Help With Affinit Parameter
Hello, IDS 11.10.FC2W1 on Solaris 10, 4 Dual core CPUS 16 GB RAM. Due to some licensing factor. i have been asked to use only 4 core's of the available 8 core's (4 Dual core CPUS). If my understanding is correct i have to bind the CPUVPS to 2 CPUS. correct me if i am wrong here, this can be done with affinity parameter VPCLASS cpu,num=8,max=8,aff=0-7,noage what should the aff= ? be so that only 4 core's will be used by informix the o/p of mpstat shows CPU 1 2 3 4 16 17 18 19 the o/p of prtdiag shows CPU 0, 16 1, 17 2, 18 3, 19 Please let me know what all should be entries in my onconfig file so that only 4 core's (2 CPU) will be used by informix . also what command's can be used to check that informix is using only 4 core's of 8 core's available or 2 CPU's form 4 available CPU's Hope this time i am atleast close enough in terms of posting a correct thread . Thanking you all in advance ! Regards, vikas
VIKAS HIVARKAR wrote: > Hello, > > IDS 11.10.FC2W1 on Solaris 10, > 4 Dual core CPUS > 16 GB RAM. > > Due to some licensing factor. i have been asked to use only 4 core's of the > available 8 core's (4 Dual core CPUS). > > If my understanding is correct i have to bind the CPUVPS to 2 CPUS. > correct me if i am wrong here, this can be done with affinity parameter > VPCLASS cpu,num=8,max=8,aff=0-7,noage > The OS (and therefore IDS) treats each core as a separate CPU (ie CPU 0 has cores treated as CPUs #0 & 16). So, your setting should be: VPCLASS cpu,num=8,max=8,aff=0-4,noage This will start up 8 CPU VPs nailed to the 4 cores starting with core 0, so 0-3. ALternatively you could use: VPCLASS cpu,num=8,max=8,aff16-4,noage To nail the VPS to the upper set of cores, 16-19. Art S. Kagel Oninit > what should the aff= ? be so that only 4 core's will be used by informix > > the o/p of mpstat shows CPU 1 2 3 4 16 17 18 19 > the o/p of prtdiag shows > CPU > 0, 16 > 1, 17 > 2, 18 > 3, 19 > > Please let me know what all should be entries in my onconfig file so that only > 4 core's (2 CPU) will be used by informix . > > also what command's can be used to check that informix is using only 4 core's > of 8 core's available or 2 CPU's form 4 available CPU's > > Hope this time i am atleast close enough in terms of posting a correct thread > .. > > Thanking you all in advance ! > > Regards, > vikas > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > See you at the IIUG Informix 2008 Conference > The Power Conference for Informix Professionals > April 27 - 30, 2008 Marriott Overland Park (Kansas City), Kansas > http://www.iiug.org/conf > Registration Now Open!! > > >
Sorry Mr Kagel again a typo this time i am pasting the o/p of mpstat CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl 0 16 1 88 15 6 414 6 21 35 0 846 1 1 0 98 1 7 1 75 42 34 334 4 28 24 0 224 2 1 0 98 2 6 1 76 2916 72 321 4 20 23 0 198 1 1 0 98 3 15 1 86 301 195 329 4 32 24 0 266 1 1 0 98 16 4 1 99 380 370 118 7 19 35 0 231 1 1 0 98 17 3 1 79 29 19 149 6 22 18 0 218 1 0 0 99 18 3 1 102 17 5 174 7 19 20 0 234 1 1 0 99 19 3 1 91 15 5 237 6 25 20 0 246 1 0 0 99 and not CPU 1 2 3 4 16 17 18 19 In this case VPCLASS cpu,num=8,max=8,aff=0-4,noage Will this remain the same ? or will it be VPCLASS cpu,num=8,max=8,aff=0-3,noage Sorry again !
It is usually recommended to leave the first CPU for the OS functions. In my environment we have a 12 CPU server running two instances and for the first instance the VPCLASS has been set as VPCLASS CPU,num=9,aff=1-17,noage If you see 8 CPUs and you want to use all the 8 CPUs, then your VPCLASS should look like VPCLASS CPU,num=8,aff=0-19,noage My recommendation will be VPCLASS CPU,num=7,aff=1-19,noage You can run a hefty query and check the CPU usage with mpstat utility. Thank you, Kannan Thirugnanam -----Original Message----- From: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] On Behalf Of VIKAS HIVARKAR Sent: Friday, March 28, 2008 12:24 PM To: ids@iiug.org Subject: Re: Help With Affinit Parameter [11737] Sorry Mr Kagel again a typo this time i am pasting the o/p of mpstat CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl 0 16 1 88 15 6 414 6 21 35 0 846 1 1 0 98 1 7 1 75 42 34 334 4 28 24 0 224 2 1 0 98 2 6 1 76 2916 72 321 4 20 23 0 198 1 1 0 98 3 15 1 86 301 195 329 4 32 24 0 266 1 1 0 98 16 4 1 99 380 370 118 7 19 35 0 231 1 1 0 98 17 3 1 79 29 19 149 6 22 18 0 218 1 0 0 99 18 3 1 102 17 5 174 7 19 20 0 234 1 1 0 99 19 3 1 91 15 5 237 6 25 20 0 246 1 0 0 99 and not CPU 1 2 3 4 16 17 18 19 In this case VPCLASS cpu,num=8,max=8,aff=0-4,noage Will this remain the same ? or will it be VPCLASS cpu,num=8,max=8,aff=0-3,noage Sorry again ! ************************************************************************ ******* Forum Note: Use "Reply" to post a response in the discussion forum. See you at the IIUG Informix 2008 Conference The Power Conference for Informix Professionals April 27 - 30, 2008 Marriott Overland Park (Kansas City), Kansas http://www.iiug.org/conf Registration Now Open!! ***** Jackson Hewitt Email Disclaimer ***** The sender believes that this E-mail and any attachments were free of any virus, worm, Trojan horse, and/or malicious code when sent. This message and its attachments could have been infected during transmission. By reading the message and opening any attachments, the recipient accepts full responsibility for taking protective and remedial action about viruses and other defects. The sender's business entity is not liable for any loss or damage arising in any way from this message or its attachments. Privileged/Confidential Information may be contained in this message. If you are not the addressee indicated in this message (or responsible for delivery of the message to such person), you may not copy or deliver this message to anyone. In such case, you should destroy this message and kindly notify the sender by reply email.
VIKAS HIVARKAR wrote: > Sorry Mr Kagel again a typo > > this time i am pasting the o/p of mpstat > CPU minf mjf xcal intr ithr csw icsw migr smtx srw syscl usr sys wt idl > 0 16 1 88 15 6 414 6 21 35 0 846 1 1 0 98 > 1 7 1 75 42 34 334 4 28 24 0 224 2 1 0 98 > 2 6 1 76 2916 72 321 4 20 23 0 198 1 1 0 98 > 3 15 1 86 301 195 329 4 32 24 0 266 1 1 0 98 > 16 4 1 99 380 370 118 7 19 35 0 231 1 1 0 98 > 17 3 1 79 29 19 149 6 22 18 0 218 1 0 0 99 > 18 3 1 102 17 5 174 7 19 20 0 234 1 1 0 99 > 19 3 1 91 15 5 237 6 25 20 0 246 1 0 0 99 > > and not CPU 1 2 3 4 16 17 18 19 > > In this case VPCLASS cpu,num=8,max=8,aff=0-4,noage Will this remain the same ? > or will it be VPCLASS cpu,num=8,max=8,aff=0-3,noage > If you want to pin the VPs to processors 0, 1, 2, & 3 then it would be the second version with 'aff=0-3'. The documentation is not clear on what it would be if you want to pin them to processors 1, 2, 3, & 16 or 16, 17, 18, & 19 as it states: "The value end_range must be larger than start_range, and all values must be less than the total number of available CPUs." But this assumes that all processor cores are sequentially numbered, and as in your case, this is rarely true. I would say the settings for the two extra cases I mentioned would be: VPCLASS cpu,num=8,max=8,aff=1-16 -and- VPCLASS cpu,num=8,max=8,aff=16-19 Also the docs (even the 11.50 Beta docs) are ambiguous as to what the proper separator should be between the start and end range. Also why call each a range when clearly each value is a single CPU number and the pair forms a range? The sample lines all use a dash '-' as we have above, but the syntax line uses a comma ',', so ... I've always used a dash myself and it seems to work. Art S. Kagel Oninit > Sorry again ! > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > See you at the IIUG Informix 2008 Conference > The Power Conference for Informix Professionals > April 27 - 30, 2008 Marriott Overland Park (Kansas City), Kansas > http://www.iiug.org/conf > Registration Now Open!! > > >