A propósito. The latest PERFKIT APAR VM64717 allows to track all share issues. Your case:
** *cp set share ptktsta4 abs 30% abs 30% limith* 02:15:11 USER PTKTSTA4: CP ABSOLUTE SHARE = 30% 02:15:11 MAXIMUM SHARE = LIMITHARD ABSOLUTE 30% 02:15:11 ZAAP ABSOLUTE SHARE = 30% 02:15:11 MAXIMUM SHARE = LIMITHARD ABSOLUTE 30% 02:15:11 IFL ABSOLUTE SHARE = 30% 02:15:11 MAXIMUM SHARE = LIMITHARD ABSOLUTE 30% 02:15:11 ICF ABSOLUTE SHARE = 30% 02:15:11 MAXIMUM SHARE = LIMITHARD ABSOLUTE 30% 02:15:11 ZIIP ABSOLUTE SHARE = 30% 02:15:11 MAXIMUM SHARE = LIMITHARD ABSOLUTE 30% *menu* *user ptktsta4* FCX115 CPU 2097 SER 0312E Interval 02:15:57 - 02:16:00 Perf. Monitor Detailed data for user PTKTSTA4 Total CPU : .0% Storage def. : 256MB Page fault rate: .0/s Superv. CPU : .0% Resident <2GB : 4506 Page read rate : .0/s Emulat. CPU : .0% Resident >2GB : 11980 Page write rate: .0/s I/O rate : .0/s Proj. WSET : 16486 Pgs moved >2GB>: .0/s DASD IO rate : .0/s Reserved pgs : 0 Main > XSTORE : .0/s UR I/O rate : .0/s Locked pages : 0 XSTORE > main : .0/s Diag. X'98' : .0/s XSTORE dedic. : 2MB XSTORE > DASD : .0/s *BLOCKIO : .0/s XSTORE pages : 0 SPOOL pg reads : .0/s Last IPL : DEV E000 DASD slots : 0 SPOOL pg writes: .0/s Config mode : LINUX IUCV X-fer/s : .0/s MDC insert rate: .0/s Base CPU type : IFL Share : 30% MDC I/O avoided: .0/s Base CPU affin.: ON Max. share : 30%H #I/O active : 0 Active : 0% PSW wait : ..% I/O act. : ..% Stacked blk : .. Page wait: ..% CF wait : ..% Eligible : ..% Stat.: EME,P03,DORM I/O wait : ..% Sim. wait: ..% Runnable : ..% Proc. %CPU %CP %EM IO/S Status Type Aff SRel. SAbs. Limit MRel. MAbs. 00 BI .0 .0 .0 .0 EME,P03,DORM IFL ON ..... 10.002 Hard ..... 10.002 01 .0 .0 .0 .0 EME,P03,DORM IFL ON ..... 10 Hard ..... 10 02 .0 .0 .0 .0 EME,P03,DORM IFL ON ..... 10 Hard ..... 10 The 10.002 value is a standard dec-hex conversion trouble. 2011/4/11 John Hanley <jhan...@courts.state.va.us> > Maxim, > Thanks for the reply. > So if the system has 6 real IFL processors and 2 real CP processors > with the virtual guest as VCONFIG=LINUX , CPUAFFINITY = ON and 3 virtual > IFL processors . > Then use 'SET SHARE guest abs 30% abs 30% limith', each virtual IFL will > get a 10% share? > -- WBR, =Maxim Bochagov IBM Russia, PTK development team
