LongEx Mainframe Quarterly - November 2022
OK, let's get this out of the way. I love RMF Monitor III. Really do. It's just brilliant. It can do amazing things. And best of all, most of my clients have it, and let me use it. You could spend a lot of time looking at all the Share presentations and other documentation on RMF and what it can do. But let me tell you the five features I use most. CPC CapacityI often logon to a new site, and need to find out information about the z/OS systems fast. RMF Monitor III regularly comes to the rescue. Let's take the CPC Capacity report: RMF V2R4 CPC Capacity Line 1 of 18 Command ===> Scroll ===> CSR Samples: 300 System: MVS1 Date: 08/27/22 Time: 00.34.00 Range: 300 Sec Partition: MVS1 3906 Model 605 Boost: N CPC Capacity: 591 Weight % of Max: **** 4h Avg: 352 Image Capacity: 591 WLM Capping %: 0.0 4h Max: 506 Group: N/A MT Mode IIP: N/A Prod % IIP: N/A AbsMSUCap: N Limit: N/A Partition --- MSU --- Cap Proc Logical Util % - Physical Util % - Def Act Def Num Effect Total LPAR Effect Total *CP 17.0 0.8 67.3 68.1 MVS1 0 316 N N N 5.0 53.1 53.4 0.3 53.1 53.4 MVS2 0 12 N N N 5.0 2.1 2.1 0.0 2.1 2.1 MVS3 0 64 N N N 5.0 10.7 10.8 0.0 10.7 10.8 MVS4 0 9 N Y N 2.0 3.6 3.6 0.0 1.4 1.4 PHYSICAL 0.4 0.4 *ICF 3.0 0.0 100 100 CF01 N N N 2.0 100 100 0.0 66.7 66.7 CH02 N N N 1.0 100 100 0.0 33.3 33.3 PHYSICAL 0.0 0.0 *IIP 7.0 0.3 12.6 13.0 MVS1 N N N 2.0 13.2 13.4 0.1 8.8 8.9 MVS2 N N N 2.0 2.2 2.2 0.0 1.5 1.5 MVS3 N N N 2.0 3.1 3.1 0.0 2.1 2.1 MVS4 N N N 1.0 0.7 0.7 0.0 0.2 0.2 PHYSICAL 0.2 0.2 It looks good, doesn't it? I find out so much information about the processor from this one screen. Let's list some of things I find useful:
2. System InformationThe System Information panel gives some of the same information: RMF V2R4 System Information Line 1 of 61 Command ===> Scroll ===> CSR Samples: 300 System: MVS1 Date: 08/27/22 Time: 00.34.00 Range: 300 Sec Partition: MVS1 3906 Model 605 Appl%: 43 Policy: POLICYA CPs Online: 5.0 Avg CPU Util%: 53 EAppl%: 47 Date: 04/26/19 AAPs Online: - Avg MVS Util%: 64 Appl% AAP: - Time: 09.42.12 IIPs Online: 2.0 Appl% IIP: 12 Group T WFL --Users-- RESP TRANS -AVG USG- -Average Number Delayed For - % TOT ACT Time /SEC PROC DEV PROC DEV STOR SUBS OPER ENQ *SYSTEM 71 661 13 0.90 3.5 7.6 2.3 1.3 0.0 0.9 0.0 0.1 *TSO 60 35 0 0.90 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 *BATCH 69 221 7 0.00 2.3 3.5 1.9 0.5 0.0 0.2 0.0 0.0 *STC 72 367 6 0.00 0.9 4.1 0.3 0.8 0.0 0.7 0.0 0.1 *ASCH 0 0 0.00 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 *OMVS 86 21 0 0.00 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 *ENCLAVE 77 16 N/A N/A 0.4 N/A 0.1 N/A 0.0 N/A N/A N/A BATCH W 67 15 6 48855 0.23 1.8 2.5 1.6 0.3 0.0 0.1 0.0 0.0 HOTBATCH S 44 2 0 32983 0.07 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 PRODBTCH S 64 7 3 74999 0.10 0.8 1.6 1.1 0.2 0.0 0.1 0.0 0.0 ONLINE W 77 227 4 282.5 87.48 0.6 2.6 0.3 0.6 0.0 0.1 0.0 0.0 CICS S 0 0 0.773 0.41 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 OTH_ONL S 0 0 23.02 0.31 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ONLINE S 0 0 41.21 85.55 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ONLTASKS S 77 227 4 17407 1.22 0.6 2.6 0.3 0.6 0.0 0.1 0.0 0.0 OS390 W 47 75 1 7801 0.21 0.2 0.8 0.2 0.2 0.0 0.7 0.0 0.1 WASz S 70 19 0 438K 0.00 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 RELAXED S 46 51 1 0.000 0.00 0.1 0.8 0.1 0.2 0.0 0.7 0.0 0.1 But look a little deeper, and we can see that:
3. Sysplex SummaryThe previous screen gave a quick view of the performance of the local z/OS system. Sometimes a Sysplex wide view of performance is handy. The Sysplex Summary helps out here: RMF V2R4 Sysplex Summary - PLXTNKX Line 1 of 432 Command ===> Scroll ===> CSR WLM Samples: 480 Systems: 5 Date: 08/15/22 Time: 05.53.00 Range: 120 Sec >>>>>>>>------------------<<<<<<<< Service Definition: wlmdef01 Installed at: 04/24/22, 01.13.39 Active Policy: POLICYA Activated at: 04/24/22, 01.13.39 ------- Goals versus Actuals -------- Trans --Avg. Resp. Time- Exec Vel --- Response Time --- Perf Ended WAIT EXECUT ACTUAL Name T I Goal Act ---Goal--- --Actual-- Indx Rate Time Time Time BATCH W 50 0.100 933.7 694.9 1454 PRDBATMD S 4 40 50 0.80 0.067 742.4 546.8 1289 TBATDISC S 0.0 0.033 1316 990.9 1783 1 5 15 0.0 N/A 0.033 1316 990.9 1783 CICS W 47 4.050 0.000 4208 4208 CICSDMED S 4 50 10000 90% N/A N/A 0.000 0.000 0.000 0.000 CICSMED S 2 78 1000 80% 90% 0.50 1.850 0.000 1264 1264 CICSREGN S 2 35 75 0.47 0.000 CICSUNC S 3 48 1000 80% 94% 0.50 2.200 0.000 6684 6684 CICSVEL2 S 3 10 35 0.29 0.000 A lot of similar information, but at the Sysplex level. In particular, I like to look at the WLM performance index: this is a real quick way of determining if there are potential performance problems happening. In this case, all performance indexes are less than 1: everything is humming along nicely. 4. Job DelayIn many ways, the Job Delay screen has been one of the best features of RMF Monitor III. When there is a performance issue with a job or started task, this will tell me why. RMF V2R4 Job Delays Command ===> Scroll ===> CSR Samples: System: MVS1 Date: 08/15/22 Time: 03.04.13 Range: 100 Sec Job: PRDCICS Primary delay: Job is waiting to use the processor. Probable causes: 1) Higher priority work is using the system. 2) Improperly tuned dispatching priorities. ------------------------- Jobs Holding the Processor -------------------------- Job: DBP1DDF Job: PRDIMS Job: PRDBAT01 Holding: 45% Holding: 13% Holding: 3% PROC Using: 48% PROC Using: 12% PROC Using: 4% DEV Using: 1% DEV Using: 4% DEV Using: 3% --------------------------- Job Performance Summary --------------------------- Service WFL -Using%- DLY IDL UKN ---- % Delayed for ---- Primary CX ASID Class P Cr % PRC DEV % % % PRC DEV STR SUB OPR ENQ Reason SO 0182 CICSPRD 1 64 28 1 20 0 68 20 0 0 0 0 0 DBP1DDF In this case, our CICS region is waiting for CPU. The biggest 'hog' is DBP1DDF (A Db2 DIST address space) and PRDIMS. The Job Delay screen is the first screen I look at if there is a current performance problem. 5. Coupling FacilityCoupling facilities are important, but often forgotten. The RMF Monitor III CF Overview screen is great to get quick information about coupling facilities used: RMF V2R4 CF Overview - PRDPLX1 Line 1 of 2 Command ===> Scroll ===> CSR Samples: 120 Systems: 5 Date: 08/15/22 Time: 05.53.00 Range: 120 Sec CF Policy: POLICYA Activated at: 08/10/22 00.47.46 --- Coupling Facility --- ----- Processor ------- Req - Storage - --- SCM --- Name Type Mod Lvl Dyn Util% Def Shr Wgt Eff Rate Size Avail Size Avail CF001 8561 T01 24 OFF 34.7 1 0 1.0 7622 110G 55G 144G 144G CF002 8561 T01 24 OFF 35.3 1 0 1.0 8498 110G 52G 144G 144G In this example, our Parallel Sysplex has two coupling facilities: each with one dedicated ICF processor (the Def column is 1: one defined. The Shr column is 0: none shared). Their CPU usage is a little high: around 35%. They are around half full, which is about a full as you want to get with coupling facilities. Sometimes I want to quickly see how a site is using the coupling facility: do they have VTAM Generic resources, or not? Do they have an active Db2 data sharing group, or not? The RMF Monitor III CF Activity view is great for this: showing each structure, what it is, and how much it is used. RMF V2R4 CF Activity - PRDPLX1 Line 238 of 252 Command ===> Scroll ===> CSR Samples: 120 Systems: 5 Date: 08/15/22 Time: 05.53.00 Range: 120 Sec CF: ALL Type ST E System CF --- Sync --- -------- Async ------- Util Rate Avg Rate Avg Chg Del Structure Name % Serv Serv % % DB2G_GBP1 CACHE AP *ALL 1.3 645 13 213.2 222 0.0 0.0 DB2G_GBP1 CACHE AS *ALL 1.3 645 13 213.2 222 0.0 0.0 ISGLOCK LOCK A - *ALL 0.0 9.2 130 1.9 561 0.0 0.0 RLS_CACHE01 CACHE A N *ALL 0.0 1.0 74 0.7 161 0.0 0.0 RLS_LOCK01 LOCK A - *ALL 0.0 0.4 100 <0.1 33 0.0 0.0 RRSD_MAIN LIST A N *ALL 0.0 0.1 262 0.3 226 0.0 0.0 RRSD_RESTART LIST A N *ALL 0.0 1.1 33 1.5 219 0.0 0.0 RRSD_RMDATA LIST A N *ALL 0.0 0.2 246 0.3 229 0.0 0.0 RRSQ_ARCHIVE LIST A N *ALL 0.0 0.2 37 0.1 656 0.0 0.0 RRSQ_DELAYED LIST A N *ALL 0.0 0.4 29 0.1 738 0.0 0.0 RRSQ_MAIN LIST A N *ALL 0.0 0.3 140 0.1 495 0.0 0.0 RRSQ_RESTART LIST A N *ALL 0.0 0.2 172 0.1 381 0.0 0.0 RRSQ_RMDATA LIST A N *ALL 0.0 0.2 173 0.1 166 0.0 0.0 SYSARC_PRCRQ_RCL LIST A N *ALL 0.0 0.0 0 0.0 0 0.0 0.0 SYSIGGCAS_ECS CACHE A N *ALL 0.0 0.0 0 0.0 0 0.0 0.0 SYSPLEX_OPERLOG LIST A N *ALL 0.0 7.6 53 3.1 656 0.0 0.0 What can we see in this example?
Using RMF Monitor IIISo, five (well, six if you include both coupling facility screens) RMF Monitor III screens that quickly give me a lot of information about a system: the CEC model, z/OS version, some features used, and if there are performance problems or not. All the RMF Monitor III screens are for a short period of time: 5 minutes in all of the above examples. So, RMF Monitor III is only good for a quick 'feel' of what's happening. If I see something interesting (like a performance issue), then I'll look at a longer period (possibly using the RMF Postprocessor) to see if this is just a 'blip', or something real. Yes, I love RMF Monitor III. |