create table tmp_number
(
val_n number,
val_n8 number(8),
val_int integer,
val_f float
);
Insert into TMP_NUMBER
(VAL_N, VAL_N8, VAL_INT, VAL_F)
Values
(0, 0, 0, 0);
Insert into TMP_NUMBER
(VAL_N, VAL_N8, VAL_INT, VAL_F)
Values
(-1, -1, -1, -1);
Insert into TMP_NUMBER
(VAL_N, VAL_N8, VAL_INT, VAL_F)
Values
(1, 1, 1, 1);
Commit;

SQL*Plus: Release 10.2.0.3.0 - Production on Thu Mar 22 15:20:58 2012

Copyright (c) 1982, 2006, Oracle.  All Rights Reserved.

Connected to:
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - 64bit
Production
With the Partitioning, OLAP, Data Mining and Real Application Testing
options

SQL> select *
     from tmp_number
     where val_n >= to_number(to_char(sysdate, 'YYYYMMDD'));

no rows selected

SQL> select *
     from tmp_number
     where val_n >= to_number(to_char(sysdate, 'YYYYMMDD'))
     or val_n = 0;

no rows selected

SQL> select *
     from tmp_number
     where val_n >= 20120322
     or val_n = 0;

     VAL_N     VAL_N8    VAL_INT      VAL_F
---------- ---------- ---------- ----------
         0          0          0          0

SQL> select *
     from tmp_number
     where val_n >= to_number(to_char(sysdate, 'YYYYMMDD'))
     or val_n <= 0;

     VAL_N     VAL_N8    VAL_INT      VAL_F
---------- ---------- ---------- ----------
         0          0          0          0
        -1         -1         -1         -1
         1          1          1          1

SQL> select *
     from tmp_number
     where val_n >= to_number(to_char(sysdate, 'YYYYMMDD'))
     or to_number(val_n) <= 0;

VAL_N     VAL_N8    VAL_INT      VAL_F
----- ---------- ---------- ----------
    0          0          0          0
   -1         -1         -1         -1

select 
      dump(val_n), dump(val_f), val_n
from tmp_number;

  
VAL_N                     VAL_F                     VAL_N
---------------------------------------------------------
Typ=2 Len=1: 128          Typ=2 Len=1: 128          0      
Typ=2 Len=2: 193,2        Typ=2 Len=2: 193,2        1
Typ=2 Len=3: 62,100,102   Typ=2 Len=3: 62,100,102   -1

Bitte beachten Sie, ICE373 nach Bern fährt heute auf Gleis 6 ein!

Mount Optionen: rw,acl,user_xattr)
$ iostat -d -x 240
...
Device:         rrqm/s   wrqm/s     r/s     w/s   rsec/s   wsec/s avgrq-sz avgqu-sz   await  svctm  %util
sda               0.01     5.28    0.07    2.00     1.33    60.57    29.89     0.01    3.48   0.72   0.15
sdb               2.92   717.37   71.91  293.34  1333.90  8087.83    25.80    65.40  179.04   2.15  78.63

data={journal|ordered|writeback}
         Specifies  the  journalling  mode for file data.  Metadata is always journaled.  To use modes other than ordered on the root filesystem, pass
         the mode to the kernel as boot parameter, e.g.  rootflags=data=journal.

         journal
                All data is committed into the journal prior to being written into the main filesystem.

         ordered
                This is the default mode.  All data is forced directly out to the main file system prior to its metadata being committed to the  jour?
                nal.

         writeback
                Data  ordering  is  not preserved - data may be written into the main filesystem after its metadata has been committed to the journal.
                This is rumoured to be the highest-throughput option.  It guarantees internal filesystem integrity, however it can allow old  data  to
                appear in files after a crash and journal recovery.
commit=nrsec
         Sync all data and metadata every nrsec seconds. The default value is 5 seconds. Zero means default

noatime
        Access timestamps are not updated when a file is read.

• Die Mountoption "noatime" brachte in Spitzenzeiten bei lokalen Mounts maximal 1-2 % Besserung


• Die zusätzliche Mountoption "commit=120" - also das Flushen der Schreib-Buffer alle 120 Sekunden hat in Nebenzeiten fast keine Besserung gebracht, in Hochlastzeiten war der Unterschied bei ca. 25% Utilization


• Die zusätzliche Mountoption "data=writeback" brachte ebenfalls in den Nebenzeit ebenfalls keine Besserung - in Hochlastzeiten war der Unterschied dann nochmals bei 25% Utilization

Natürlich bringen derartige Optimierungen nicht nur Vorteile - die Nachteile (nach einem Linux-Crash würden veraltete Daten im Filesystem auftauchen) waren aber in meiner Situation aber kein Problem...
Nun sieht es selbst in Spitzenzeiten noch ganz annehmbar aus:

$ iostat -d -x 240
Time: 20:56:12
Device:         rrqm/s   wrqm/s     r/s     w/s   rsec/s   wsec/s avgrq-sz avgqu-sz   await  svctm  %util
sda               0.00     5.91    0.02    2.14     0.13    66.83    31.03     0.01    4.95   1.76   0.38
sdb               2.58   351.85   54.17  182.47  1063.50  8283.43    22.59    34.01  143.72   2.29  54.17