Oracle Database Discussions

hey guys ....

i get the following errors wher performing an rman db backup (db 10.2.0 , on windows server 2003 enterprise edition):

unable to queue I/O
RMAN-03009: failure of backup command on ORA_DISK_1 channel at 10/30/2008 23:32:24
ORA-19502: write error on file "\\10.10.1.214\BK_F\abs10G\FULL\6AJUFS85_1_1", blockno 6145 (blocksize=512)
ORA-27091: unable to queue I/O
OSD-04008: WriteFile() failure, unable to write to file
O/S-Error: (OS 1450) Insufficient system resources exist to complete the requested service.
ORA-19502: write error on file "\\10.10.1.214\BK_F\abs10G\FULL\6AJUFS85_1_1", blockno 1 (blocksize=512)
ORA-27091: unable to queue I/O

after doing some research , it seems like the problem is caused by the /3gb switch i'm using in the windows boot.ini as follows:

[boot loader]
timeout=30
default=multi(0)disk(0)rdisk(0)partition(1)\WINDOWS
[operating systems]
multi(0)disk(0)rdisk(0)partition(1)\WINDOWS="Windows Server 2003, Enterprise" /fastdetect /3GB /PAE


( metalink NOTE:297498.1 : Resolving Instance Evictions on Windows Platforms was really helpful)

using the performance monitor as described in the note , here's an average of the output on a moderately busy day:
pool paged bytes=131460000
pool nonpaged bytes=42738000
free system page table entries=900

now as indicated in the note , i intend to add the /userva switch too set to an apropriate value , but before doing that on the weekend , does anybody have any recommendations ?
do u think the /pae flag will cause a problem ? & won't the memory reduced from the 3gb & added to the os kernel's PTE's force me to reduce any of the following components?

3GB =
a) AWE_WINDOW_MEMORY (used to address/map the indirect data buffers) +
b) non-buffer_cache part of the SGA (Shared Pool, Large Pool, Java Pool, Streams Pool, Log buffer) +
c) Oracle Process overhead +
d) (Thread stack x number of threads(ie Oracle processes)) +
c) PGA

thank you in advance...

Edited by: user6739264 on Nov 2, 2008 3:40 AM