- Community Home
- >
- Servers and Operating Systems
- >
- Operating Systems
- >
- Operating System - HP-UX
- >
- Memory question
Categories
Company
Local Language
Forums
Discussions
Forums
- Data Protection and Retention
- Entry Storage Systems
- Legacy
- Midrange and Enterprise Storage
- Storage Networking
- HPE Nimble Storage
Discussions
Forums
Discussions
Discussions
Discussions
Forums
Discussions
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
- BladeSystem Infrastructure and Application Solutions
- Appliance Servers
- Alpha Servers
- BackOffice Products
- Internet Products
- HPE 9000 and HPE e3000 Servers
- Networking
- Netservers
- Secure OS Software for Linux
- Server Management (Insight Manager 7)
- Windows Server 2003
- Operating System - Tru64 Unix
- ProLiant Deployment and Provisioning
- Linux-Based Community / Regional
- Microsoft System Center Integration
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Community
Resources
Forums
Blogs
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-18-2002 06:58 AM
03-18-2002 06:58 AM
I've an old HP 9000/t600 server (running HP-UX 11.00) with about 1,3 Gb of Physical memory.
Output of swapinfo is:
Kb Kb Kb PCT START/ Kb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 2097152 342440 1754712 16% 0 - 1 /dev/vg00/lvol2
dev 294912 294848 64 100% 0 - 0 /dev/vg00/lvol10
reserve - 1724160 -1724160
memory 935036 463400 471636 50%
total 3327100 2824848 502252 85% - 0 -
Output of ipcs -ma is:
Shared Memory:
m 0 0x4118061a --rw-rw-rw- root root root root 0 348 758 758 9:44:52 9:44:52 9:44:46
m 1 0x4e0c0002 --rw-rw-rw- root root root root 1 31040 758 758 9:44:49 9:44:52 9:44:46
m 2 0x411c6371 --rw-rw-rw- root root root root 1 8192 758 770 9:44:49 9:44:46 9:44:46
m 96003 0x29c905a4 --rw-r----- oracle dba oracle dba 11 697331712 2590 15758 11:47:26 12:24:56 9:53:34
m 404 0x3f2db010 --rw-r----- oracle dba oracle dba 6 14794752 28391 6851 18:57:59 0:19:44 12:05:50
m 5 0xe8199378 --rw-r----- oracle dba oracle dba 15 259022848 2777 1449 15:33:48 15:35:54 9:55:25
m 3206 0xcf438aa0 --rw-r----- oracle dba oracle dba 6 107790336 26533 19895 14:01:25 14:25:02 10:30:29
m 208 0x88547a88 --rw-r----- oracle dba oracle dba 6 29253632 21923 17214 10:29:51 10:31:54 10:23:57
(... a lot of SGA)
Output of top (Memory line) is:
Memory: 64744K (39904K) real, 1138848K (1098144K) virtual, 92392K free
When "free" is about 4 Mb (same as the setting of our minfree), we got the paging activity and then "free" goes to 32 Mb (same as the setting of our lotsfree). When this happens, "real" is never very high (about 70- 80 Mb). Our dcb_max_pct is 50.
Is it normal that "free" is so small (It is caused by the Oracle's SGA?) ?
Thanks in advance for your help.
Francesca.
Solved! Go to Solution.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-18-2002 07:06 AM
03-18-2002 07:06 AM
Re: Memory question
Sandip
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-18-2002 07:13 AM
03-18-2002 07:13 AM
SolutionI see a couple of things that are bad. It appears that you have about 1.1GB of shared memory allocated in a machine with only 1.3GB of memory. You need to reduce the size of your SGA's so that your machine is not paging out. The performance hit you take from page-outs far exceeds the benefits of SGA buffer cache hits. I also note that you have pseudo-swap enabled (swapmem_on=1). There is no reason to run pseudo-swap when you havbe more swap than memory. You should cetainly reduce the dbc_max_pct to no more than about 15% (10% is probably better) or use a non-zero bufpages value to set a static buffer cache of about 200MB (bufpages=51200) as a reasonable starting point. In any event, you need to reduce your SGA's.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-18-2002 07:46 AM
03-18-2002 07:46 AM
Re: Memory question
I agree with Clay for the size of the actuel SGA 1.1 GB for a total of memory of 1.3 GB IS really too much and will prevent all other application to be able to run correctly.
. You must reduce the amount of your SGA to a suitable value that match with the other running applications on your system.
You may also increase your swap space to 3 to 4 times the amount of memory: that's makes from 3.9 to 4.2 GB to have good performance.
I would leave the swap_on kernel parameter to value "1" to speed up processes if it's necessary ( your process will use a specific amount of memory space like a swap area ).
Good luck
Magdi
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-18-2002 10:29 AM
03-18-2002 10:29 AM
Re: Memory question
Bill Hassell, sysadmin
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-18-2002 10:17 PM
03-18-2002 10:17 PM
Re: Memory question
According to the SWAP issue :
In the book " Essential System Administration" by Aeleen Frisch - O'Reilly second edition - page 315 under " How much paging space " it's written :
"For real production environments running programs with very large memory requirements, you'll see better performancee with three or four times the amount of physical memory"
----
Swap is used to page-out process whenever they are asking for I/Os, but if the OS didn't find enough space to page-out these process such error message will appear.
The output of the command "swapinfo -tam" line "reserve" gives us the last tentative of allocation space in swap.
Magdi