinions on this matter on the web, and in Mos. That is why i think taking inventory   and a plan might still be useful.

Details:

Database at hand is a 10.2.3.0 EE. on a Solaris box.

Looking under the hood i see:

SQL> select GROUP#,THREAD#,bytes/1024/1024 MB,members, STATUS from v$log order by 2,1 ;
GROUP#    THREAD#         MB    MEMBERS STATUS
 ---------- ---------- ---------- ---------- ----------------
 1          1       2000          2 INACTIVE
 2          1       2000          2 INACTIVE
 3          1       2000          2 CURRENT
 4          1       2000          2 INACTIVE
 5          1       2000          2 INACTIVE

Database is 1.24 TB

and looking at:

show parameter db_writer_processes
NAME                                 TYPE        VALUE
 ------------------------------------ ----------- ------------------------------
 db_writer_processes                  integer     2
Switches per hour:
Date            Switches
 ------------- ----------
 2013-03-13 02          8
 2013-03-13 03          7
 2013-03-13 04          3
 2013-03-13 05         13
 2013-03-13 06          9
 2013-03-13 07          2
 2013-03-13 08          2
 2013-03-13 09          9
 2013-03-13 10          1

According to MOS note Alert Log Messages: Private Strand Flush Not Complete [ID 372557.1] i should/could consider increasing the  db_writer_processes  at least from the current 2 to  4..  Which  sounds like a plan 2 Me.

Happy reading.

Mathijs