HP-UX HB v13.00 Ch-22 - Performance and Tuning

HP-UX Handbook – Rev 13.00 Page 14 (of 19)
Chapter 22 Performance & Tuning
October 29, 2013
# cat /etc/fstab
to determine the file system type associated with the lvol/ mountpoint
# bdf
to see if this volumes file systems are reaching capacity full ( > 85%)
If installed Glance can be used to identify I/O by file system
How to improve disk I/O ?
1. Reduce the volume of data on the disk path to till the run queue is less than 1.0
2. Stripe the data across disks to improve I/O speed and decrease path overload.
3. If you are using Online JFS , run fsadm –e to defragment the extents.
4. Tune the buffer cache / file cache:
Check the cache use with sar – b , ideally the writes to cache (%wcache) should be at
least 75%.
If the %wcache is low , consider the size of the I/O’s and the cache threshold before
tuning the cache size. If these are OK , then the cache size can be reduced .
Example:
05:00:00 bread/s lread/s %rcache bwrit/s lwrit/s %wcache pread/s pwrit/s
Average 0 195 100 131 216 39 0 0
Most data bases buffer internally , for example Oracle uses the SGA buffer pool .
Modern arrays also have onboard cache , these two areas should be taken into
consideration when sizing the dynamic buffer cache/file cache
5. Consider direct mount options, these are available only with the Online JFS product.
The VxFS direct mount options are mincache=direct and convosync=direct, these
should be used together to avoid mixed I/O This can be tested without unmounting
the file system :
To change mount options on the fly:
mount -F vxfs -o remount, mincache=direct, convosync=direct
/dev/vg_name/lvol_name /mount_point
to return to default : mount -F vxfs -o remount /dev/vg_name/lvol_name
/mount_point
6. If you are using raw logical volumes, consider implementing asynchronous I/O.
The difference between the async I/O and synchronous I/O is that async does not
wait for confirmation of the write before moving on to the next task. This does
increase the speed of the disk performance at the expense of robustness.
Synchronous I/O waits for acknowledgement of the write (or fail) before continuing