- Community Home
- >
- Servers and Operating Systems
- >
- Operating Systems
- >
- Operating System - HP-UX
- >
- memory issue
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
08-10-2001 02:29 AM
08-10-2001 02:29 AM
memory issue
I've recently take over administration of a L-class 2000 server running Apache and Oracle apps server 11i (it's also been installed as 64bit), I suspect that the configuration is not being optimised and the performance maybe poor.
I've read several threads regarding swap and pseudo_swap, I would appreciated if anyone could advise me on weather I should enable pseudo swap and what I should set my dbc_max_pct kernel parameter too.
The current setting of the system is,
2GB Physical memory
3GB Disk swap
dbc_max_pct= 40
Pseudo swap is on
I ran glance and got the following reads.
Memory Utilisation = 100%
Swap Utilisation = 53%
Pseudo_swap = 1.5GB of which 866mb is utilised.
Thanks
Michael
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-10-2001 02:42 AM
08-10-2001 02:42 AM
Re: memory issue
as far as I know this situation indicates that the system is swapping and you need to install additional amount of phisical memory. You can perform swapinfo and sar with -w option to watch swapping
later,
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-10-2001 02:59 AM
08-10-2001 02:59 AM
Re: memory issue
UNIX95= ps -aef -o "pid,user,comm,sz" | sort +3.0 -n
List process in order of memory comsum.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-10-2001 04:53 AM
08-10-2001 04:53 AM
Re: memory issue
My first suggestion would be
dbc_max_pct is high, as you have 2Gb of memory as value of 200-300B for buffer cache would be ok i guess.
so 10-15% would be fine I guess.
..BPK...
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-10-2001 05:26 AM
08-10-2001 05:26 AM
Re: memory issue
You can estimate total resident memory size (RSS) on the system by
$ ps -elf |awk '{print $10}'
it gives RSS for all processes in 4MB pages.
Rest of the memory are used by pseudo swap/buffer cache. Dynamic buffer cache has lower priority than processes - it will release memory if a process requests it.
You may reduce dbc_max_pct from 40 to 10 or 15.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-11-2001 02:58 AM
08-11-2001 02:58 AM
Re: memory issue
As the memory is big enough..U can obviously
enable psudeo_swap..
U can try
# kmtune -s swapmem_on = 1
and then rebuild the kernel..
U can try setting the
dbc_min_pct = 5 and dbc_mac_pct_25.
and then increase the dbc_max_pct according
to the performance..
Also make sure
nbuf and bufpages are set to 0 so that
Dynamic Buffer will be enabled..
Hope this helps
Sundar
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-12-2001 08:17 AM
08-12-2001 08:17 AM
Re: memory issue
In your case, I think you will actually see better performance if you disable dynamic buffer cache and set bufpages to about 80000
(~320 MB) and that is very, very generous.
If you measure, you find in the vast majority of cases that buffer cache hit rates and performance increase very,very slowly past this value. I suspect that you could drop it even more (~200 MB) and still get very good performance. If your are using raw/io or vxfs mount option convosync=direct,mincache=direct which bypass the unix buffers, you can drop even lower. Typically, it's better to use that memory in the SGA so that reducing buffer cache lets up pump up the size of the buffers within the SGA.
My 2 cents, Clay