- Community Home
- >
- Servers and Operating Systems
- >
- Operating Systems
- >
- Operating System - HP-UX
- >
- Re: Slow IO Performance for Oracle on 11.31 versus...
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
Discussions
Discussions
Forums
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
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
тАО11-07-2010 03:55 PM
тАО11-07-2010 03:55 PM
Slow IO Performance for Oracle on 11.31 versus 11.23
I made a comparison on Oracle query response time on Oracle 10g database on HP-UX 11.31 and 11.23 (production). The database on 11.31 in a production copy.
The response time of query on 11.31 is 3 times slower on 11.31. I ran Oracle AWR reports on both servers, confirmed that they both using the same Oracle execution plan and the same number of the Oracle block read requests. The only different between them is the average IO wait on 11.31 is higher.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО11-07-2010 07:39 PM
тАО11-07-2010 07:39 PM
Re: Slow IO Performance for Oracle on 11.31 versus 11.23
Shibin
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО11-07-2010 08:04 PM
тАО11-07-2010 08:04 PM
Re: Slow IO Performance for Oracle on 11.31 versus 11.23
Looks like you did your homework on this problem.
When you mention average IO wait, is that as per Oracle, or as per the OS?
Too few folks realize that an accute CPU shortage will cause Oracle to report slow IO. This is because for Oracle to 'see' the IO complete, the initiating process needs to be scheduled again, so there is some CPU and possibly a lot of scheduling time included in the IO time as reported by Oracle. Admittedly, typically an application would need to be in the high cpu load ranges for this to play a role.
What does oracle report as the main component for the transaction (wait_time)?
Is the cpu time for top queries similar (and significant?)
Are the servers and storage sufficiently comparable?
Hope this help some,
Hein van den Heuvel.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО11-07-2010 11:15 PM
тАО11-07-2010 11:15 PM
Re: Slow IO Performance for Oracle on 11.31 versus 11.23
ASM? Files? Logical vols - How many? Physical disks or RAID'd LUNs on external storage, how many and of what type etc etc...
Specificaly what sort of storage array, and how were you managing MPIO on the 11.23 system?
HTH
Duncan
I am an HPE Employee
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО11-08-2010 06:08 AM
тАО11-08-2010 06:08 AM
Re: Slow IO Performance for Oracle on 11.31 versus 11.23
While I do not like 'shooting from the hip', in this case I would first check the SCSI queue depth. The default is just 8 which is often not nearly enough for a lun with many physical disks behind it. Perhaps it was adjusted on the 11.23 system and not on the 11.31 ?
google for queue_depth for details.
It may need to be in the 42.
Or 50 or 60 or 100. The science is too hard. If you have an important, repeateable query then try a few settings (8, 16, 32, 64, 128) and look for diminishing returns. Tweak test values for binary search if it matters a lot.
fwiw,
Hein
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО11-08-2010 01:42 PM
тАО11-08-2010 01:42 PM