Examination of Virtual Tape Subsystems Ned Diehl The Information Systems Manager, Inc. ned.diehlperfman perfman 610-865

0
0
1736 days ago, 914 views
PowerPoint PPT Presentation
Trademarks (Omissions are inadvertent). ISMPerfManIBMParallel SysplexRMFOS/390. z/OSMagstarVSMStorageTekHSCNearline. Targets. Examine wellsprings of execution investigation and scope quantification information for virtual tape subsystems (VT) in an OS/390 environmentPrimary concentrate on equipment implementationsIBM VTSStorageTek VSMPresent graphical cases System as opposed to application center.

Presentation Transcript

Slide 1

Investigation of Virtual Tape Subsystems Ned Diehl The Information Systems Manager, Inc. ned.diehl@perfman.com www.perfman.com 610-865-0300 NCACMG Vienna, Virginia 5 June 2002

Slide 2

ISM PerfMan IBM Parallel Sysplex RMF OS/390 z/OS Magstar VSM StorageTek HSC Nearline Trademarks (Omissions are accidental)

Slide 3

Objectives Discuss wellsprings of execution examination and scope quantification information for virtual tape subsystems (VT) in an OS/390 condition Primary concentrate on equipment usage IBM VTS StorageTek VSM Present graphical cases System instead of use center

Slide 4

Contents Introduction Virtual Tape Structure Data Sources Key Performance Metrics Recommendations Summary References

Slide 5

Introduction Solves numerous customary tape issues Small informational collections Low movement open informational collections Allocation sit tight for tape drives Weakest with "great" tape action Full volumes High exchange rates Exploits high limit tapes VT evacuates execution obstructions

Slide 6

Introduction Consider debacle recuperation VTS Peer-to-Peer Clustered VTSS Configuration Mount administration contemplations Deferred mount ought to by and large be kept away from Premount can be alluring Not yet everything for everybody – except arriving Fewer issue datasets with current executions

Slide 7

VT Structure General Looks (practically) like 3490E to host Combination of equipment and programming RISC processor Library of high limit tapes RAID DASD cushion or CACHE Tape drives Amount of host programming shifts with usage

Slide 8

VT Structure Logical Components

Slide 9

VT Structure VTS - Physical Components

Slide 10

MVC VTV MVC VTV MVC VTV VT Structure VSM - Physical Architecture HSC VTCS VSM Control & Data Path Control Path CDS Migrate/Recall VTSS Physical tape drives Physical tape volumes Physical Libraries & spaces Virtual tape drives Virtual tape volumes

Slide 11

Data Sources Application SMF 30, 72, 14, & 15 Hardware SMF 94 - tape library measurements STK client SMF record SMF 14 & 15 – informational index action SMF 21 – blunder insights by volume RMF 73 - channel way action RMF 74-1 - gadget action RMF 78-3 - I/O lining Real time controller interface

Slide 12

Data Sources SMF 94 Easy to work with however not adaptable Hourly synopsis from VTS or Library Difficult to synchronize (don't utilize SMF94HHI) Identical information to all joined recording pictures Base sections mirror all library insights ATL, mount, get off, launch, & embed Native and VTS action Identical information in various records with an alternate serial number if VTS and local drives connected VTS fragments reflect single VTS, import/send out, improved measurements

Slide 13

Data Sources SMF 94 One record created for every Library with local drives, distinguished by ATL serial (SMF94SNO). Contains base portions. VTS, recognized by VTS serial (SMF94SNO & SMF94VLS). Contains base and VTS portions. With distributed, three (neighborhood) or four (remote) VTSs included Only two have physical tape No simple approach to partner them Generally reliable with RMF permits better detail of practically identical measurements

Slide 14

Data Sources SMF 94 Data Issues Hour file (SMF94HHI) is interim end Drives accessible (SMF94VTA) not generally reset after administration Drives utilized (SMF94VTV, VTN, & VTX) can be more noteworthy than accessible Alignment issue with early Import/Export insights (SMF94ACA & SMF94ACB) Duplicate library information with an alternate serial number if VTS and local drives joined Average age in reserve (SMF94VCA) has had changed definitions

Slide 15

Data Sources SMF 94 Data Issues Tape information exchange (SMF94VTR & SMF94VTW) revealed as zero in some current specimens Max store volume age (S94MTVCA) recorded as seconds, archived as minutes Backstore pressure proportion (S94BSRAT) some of the time announced as short of what one Recall throttle percent (S94RCPRT) here and there more noteworthy than 100

Slide 16

Data Sources SMF 94 Data Issues Reference Flash 10124 dated 20 November 2001. Reviews may be recorded as hits. Affected measurements: S94MAXCH Maximum Cache Hit Mount Time S94AVGCH Average Cache Hit Mount Time SMF94VMH Number of Cache Hit Mounts S94MAXRM Maximum Recall Mount Time S94AVGRM Average Recall Mount Time SMF94VMS Number of Recall Mounts SMF94VPS Number of Physical Mounts for Recall

Slide 17

Data Sources STK User Record Complex to work with Many subtypes Repeating sections Allows detail examination Configuration information CU and gadget occupied Volume level information VTSS interim of 15 minutes Grouping and synopsis required Some coherent measurements require numerous subtypes

Slide 18

Data Sources STK User Record Interval measurements by VTSS (not picture) to each recording OS/390 Subsystem, channel interface, and RTD execution Values change with OS/390 recording time Logically indistinguishably however physically extraordinary Event measurements recorded once VTV mount, get off, erase, relocate, review, development, reproduce RTD mount, get off, differ MVC status

Slide 19

Data Sources STK Data Issues Must process occasion records from all recording pictures Mount occasions not really recorded to unique asking for picture Dismount occasions not really recorded to same picture as mount Might be additional mounts or descent Calculated circumstances may be expansive or negative Some time values in complex organizations

Slide 20

Data Sources STK Data Issues RMF consistency shifts Most measurements close over different examples I/O Rates Connect time Allocated time Total mount time Significant variety with mount tallies and along these lines normal mount time Probably a code level issue

Slide 21

Key Performance Metrics Application Service levels Hardware Data exchange Device utilization Virtual mount time Mount miss (review) rate Storage use Tape volume age (in DASD cradle or reserve)

Slide 22

Key Performance Metrics Application Performance Should have sensible and quantifiable execution targets Objectives will fluctuate with Installation Application VT Time & day

Slide 23

Key Performance Metrics Data Transfer Good to report work performed Mounts and I/O rates likewise alternatives Saturation differs with condition SMF 14, 15, and 21 give read and compose RMF 73 gives read and keep in touch with FICON. VTS gives information exchange Host read (SMF94VBR) and compose (SMF94VBW) are great throughput pointers Read (SMF94VTR) and compose (SMF94VTW) amongst store and genuine tapes

Slide 24

Key Performance Metrics Data Transfer VTSS gives numerous information exchange measurements however suppositions are required RTD bytes read (SMF20BTR) and composed (SMF20BTW) RTD associate (SMF20DCT) and use (SMF20DUT) Host and RTD channel interface occupied (SMF11CUB) VTV get off has VTV estimate (SMF14VSZ) Probably best throughput marker

Slide 25

Key Performance Metrics Data Transfer

Slide 26

Key Performance Metrics Data Transfer

Slide 27

Key Performance Metrics Device Usage RMF 74-1 gives virtual by gadget or gathering. Permits division of mount and distribution parts (e.g. associate, pend, hold up). VTS gives least, most extreme, normal, and arranged for both virtual and physical gadgets. All qualities are whole numbers. VTS normal (SMF94VTV) and most extreme (SMF94VTX) physical are useful for slanting While not typically an issue, virtual gadget utilize ought not be overlooked.

Slide 28

Key Performance Metrics Device Usage Difficult to figure VTSS least, greatest, normal, and accessible for either virtual or physical gadgets Mount, get off, and differ subtypes must be assembled and compressed RTDs can be statically shared (administrator charge) by different VTSSs and OS/390 VTSS RTD associate and use, which were beforehand examined, are helpful

Slide 29

Key Performance Metrics Device Usage

Slide 30

Key Performance Metrics Device Usage

Slide 31

Key Performance Metrics Device Usage

Slide 32

Key Performance Metrics Device Usage (RMF)

Slide 33

Key Performance Metrics Virtual Mount Time RMF 74-1 gives normal mount time by gadget or gathering VTS gives least, most extreme, and normal virtual and physical mount times Average virtual (SMF94VRA) is useful for slanting High greatest virtual (SMF94VRX) regularly connects with issues

Slide 34

Key Performance Metrics Virtual Mount Time VTSS mount time ascertained from mount occasion records (SMF13MET - SMF13MST) Min, max, normal and tallies require gathering and rundown Average may be misshaped by additional mounts VTSS gives a few conceivably intriguing identifiers Job, step, DSN, VTV administration class Scratch, existing

Slide 35

Key Performance Metrics Virtual Mount Time Mount times tend to increment with: Reduced tape volume age on DASD (SMF94VCA) Increased mount miss rate (SMF94VMS) Increased normal genuine drives mounted (SMF94VTV) Published targets: Daily normal under 30 seconds Hourly normal under 300 seconds Maximum under 900 seconds

Slide 36

Key Performance Metrics Virtual Mount Time (RMF)

Slide 37

Key Performance Metrics Virtual Mount Time

Slide 38

Key Performance Metrics Max Virtual Mount Time

Slide 39

Key Performance Metrics Average Virtual Mount Time

Slide 40

Key Performance Metrics Mount Miss Rate VTS gives mounts by sort Fast prepared (SMF94VFR) Specific mount hits (SMF94VMH) Specific mount misses (SMF94VMS) Related to normal time on DASD (SMF94VCA) and application cycles Published focus of miss rate under 20% Miss % = 100 * Misses/Total Virtual Mounts Miss % = 100 * VMS/(VFR + VMH + VMS)

Slide 41

Key Performance Metrics Mount Miss Rate VTSS mount hit pointer included late refresh (SMF13RCI) Scratch mounts demonstrated (SMF13VMT) If more seasoned information, could accept a hit if computed mount time not as much as an edge

Slide 42

Key Performance Metrics Mount Miss Rate

Slide 43

Key Performance Metrics Specific Mount Miss Rate

Slide 44

Key Performa

SPONSORS