About RA_PAGES

Hi, I've a DSS style Server, I have this parameters in my ONCONFIG file: # Read Ahead Variables RA_PAGES 1000 # Number of pages to attempt to readahead RA_THRESHOLD 500 # Number of pages left before next group My AIO's working to much. Output of "onstat -g iov" : AIO I/O vps: class/vp s io/s totalops dskread dskwrite dskcopy wakeups io/wup errors msc 0 i 0.0 936 0 0 0 843 1.1 0 aio 0 s 9.4 743188 682740 60408 0 260139 2.9 0 aio 1 s 8.8 694198 642324 51860 0 214580 3.2 0 aio 2 s 8.4 669338 620532 48805 0 193024 3.5 0 aio 3 s 8.3 655764 612561 43202 0 181597 3.6 0 aio 4 s 8.1 643922 603802 40119 0 171931 3.7 0 aio 5 s 8.0 633544 598318 35225 0 163036 3.9 0 aio 6 s 7.8 621254 588774 32479 0 152565 4.1 0 aio 7 s 7.7 609731 580301 29429 0 142759 4.3 0 aio 8 s 7.5 597762 571123 26638 0 133127 4.5 0 aio 9 s 7.4 584992 560647 24344 0 123137 4.8 0 aio 10 s 7.2 571072 549016 22054 0 113977 5.0 0 aio 11 s 7.1 559445 539354 20089 0 105207 5.3 0 aio 12 s 6.9 548570 529858 18711 0 96963 5.7 0 aio 13 s 6.8 538166 520690 17475 0 89020 6.0 0 aio 14 s 6.7 530005 513450 16555 0 83058 6.4 0 aio 15 i 6.6 526492 510207 16285 0 78123 6.7 0 pio 0 i 0.0 2892 0 2892 0 2893 1.0 0 lio 0 i 0.1 10362 0 10362 0 10363 1.0 0 My system is: IDS 7.31.UC2 on HPUX-11.00. (HP-9000/L1000 One CPU 440Mhz) Load average of system when the motor is running is 16 (note: aio's = 16) Well, after of this situation, I've decided to change number of aio's. onmode -p +20 aio Well, After some time, I've observed: The value io/s down to 2 to 4. and the load of system grown to 36. (note aio's = 36) Well. What of the two situations is better ?. 1.- The overhead in the system with a lot of aio's running. or 2.- The overhead in the Server with few aios.

How did you arrive at your settings for RA_PAGES and RA_THRESHOLD??? Thanks in advance, Tim Informatica wrote: > > Hi, > > I've a DSS style Server, I have this parameters in my ONCONFIG file: > > # Read Ahead Variables > > RA_PAGES 1000 # Number of pages to attempt to read> ahead > RA_THRESHOLD 500 # Number of pages left before next group> -- . .- .-- .--- .---- Tim Schaefer .----- tschaefe@bellsouth.net .---- http://www.inxutil.com .--- .-- .- .

Informatica wrote: > > Hi, > > I've a DSS style Server, I have this parameters in my ONCONFIG file: > > # Read Ahead Variables > > RA_PAGES 1000 # Number of pages to attempt to read> ahead > RA_THRESHOLD 500 # Number of pages left before next group> > My AIO's working to much. Output of "onstat -g iov" : > > AIO I/O vps: > class/vp s io/s totalops dskread dskwrite dskcopy wakeups io/wup > errors > msc 0 i 0.0 936 0 0 0 843 > 1.1 0 > aio 0 s 9.4 743188 682740 60408 0 260139 > 2.9 0 > aio 1 s 8.8 694198 642324 51860 0 214580 > 3.2 0 > aio 2 s 8.4 669338 620532 48805 0 193024 > 3.5 0 > aio 3 s 8.3 655764 612561 43202 0 181597 > 3.6 0 > aio 4 s 8.1 643922 603802 40119 0 171931 > 3.7 0 > aio 5 s 8.0 633544 598318 35225 0 163036 > 3.9 0 > aio 6 s 7.8 621254 588774 32479 0 152565 > 4.1 0 > aio 7 s 7.7 609731 580301 29429 0 142759 > 4.3 0 > aio 8 s 7.5 597762 571123 26638 0 133127 > 4.5 0 > aio 9 s 7.4 584992 560647 24344 0 123137 > 4.8 0 > aio 10 s 7.2 571072 549016 22054 0 113977 > 5.0 0 > aio 11 s 7.1 559445 539354 20089 0 105207 > 5.3 0 > aio 12 s 6.9 548570 529858 18711 0 96963 > 5.7 0 > aio 13 s 6.8 538166 520690 17475 0 89020 > 6.0 0 > aio 14 s 6.7 530005 513450 16555 0 83058 > 6.4 0 > aio 15 i 6.6 526492 510207 16285 0 78123 > 6.7 0 > pio 0 i 0.0 2892 0 2892 0 2893 > 1.0 0 > lio 0 i 0.1 10362 0 10362 0 10363 > 1.0 0 > > My system is: > > IDS 7.31.UC2 on HPUX-11.00. (HP-9000/L1000 One CPU 440Mhz) > > Load average of system when the motor is running is 16 (note: aio's = > 16) > > Well, after of this situation, I've decided to change number of aio's. > > onmode -p +20 aio> > Well, After some time, I've observed: > > The value io/s down to 2 to 4. and the load of system grown to 36. (note > aio's = 36) > > Well. What of the two situations is better ?. > > 1.- The overhead in the system with a lot of aio's running. or > 2.- The overhead in the Server with few aios. Or the bigger question (to me): How did you arrive at those values for your RA variables?? -- John Carlson Informix DBA WHSmith USA #include std_disclaimer.h /* These are my opinions, not my company's opinion */

Informatica wrote: > > Hi, > > I've a DSS style Server, I have this parameters in my ONCONFIG file: > > # Read Ahead Variables > > RA_PAGES 1000 # Number of pages to attempt to read> ahead > RA_THRESHOLD 500 # Number of pages left before next group> [cutting] These seem a little large, how were they derived, dropping back to 32/16 (ish) sound help a little -- Paul Watson # WF Software # If it was easy Tel ++44 1436 674729 # Everybody could do it Fax ++44 1436 678693 # www.wfsoftware.com/informix #

> > RA_PAGES 1000 # Number of pages to attempt to read> > ahead > > RA_THRESHOLD 500 # Number of pages left before nextgroup > > Or the bigger question (to me): > How did you arrive at those values for your RA variables?? > -- > John Carlson > Informix DBA > WHSmith USA > Definitely. In the Performance Tuning guide, they state quite clearly that the RA_PAGES shouldn't be set higher than 32, for this very reason. Try RA_PAGES at 32 and RA_THRESHOLD at 16 or 24 so so. What's happening is each of the aoi threads is being forced to read 1000 * PAGE_SIZE, or probably either 2 or 4 MB into memory for every 500 read. I bet your ra pages used is in the gutter. -- # unrm / ksh: unrm: not found # man cpio Sent via Deja.com http://www.deja.com/ Before you buy.

mars1972@my-deja.com wrote: > > > RA_PAGES 1000 # Number of pages to attempt to read> > > ahead > > > RA_THRESHOLD 500 # Number of pages left before next> group > > > > Or the bigger question (to me): > > How did you arrive at those values for your RA variables?? > > -- > > John Carlson > > Informix DBA > > WHSmith USA > > > > Definitely. In the Performance Tuning guide, they state quite clearly > that the RA_PAGES shouldn't be set higher than 32, for this very reason. > Try RA_PAGES at 32 and RA_THRESHOLD at 16 or 24 so so. What's > happening is each of the aoi threads is being forced to read 1000 * > PAGE_SIZE, or probably either 2 or 4 MB into memory for every 500 read. > I bet your ra pages used is in the gutter. > --- Sorry, I haven't the documentation 'Performance Tuning Guide', I've only the 'Performance Guide', where I've used the follow formula, in page 3-60: RA_PAGES = (BUFFERS * bp_fact) / ( 2 * large_queries) + 2 RA_THRESHOLD = (BUFFERS * bp_fact) / ( 2 * large_queries) - 2 bp_fact = Portion of data buffers to use large scans that require read-ahead i.e.: 75 percent off buffers. large_queries = number of concurrent queries that require read-ahead that you intend to support. This information is OLD or is more new than you have ?. If my information is not correct, then I will use RA_PAGES = 32 and RA_THRESHOLD = 24. --- > > Sent via Deja.com http://www.deja.com/ > Before you buy.

The following link will take you to the web page at Informix where you can DOWNLOAD any documentation available there without charge: http://www.informix.com/answers "Dpto. Informatica" <navidul@retemail.es> wrote in message news:3882FE89.7A6C68AA@retemail.es... > mars1972@my-deja.com wrote: > > > > > RA_PAGES 1000 # Number of pages to attempt to read> > > > ahead > > > > RA_THRESHOLD 500 # Number of pages left before next> > group > > > > > > Or the bigger question (to me): > > > How did you arrive at those values for your RA variables?? > > > -- > > > John Carlson > > > Informix DBA > > > WHSmith USA > > > > > > > Definitely. In the Performance Tuning guide, they state quite clearly > > that the RA_PAGES shouldn't be set higher than 32, for this very reason. > > Try RA_PAGES at 32 and RA_THRESHOLD at 16 or 24 so so. What's > > happening is each of the aoi threads is being forced to read 1000 * > > PAGE_SIZE, or probably either 2 or 4 MB into memory for every 500 read. > > I bet your ra pages used is in the gutter. > > > > --- > > Sorry, I haven't the documentation 'Performance Tuning Guide', I've > only > the 'Performance Guide', where I've used the follow formula, in page 3-60: > > RA_PAGES = (BUFFERS * bp_fact) / ( 2 * large_queries) + 2 > RA_THRESHOLD = (BUFFERS * bp_fact) / ( 2 * large_queries) - 2 > > bp_fact = Portion of data buffers to use large scans that require > read-ahead > i.e.: 75 percent off buffers. > > large_queries = number of concurrent queries that require read-ahead that > you intend to support. > > This information is OLD or is more new than you have ?. > > If my information is not correct, then I will use RA_PAGES = 32 and > RA_THRESHOLD = 24.> > --- > > > > > Sent via Deja.com http://www.deja.com/ > > Before you buy. >