Home » RDBMS Server » Server Administration » Memory resizing (Oracle 10g on Solaris 10)
Memory resizing [message #417014] Wed, 05 August 2009 13:36 Go to next message
prvnrk
Messages: 2
Registered: September 2007
Junior Member
Hi,

We are running Oracle 10g on Solaris 10 with 16GB RAM. We got problem with memory and had to reduce to 8GB temporarily. But when we start instance, Server becomes very slow (and freezes at times). I think we need to reduce the memory size in oracle level also (8GB RAM is good enough for our environment) but i dont know how to configure this.

Please advise me how to fix this.


TIA
Prvn
Re: Memory resizing [message #417022 is a reply to message #417014] Wed, 05 August 2009 14:09 Go to previous messageGo to next message
BlackSwan
Messages: 26766
Registered: January 2009
Location: SoCal
Senior Member
>Server becomes very slow (and freezes at times).
What clues exist in alert_SID.log file?

post results from following

SQL> SHOW SGA
Re: Memory resizing [message #417032 is a reply to message #417022] Wed, 05 August 2009 14:40 Go to previous messageGo to next message
prvnrk
Messages: 2
Registered: September 2007
Junior Member
Thanks for reply;

Not much clues in alert_SID.log except for the below:

ORA-19815: WARNING: db_recovery_file_dest_size of 26843545600 bytes is 85.57% us
ed, and has 3874365440 remaining bytes available.


SQL> show SGA

Total System Global Area 4294967296 bytes
Fixed Size 2035592 bytes
Variable Size 2197819512 bytes
Database Buffers 2080374784 bytes
Redo Buffers 14737408 bytes
SQL>

Re: Memory resizing [message #417034 is a reply to message #417014] Wed, 05 August 2009 14:47 Go to previous messageGo to next message
BlackSwan
Messages: 26766
Registered: January 2009
Location: SoCal
Senior Member
>Server becomes very slow (and freezes at times).
How long after starting DB does slowdown occur?

>I think we need to reduce the memory size in oracle level
I disagree. Nominally a 4GB SGA should not be a problem on 8GB RAM

What does "top" show when slowdown occurs?

post appropriate sar data from last slowdown
Re: Memory resizing [message #417532 is a reply to message #417014] Mon, 10 August 2009 01:41 Go to previous message
trantuananh24hg
Messages: 744
Registered: January 2007
Location: Ha Noi, Viet Nam
Senior Member
Dear BlackSwan!

"top" command is not in Solaris.

Use prstat -T (-t, see man prstat).
Use vmstat (second to repeat) (times to repeat), eg: vmstat 2 100, to see how many resource for users, system and idle, how many queues will be executed.. ?

Eg:
$ prstat -T
   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP
 27994 oracle     10G   10G sleep   59    0   0:00:35 9.2% oracle/11
 10910 oracle     10G   10G sleep   59    0   0:19:01 2.6% oracle/11
 11891 oracle     10G   10G sleep   59    0   0:31:38 1.7% oracle/11
 24489 oracle     10G   10G sleep   60    0   0:00:10 0.8% oracle/11
 20849 oracle     10G   10G sleep   50    0   0:04:07 0.8% oracle/11
 23240 oracle     10G   10G sleep   59    0   0:08:13 0.7% oracle/1
  5698 oracle     10G   10G sleep   49    0   7:42:35 0.6% nmccollector/1
  5999 oracle    324M  195M sleep   29   10   2:53:14 0.5% java/35
 25665 oracle     10G   10G sleep   49    0   0:12:38 0.4% oracle/11
    31 oracle     10G   10G sleep   59    0   0:14:30 0.4% oracle/11
 28022 oracle     10G   10G sleep   59    0   0:00:02 0.3% oracle/11
 11543 oracle     10G   10G sleep   59    0   0:10:52 0.3% oracle/11
 19610 oracle     10G   10G sleep   49    0   0:04:42 0.3% oracle/11
 24263 oracle     10G   10G sleep   59    0   0:01:13 0.3% oracle/1
  5396 oracle     10G   10G sleep   59    0   0:18:51 0.2% oracle/12
TASKID    NPROC  SIZE   RSS MEMORY      TIME  CPU PROJECT
  2256      219 2214G 2204G   100%  10:56:09  20% oracle
  1371        3   10G   10G   0.5%   7:42:35 0.6% oracle
  2252        3  366M  221M   0.0%   3:08:01 0.5% oracle
  2520        2 6408K 5032K   0.0%   0:01:23 0.1% oracle
    72       19   80M 8176K   0.0%  11:58:16 0.1% oracle
Total: 302 processes, 1190 lwps, load averages: 0.65, 0.52, 0.45




$ vmstat 2 100
 kthr      memory            page            disk          faults      cpu
 r b w   swap  free  re  mf pi po fr de sr m1 m1 m1 m2   in   sy   cs us sy id
 0 0 40 17995728 2643776 107 526 1217 5 6 0 43 0 0 0 0 1972 4202 1412  9  1 90
 0 0 37 16624032 1720584 4 129 4 8 0  0  0  0  0  0  0 3900 9099 1743  9  3 88
 0 0 37 16623392 1719936 1 545 16 0 0 0  0  0  0  0  0 5358 21906 2524 10 3 87
 0 0 37 16624032 1720568 2 487 8 0 0  0  0  0  0  0  0 5081 3135 2268 10  1 89
 0 0 37 16624288 1720832 0 171 0 0 0  0  0  0  0  0  0 5232 11184 2221 5  2 93
 0 0 37 16625000 1721600 2 43 0 0  0  0  0  0  0  0  0 3917 3226 2031  4  2 94
 0 0 37 16634368 1730168 89 809 32 0 0 0 0  0  0  0  0 2998 4431 1878  8  2 90



kthr->/r: queue for executing
cpu-> /us: user, /sy: sys, /id: idle

At the end, analyze statspack to detect what's problem exactly.

Previous Topic: data dictionary views
Next Topic: About background processes
Goto Forum:
  


Current Time: Fri Dec 06 16:50:59 CST 2024