Re: WLM and DDF enclaves
And possibly checking to see if any zIIP eligible work is actually on the Speciality Processor or can go to a SP. Also sounds like a 4HRA event where I would have a look at makeup during 4HRA to see what Service Classes are contributing, and drill down and tune from there, if necessary.HTHSent from my Verizon, Samsung Galaxy smartphone Original message From: Martin Packer Date: 3/14/24 12:10 (GMT-05:00) To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: WLM and DDF enclaves They should be below Db2 Engine in the hierarchy…IRLM should be in SYSSTC.DBM1, MSTR, DIST should be in Importance 1, CPU Critical.DDF transactions should be Importance 2 on downwards.Right now you have DDF work fighting with Db2 – in whatever shaped boxing ring. 😊Cheers, MartinFrom: IBM Mainframe Discussion List on behalf of Jaime Fernandez <05f045ad6641-dmarc-requ...@listserv.ua.edu>Date: Thursday, 14 March 2024 at 14:26To: IBM-MAIN@LISTSERV.UA.EDU Subject: [EXTERNAL] WLM and DDF enclavesHi fellow listers,Newcomer here, so... I hope my question makes sense.We have some amount of business critical distributed online transactions, running on enclaves into a Db2 zOS subsystem. Our current WLM configuration assigns them a specific service class with importance 1 and a velocity goal.We have been observing quite poor performance index for this service class (and awful response times for those transactions) whenever the lpar shows high cpu usage (on or around 100%) and specially when capping is on. Performance index stays between 1 and 2 most of the time, with peaks over 40.From what I have read so far, I understand velocity goals can be tricky for short-term processes, and that gets worse with fewer logical processors. Which happens to be our case, last year we brought in a z16 with less capacity and processors than the previous machine.So, in your experience, could be a good idea switching to a response time goal?Any insights would be most appreciated.Jaime--For IBM-MAIN subscribe / signoff / archive access instructions,send email to lists...@listserv.ua.edu with the message: INFO IBM-MAINUnless otherwise stated above:IBM United Kingdom LimitedRegistered in England and Wales with number 741598Registered office: PO Box 41, North Harbour, Portsmouth, Hants. PO6 3AU--For IBM-MAIN subscribe / signoff / archive access instructions,send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: WLM and DDF enclaves
They should be below Db2 Engine in the hierarchy… IRLM should be in SYSSTC. DBM1, MSTR, DIST should be in Importance 1, CPU Critical. DDF transactions should be Importance 2 on downwards. Right now you have DDF work fighting with Db2 – in whatever shaped boxing ring. 😊 Cheers, Martin From: IBM Mainframe Discussion List on behalf of Jaime Fernandez <05f045ad6641-dmarc-requ...@listserv.ua.edu> Date: Thursday, 14 March 2024 at 14:26 To: IBM-MAIN@LISTSERV.UA.EDU Subject: [EXTERNAL] WLM and DDF enclaves Hi fellow listers, Newcomer here, so... I hope my question makes sense. We have some amount of business critical distributed online transactions, running on enclaves into a Db2 zOS subsystem. Our current WLM configuration assigns them a specific service class with importance 1 and a velocity goal. We have been observing quite poor performance index for this service class (and awful response times for those transactions) whenever the lpar shows high cpu usage (on or around 100%) and specially when capping is on. Performance index stays between 1 and 2 most of the time, with peaks over 40. From what I have read so far, I understand velocity goals can be tricky for short-term processes, and that gets worse with fewer logical processors. Which happens to be our case, last year we brought in a z16 with less capacity and processors than the previous machine. So, in your experience, could be a good idea switching to a response time goal? Any insights would be most appreciated. Jaime -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN Unless otherwise stated above: IBM United Kingdom Limited Registered in England and Wales with number 741598 Registered office: PO Box 41, North Harbour, Portsmouth, Hants. PO6 3AU -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: WLM and DDF enclaves
You might try to get WLM/RMF reports on response times to get a base line Colin On Thu, 14 Mar 2024 at 14:26, Jaime Fernandez < 05f045ad6641-dmarc-requ...@listserv.ua.edu> wrote: > Hi fellow listers, > > Newcomer here, so... I hope my question makes sense. > > We have some amount of business critical distributed online transactions, > running on enclaves into a Db2 zOS subsystem. Our current WLM configuration > assigns them a specific service class with importance 1 and a velocity > goal. > > We have been observing quite poor performance index for this service class > (and awful response times for those transactions) whenever the lpar shows > high cpu usage (on or around 100%) and specially when capping is on. > Performance index stays between 1 and 2 most of the time, with peaks over > 40. > > From what I have read so far, I understand velocity goals can be tricky > for short-term processes, and that gets worse with fewer logical > processors. Which happens to be our case, last year we brought in a z16 > with less capacity and processors than the previous machine. > > So, in your experience, could be a good idea switching to a response time > goal? > > Any insights would be most appreciated. > > Jaime > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
WLM and DDF enclaves
Hi fellow listers, Newcomer here, so... I hope my question makes sense. We have some amount of business critical distributed online transactions, running on enclaves into a Db2 zOS subsystem. Our current WLM configuration assigns them a specific service class with importance 1 and a velocity goal. We have been observing quite poor performance index for this service class (and awful response times for those transactions) whenever the lpar shows high cpu usage (on or around 100%) and specially when capping is on. Performance index stays between 1 and 2 most of the time, with peaks over 40. From what I have read so far, I understand velocity goals can be tricky for short-term processes, and that gets worse with fewer logical processors. Which happens to be our case, last year we brought in a z16 with less capacity and processors than the previous machine. So, in your experience, could be a good idea switching to a response time goal? Any insights would be most appreciated. Jaime -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
