Re: CFCC Performance
I agree the original question was a bit confusing, but to answer it directly: -> System-managed duplexing is definitely not deprecated, but yes there is an "overhead" to run system-managed duplexing for structures, which is why a new enhancement was recently released on z14 (CFLevel 22 and up) called Asynchronous coupling facility duplexing for lock structures, which for now only DB2 IRLM can exploit when at the right level. This async version of the protocol makes the service time for a CF request to a duplexed DB2 lock structure much closer to that of the same structure running simplex. This makes it much more feasible to run duplexed, especially at distance. -> Db2 is still the only exploiter of and still supports user-managed duplexing as of Db2 V12, so again not deprecated. If it helps, with regards to Coupling Thin Interrupts, I agree the recommendation should be to use them for your shared engine CF configurations. You can find more information to this effect in –> the PR/SM Planning Guide (for EC12 or later). z14 version is PDF SB10-7169-02.pdf available as of October 18 2018 –> the excellent White Paper 102400 – ‘Coupling Thin Interrupts and Coupling Facility Performance in Shared Processor Environments’ by Barbara Weiler at https://www-03.ibm.com/support/techdocs/atsmastr.nsf/WebIndex/WP102400 Hope this helps. Kieron Hinds IBM Z Platform Evaluation Test, IBM Systems From: Jesse 1 Robinson To: IBM-MAIN@LISTSERV.UA.EDU Date: 12/19/2018 11:29 AM Subject: Re: CFCC Performance Sent by:IBM Mainframe Discussion List Thin interrupts were introduced as a performance booster for shared CF engines. However when we recently upgraded both CECs to z14 and z13, we discovered that we can no longer live with 'fat' interrupts. In a new CEC, thin interrupts must be enabled by CF command; there is no profile to carry over the old value to a new CEC. Thin interrupts come into play only when a CF engine is shared among multiple sysplexes, which we have done for years with Development and Sandbox. Without (re)enabling thin interrupts, we found our Sandbox plex to be unusable. By simply turning thin interrupts (back) on, Sandbox performance returned to very respectable. For dedicated engines, thin interrupts are not even an option. The enabling command will be rejected. . . J.O.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 323-715-0595 Mobile 626-543-6132 Office ⇐=== NEW robin...@sce.com -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Vernooij, Kees (ITOP NM) - KLM Sent: Wednesday, December 19, 2018 12:00 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: (External):Re: CFCC Performance So do we, 2 CFs per Sysplex (you "cannot" run a production site with only 1). I am a little confused about your questions: System managed rebuild is standard and always active. This will rebuild structures in another CF in case of problems. It is transparent to the application, be it with some delays in structure availability during the rebuild. Structure duplexing comes in 2 flavors: user-managed and system-managed. User-managed is done by the application, if it supports it, e.g. DB2. If not, you can use system-managed structure duplexing. This comes with some cost of inter-CF communication. We do not use it, because we don't have structures that require 101% availability. Thin interrupts are an enhancement that only provides improvement and exists for several CPU generations, so it should be beyond doubt. Kees. > -Original Message- > From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] > On Behalf Of Mike Schwab > Sent: 18 December, 2018 19:23 > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: CFCC Performance > > We had 4 CFs. Two for the TestPlex and two for the Production Plex. > > On Tue, Dec 18, 2018 at 10:37 AM Allan Staller > wrote: > > > > I am in the process or reconfiguring from a Single CF to multiple > > CFs > (per partition) to eliminate the single point of failure. > > The subject of CFCC Thin Interrupts has been evaluated and will most > likely be implemented. > > > > As part of this process the discussion of DUPLEX CF structure vs. > > use > of SYSTEM MANAGED Rebuild has also come up. > > > > I have heard the use of duplexing has been deprecated in favor of > system managed rebuild for performance reasons. > > CFCC Thin Interrupts might offset some of the (alleged) performance > penalty. > > > > Can anyone point me to any documentation or contacts on the subject? > > Searches if IBM Techdocs, Redbooks, ResourceLink, etc. have produced > very limited information. Most of which seem to be leaning in the > SYSTEM MANAGED R
Re: CFCC Performance
Thin interrupts were introduced as a performance booster for shared CF engines. However when we recently upgraded both CECs to z14 and z13, we discovered that we can no longer live with 'fat' interrupts. In a new CEC, thin interrupts must be enabled by CF command; there is no profile to carry over the old value to a new CEC. Thin interrupts come into play only when a CF engine is shared among multiple sysplexes, which we have done for years with Development and Sandbox. Without (re)enabling thin interrupts, we found our Sandbox plex to be unusable. By simply turning thin interrupts (back) on, Sandbox performance returned to very respectable. For dedicated engines, thin interrupts are not even an option. The enabling command will be rejected. . . J.O.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 323-715-0595 Mobile 626-543-6132 Office ⇐=== NEW robin...@sce.com -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Vernooij, Kees (ITOP NM) - KLM Sent: Wednesday, December 19, 2018 12:00 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: (External):Re: CFCC Performance So do we, 2 CFs per Sysplex (you "cannot" run a production site with only 1). I am a little confused about your questions: System managed rebuild is standard and always active. This will rebuild structures in another CF in case of problems. It is transparent to the application, be it with some delays in structure availability during the rebuild. Structure duplexing comes in 2 flavors: user-managed and system-managed. User-managed is done by the application, if it supports it, e.g. DB2. If not, you can use system-managed structure duplexing. This comes with some cost of inter-CF communication. We do not use it, because we don't have structures that require 101% availability. Thin interrupts are an enhancement that only provides improvement and exists for several CPU generations, so it should be beyond doubt. Kees. > -Original Message- > From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] > On Behalf Of Mike Schwab > Sent: 18 December, 2018 19:23 > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: CFCC Performance > > We had 4 CFs. Two for the TestPlex and two for the Production Plex. > > On Tue, Dec 18, 2018 at 10:37 AM Allan Staller > wrote: > > > > I am in the process or reconfiguring from a Single CF to multiple > > CFs > (per partition) to eliminate the single point of failure. > > The subject of CFCC Thin Interrupts has been evaluated and will most > likely be implemented. > > > > As part of this process the discussion of DUPLEX CF structure vs. > > use > of SYSTEM MANAGED Rebuild has also come up. > > > > I have heard the use of duplexing has been deprecated in favor of > system managed rebuild for performance reasons. > > CFCC Thin Interrupts might offset some of the (alleged) performance > penalty. > > > > Can anyone point me to any documentation or contacts on the subject? > > Searches if IBM Techdocs, Redbooks, ResourceLink, etc. have produced > very limited information. Most of which seem to be leaning in the > SYSTEM MANAGED Rebuild direction. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: CFCC Performance
So do we, 2 CFs per Sysplex (you "cannot" run a production site with only 1). I am a little confused about your questions: System managed rebuild is standard and always active. This will rebuild structures in another CF in case of problems. It is transparent to the application, be it with some delays in structure availability during the rebuild. Structure duplexing comes in 2 flavors: user-managed and system-managed. User-managed is done by the application, if it supports it, e.g. DB2. If not, you can use system-managed structure duplexing. This comes with some cost of inter-CF communication. We do not use it, because we don't have structures that require 101% availability. Thin interrupts are an enhancement that only provides improvement and exists for several CPU generations, so it should be beyond doubt. Kees. > -Original Message- > From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On > Behalf Of Mike Schwab > Sent: 18 December, 2018 19:23 > To: IBM-MAIN@LISTSERV.UA.EDU > Subject: Re: CFCC Performance > > We had 4 CFs. Two for the TestPlex and two for the Production Plex. > > On Tue, Dec 18, 2018 at 10:37 AM Allan Staller > wrote: > > > > I am in the process or reconfiguring from a Single CF to multiple CFs > (per partition) to eliminate the single point of failure. > > The subject of CFCC Thin Interrupts has been evaluated and will most > likely be implemented. > > > > As part of this process the discussion of DUPLEX CF structure vs. use > of SYSTEM MANAGED Rebuild has also come up. > > > > I have heard the use of duplexing has been deprecated in favor of > system managed rebuild for performance reasons. > > CFCC Thin Interrupts might offset some of the (alleged) performance > penalty. > > > > Can anyone point me to any documentation or contacts on the subject? > > Searches if IBM Techdocs, Redbooks, ResourceLink, etc. have produced > very limited information. Most of which seem to be leaning in the SYSTEM > MANAGED Rebuild direction. > > > > Thanks in advance, > > ::DISCLAIMER:: > > -- > > > > > The contents of this e-mail and any attachment(s) are confidential and > intended for the named recipient(s) only. E-mail transmission is not > guaranteed to be secure or error-free as information could be > intercepted, corrupted, lost, destroyed, arrive late or incomplete, or > may contain viruses in transmission. The e mail and its contents (with > or without referred errors) shall therefore not attach any liability on > the originator or HCL or its affiliates. Views or opinions, if any, > presented in this email are solely those of the author and may not > necessarily reflect the views or opinions of HCL or its affiliates. Any > form of reproduction, dissemination, copying, disclosure, modification, > distribution and / or publication of this message without the prior > written consent of authorized representative of HCL is strictly > prohibited. If you have received this email in error please delete it > and notify the sender immediately. Before opening any email and/or > attachments, please check them for viruses and other defects. > > -- > > > > > > > -- > > For IBM-MAIN subscribe / signoff / archive access instructions, > > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN > > > > -- > Mike A Schwab, Springfield IL USA > Where do Forest Rangers go to get away from it all? > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attach
Re: CFCC Performance
I'm somewhat puzzled by this thread. Could mean that I'm missing a basic point. We've been doing parallel sysplex since the mid-90s and therefore gone through a number of advances in technology. So let me toddle a ways down Memory Lane. In the beginning, there was no duplexing. IBM promised it early but ran into technical difficulties and advised customers not to use it. Meanwhile the DB2 folks couldn't/didn't want to wait and so invented their own duplexing, but of course it worked only for DB2. Eventually IBM got 'system managed duplexing' (as opposed to application [DB2] duplexing) working and gave the green light. In order to enable system managed duplexing, you have to format the CFRM couple data set with "ITEM NAME(SMDUPLEX) NUMBER(1)". Once that's done, duplexing will be IN EFFECT for any supported structure whose policy definition allows it. Without duplexing in effect, a backup structure may take over but it may not be entirely in sync with the lost one. We have formatted CFRM couple data set with both SMREBLD and SMDUPLEX. They are not mutually exclusive. The effect is that for all structures allowed by policy, two structures are maintained in sync with each other. If either one fails, the other takes over automatically. On two widely separate occasions we lost an entire CEC containing duplexed ICFs. In both cases, once the CEC was repaired, the sysplex was reIPLed with no data loss. That transparent recovery is worth more than any performance degradation that duplexing might entail. . . J.O.Skip Robinson Southern California Edison Company Electric Dragon Team Paddler SHARE MVS Program Co-Manager 323-715-0595 Mobile 626-543-6132 Office ⇐=== NEW robin...@sce.com -Original Message- From: IBM Mainframe Discussion List [mailto:IBM-MAIN@LISTSERV.UA.EDU] On Behalf Of Allan Staller Sent: Tuesday, December 18, 2018 10:50 AM To: IBM-MAIN@LISTSERV.UA.EDU Subject: (External):Re: CFCC Performance Mike, Are you using Structure Duplexing and/or System managed rebuild? Al -Original Message- From: IBM Mainframe Discussion List On Behalf Of Mike Schwab Sent: Tuesday, December 18, 2018 12:23 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: CFCC Performance We had 4 CFs. Two for the TestPlex and two for the Production Plex. On Tue, Dec 18, 2018 at 10:37 AM Allan Staller wrote: > > I am in the process or reconfiguring from a Single CF to multiple CFs (per > partition) to eliminate the single point of failure. > The subject of CFCC Thin Interrupts has been evaluated and will most likely > be implemented. > > As part of this process the discussion of DUPLEX CF structure vs. use of > SYSTEM MANAGED Rebuild has also come up. > > I have heard the use of duplexing has been deprecated in favor of system > managed rebuild for performance reasons. > CFCC Thin Interrupts might offset some of the (alleged) performance penalty. > > Can anyone point me to any documentation or contacts on the subject? > Searches if IBM Techdocs, Redbooks, ResourceLink, etc. have produced very > limited information. Most of which seem to be leaning in the SYSTEM MANAGED > Rebuild direction. -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
Re: CFCC Performance
Mike, Are you using Structure Duplexing and/or System managed rebuild? Al -Original Message- From: IBM Mainframe Discussion List On Behalf Of Mike Schwab Sent: Tuesday, December 18, 2018 12:23 PM To: IBM-MAIN@LISTSERV.UA.EDU Subject: Re: CFCC Performance We had 4 CFs. Two for the TestPlex and two for the Production Plex. On Tue, Dec 18, 2018 at 10:37 AM Allan Staller wrote: > > I am in the process or reconfiguring from a Single CF to multiple CFs (per > partition) to eliminate the single point of failure. > The subject of CFCC Thin Interrupts has been evaluated and will most likely > be implemented. > > As part of this process the discussion of DUPLEX CF structure vs. use of > SYSTEM MANAGED Rebuild has also come up. > > I have heard the use of duplexing has been deprecated in favor of system > managed rebuild for performance reasons. > CFCC Thin Interrupts might offset some of the (alleged) performance penalty. > > Can anyone point me to any documentation or contacts on the subject? > Searches if IBM Techdocs, Redbooks, ResourceLink, etc. have produced very > limited information. Most of which seem to be leaning in the SYSTEM MANAGED > Rebuild direction. > > Thanks in advance, > ::DISCLAIMER:: > -- > -- > -- > > The contents of this e-mail and any attachment(s) are confidential and > intended for the named recipient(s) only. E-mail transmission is not > guaranteed to be secure or error-free as information could be intercepted, > corrupted, lost, destroyed, arrive late or incomplete, or may contain viruses > in transmission. The e mail and its contents (with or without referred > errors) shall therefore not attach any liability on the originator or HCL or > its affiliates. Views or opinions, if any, presented in this email are solely > those of the author and may not necessarily reflect the views or opinions of > HCL or its affiliates. Any form of reproduction, dissemination, copying, > disclosure, modification, distribution and / or publication of this message > without the prior written consent of authorized representative of HCL is > strictly prohibited. If you have received this email in error please delete > it and notify the sender immediately. Before opening any email and/or > attachments, please check them for viruses and other defects. > -- > -- > -- > > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, send > email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? -- 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: CFCC Performance
We had 4 CFs. Two for the TestPlex and two for the Production Plex. On Tue, Dec 18, 2018 at 10:37 AM Allan Staller wrote: > > I am in the process or reconfiguring from a Single CF to multiple CFs (per > partition) to eliminate the single point of failure. > The subject of CFCC Thin Interrupts has been evaluated and will most likely > be implemented. > > As part of this process the discussion of DUPLEX CF structure vs. use of > SYSTEM MANAGED Rebuild has also come up. > > I have heard the use of duplexing has been deprecated in favor of system > managed rebuild for performance reasons. > CFCC Thin Interrupts might offset some of the (alleged) performance penalty. > > Can anyone point me to any documentation or contacts on the subject? > Searches if IBM Techdocs, Redbooks, ResourceLink, etc. have produced very > limited information. Most of which seem to be leaning in the SYSTEM MANAGED > Rebuild direction. > > Thanks in advance, > ::DISCLAIMER:: > -- > The contents of this e-mail and any attachment(s) are confidential and > intended for the named recipient(s) only. E-mail transmission is not > guaranteed to be secure or error-free as information could be intercepted, > corrupted, lost, destroyed, arrive late or incomplete, or may contain viruses > in transmission. The e mail and its contents (with or without referred > errors) shall therefore not attach any liability on the originator or HCL or > its affiliates. Views or opinions, if any, presented in this email are solely > those of the author and may not necessarily reflect the views or opinions of > HCL or its affiliates. Any form of reproduction, dissemination, copying, > disclosure, modification, distribution and / or publication of this message > without the prior written consent of authorized representative of HCL is > strictly prohibited. If you have received this email in error please delete > it and notify the sender immediately. Before opening any email and/or > attachments, please check them for viruses and other defects. > -- > > -- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN -- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
CFCC Performance
I am in the process or reconfiguring from a Single CF to multiple CFs (per partition) to eliminate the single point of failure. The subject of CFCC Thin Interrupts has been evaluated and will most likely be implemented. As part of this process the discussion of DUPLEX CF structure vs. use of SYSTEM MANAGED Rebuild has also come up. I have heard the use of duplexing has been deprecated in favor of system managed rebuild for performance reasons. CFCC Thin Interrupts might offset some of the (alleged) performance penalty. Can anyone point me to any documentation or contacts on the subject? Searches if IBM Techdocs, Redbooks, ResourceLink, etc. have produced very limited information. Most of which seem to be leaning in the SYSTEM MANAGED Rebuild direction. Thanks in advance, ::DISCLAIMER:: -- The contents of this e-mail and any attachment(s) are confidential and intended for the named recipient(s) only. E-mail transmission is not guaranteed to be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or may contain viruses in transmission. The e mail and its contents (with or without referred errors) shall therefore not attach any liability on the originator or HCL or its affiliates. Views or opinions, if any, presented in this email are solely those of the author and may not necessarily reflect the views or opinions of HCL or its affiliates. Any form of reproduction, dissemination, copying, disclosure, modification, distribution and / or publication of this message without the prior written consent of authorized representative of HCL is strictly prohibited. If you have received this email in error please delete it and notify the sender immediately. Before opening any email and/or attachments, please check them for viruses and other defects. -- -- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
