See metalink article 200266.1 Increasing Usable Address Space for Oracle
on 32-bit Linux
-----Original Message-----
From: Lucas Brasilino [mailto:lucas.brasilino@(protected)]
Sent: Monday, October 23, 2006 11:19 AM
To: suse-oracle@(protected)
Subject: [suse-oracle] Oracle 9i can't use more than 1.8Gb SGA in Suse
Enterprise Server 9
Hi folks:
I'm trying to make Oracle 9i an SGA with more than 1.8Gb, but I was not
successful, even googling around and trying a lot of configurations both
in Oracle and Linux Kernel, which I'll describe (some that I remember).
Our system:
SUSE LINUX Enterprise Server 9
# uname -r
2.6.5-7.244-bigsmp
Oracle 9i:
Oracle9i Enterprise Edition Release 9.2.0.8.0 - Production With the
Partitioning, OLAP and Oracle Data Mining options JServer Release
9.2.0.8.0 - Production
Two CPU: Intel(R) Xeon(TM) CPU 3.60GHz (32bits) SMP Kernel
# free -m
total used free shared buffers
cached
Mem: 4048 3905 143 0 53
611
-/+ buffers/cache: 3240 807
Swap: 6149 0 6149
# mount
<snippet>
tmpfs on /dev/shm type tmpfs (rw,nr_blocks=8g) </snippet>
# cat /proc/meminfo (after trying to setting sysctl vm.nr_hugepages to
2048):
MemTotal: 4145552 kB
MemFree: 146548 kB
Buffers: 54464 kB
Cached: 626072 kB
SwapCached: 0 kB
Active: 232044 kB
Inactive: 519596 kB
HighTotal: 3276592 kB
HighFree: 5648 kB
LowTotal: 868960 kB
LowFree: 140900 kB
SwapTotal: 6297440 kB
SwapFree: 6296960 kB
Dirty: 0 kB
Writeback: 0 kB
Mapped: 137964 kB
Slab: 29520 kB
Committed_AS: 968488 kB
PageTables: 3080 kB
VmallocTotal: 112632 kB
VmallocUsed: 41360 kB
VmallocChunk: 70744 kB
HugePages_Total: 1492
HugePages_Free: 899
Hugepagesize: 2048 kB
# sysctl vm
<snippet>
vm.disable_cap_mlock = 1
vm.nr_hugepages = 1492
</snippet>
# sysctl kernel
<snippet>
kernel.sem = 250 32000 100 128
kernel.msgmnb = 65536
kernel.msgmni = 2878
kernel.msgmax = 8192
kernel.shmmni = 100
kernel.shmall = 2097152
kernel.shmmax = 2147483648 (half of the physical RAM) </snippet>
# cat ini<our instance>.ora
db_files = 200
open_cursors = 300
max_enabled_roles = 100
db_file_multiblock_read_count = 125
shared_pool_size = 700000000
db_cache_size = 1024000000 #
java_pool_size = 20000000
#use_indirect_data_buffers=true
hash_join_enabled=true
large_pool_size = 20000000
parallel_automatic_tuning = TRUE
parallel_max_servers = 5
pga_aggregate_target = 25165824
db_block_size = 8192
Well, I think I haven't forgotten any important information.
I have tried to:
* mount tmpfs with 'nr_blocks=8g' mouting option
* set "disable_vm_cap_mlock" to 1
* set DISABLE_HUGETLBFS=1 environment variable
* 'oracle' process is setuid
* use_indirect_data_buffers=true in our init<instance>.ora
* my last try was to set vm.nr_hugepages to 2048 since my HugePagesize
is 2048Kb.
Those configurations above was made with vm.nr_hugepages=0.
So, any feelings where I'm messing things up ?
thanks a lot in advance.
Lucas Brasilino
--
To unsubscribe, email: suse-oracle-unsubscribe@(protected)
commands, email: suse-oracle-help@(protected)
http://www.suse.com/oracle/ before posting
--
To unsubscribe, email: suse-oracle-unsubscribe@(protected)
For additional commands, email: suse-oracle-help@(protected)
Please see http://www.suse.com/oracle/ before posting
