HyperSwap
HyperSwap对跨Sysplex的共享数据的影响1. 测试场景
1.1 测试环境
硬件: 1台z990主机/2台DS8000/FICON连接
软件: NetView v5.1/SA v2.2/GDPS PPRC v3.2
1.2 系统结构
PLEXP1
- 包含BP0M/BP0D/BP0J系统, 其中BP0M是GDPS/PPRC的K系统
- 使用两台DS8000作为存储设备
- 系统看到的地址在主盘是8800-8F20
- 系统看到的地址在从盘是9800-9F1F, 主盘到从盘形成PPRC Duplex关系
CSCPLEXT
- 包含T101/T102系统
- 使用的存储设备包括ESS800和PLEXP1共用主盘的DS8000
- 系统看到ESS800上的盘地址范围是B000-B15F
- 系统看到DS8000上的盘地址范围是8800-8F20, 即PLEXP1系统的主盘
PLEXP1和CSCPLEXT这两个不同的Sysplex, 具备跨Sysplex的数据共享能力.
1.3 GEOPARM
1.4 测试程序
1.4.1 生成测试数据
PLEXP1系统运行如下Rexx产生用于测试数据的跨Sysplex共享数据:
通过以下JCL提交Rexx来生成测试数据:
产生的测试数据如下所示:
1.4.2 跨Sysplex读取数据
在CSCPLEXT系统上调用如下Rexx读取在PLEXP1系统上生成的共享数据, 并保存在CSCPLEXT所有的ESS800的卷上.
通过以下JCL调用Rexx程序:
2. 测试流程和结果分析
(1) 在BP0D系统上提交作业BJSP02.REXX.EXEC(@2JWRTDS)产生测试数据
(2) 在T101系统提交作业, 读取共享数据:
TSCLJ.REXX.EXEC(@3JRDDS5)
TSCLJ.REXX.EXEC(@3JRDDS6)
(3) 执行GDPS脚本DASD1_MAINT, 进行HyperSwap操作.
SDF LOG
17:03:04 SYSPLEX='CDS SITE2' STARTED
17:03:05 SYSPLEX='CDS SITE2' ENDED RC=0
17:03:05 DASD='SWITCH HYPERSW AP SUSPEND' STARTED
17:03:08 HYPERSW AP RESUME COMPLETE
17:03:36 DASD='SWITCH HYPERSW AP SUSPEND' ENDED RC=0 17:03:36 IPLTYPE='BP0D MODE=SITE2' STARTED
17:03:36 IPLTYPE='BP0D MODE=SITE2' ENDED RC=0
17:03:36 IPLTYPE='BP0H MODE=SITE2' STARTED
17:03:36 IPLTYPE='BP0H MODE=SITE2' ENDED RC=0
17:03:36 IPLTYPE='BP0J MODE=SITE2' STARTED
17:03:36 IPLTYPE='BP0J MODE=SITE2' ENDED RC=0
17:03:36 DASD1_MAINT PLANNED/STANDARD ACTION ENDED
NetView Log
17:03:05 M ===> GEO554I HYPERSWAP PHASE 3 - QUIESCE DISK I/O 17:03:06 C VPCEACTB BA TCH 1 CGROUP 800808 909808 FREEZE 17:03:06 C VPCEACTB BA TCH 2 CGROUP 800909 909909 FREEZE 17:03:06 C VPCEACTB BA TCH 3 CGROUP 800A0A 909A0A FREEZE 17:03:06 C VPCEACTB BA TCH 4 CGROUP 800B0B 909B0B FREEZE 17:03:06 C VPCEACTB BA TCH 5 CGROUP 800C0C 909C0C FREEZE 17:03:06 C VPCEACTB BA TCH 6 CGROUP 800D0D 909D0D FREEZE 17:03:06 C VPCEACTB BA TCH 7 CGROUP 800E0E 909E0E FREEZE
17:03:06 C VPCEACTB BA TCH 8 CGROUP 800F0F 909F0F FREEZE
……
17:03:08 M ===> GEO555I HYPERSWAP PHASE 3 - QUIESCE COMPLETE (4) 在跨Sysplex共享磁盘Freeze的时候, 读取共享数据出现停顿
从SDSF 输出里看到, 程序在停顿2分钟后继续运行:
→Stop Position
Current Time:17:03:06.150518 ; Current line:3820
→Restart Position
Current Time:17:05:06.275556 ; Current line:3821
→Duration
2分钟
(5) 执行脚本RESYNC#, 完成场地2磁盘到场地1磁盘的PPRC再同步SDF LOG:
17:12:21 DASD='START SECONDARY NOFLASH' STARTED
17:13:48 MIRRORING STATUS IS NOW OK
17:13:49 DASD='START SECONDARY NOFLASH' ENDED RC=0
17:13:49 RESYNC# PLANNED/STANDARD ACTION ENDED
(6) 程序执行完毕, 观察程序结果
→切换时刻快照
TSCLJ.ICBC.TEST.DATA1片断:
R.7633 RT.11:00:59.446607 >>> W.3817 WT.17:03:06.055992
R.7635 RT.11:00:59.446677 >>> W.3818 WT.17:03:06.086105
R.7637 RT.11:00:59.446746 >>> W.3819 WT.17:03:06.118951
R.7639 RT.11:00:59.446816 >>> W.3820 WT.17:03:06.150518
R.7641 RT.11:00:59.446885 >>> W.3821 WT.17:05:06.275556
R.7643 RT.11:00:59.446954 >>> W.3822 WT.17:05:06.307117
R.7645 RT.11:00:59.447025 >>> W.3823 WT.17:05:06.340359
R.7647 RT.11:00:59.447094 >>> W.3824 WT.17:05:06.373582
TSCLJ.ICBC.TEST.DATA2片断:
R.7624 RT.11:00:59.446294 >>> W.3812 WT.17:03:06.045963
R.7626 RT.11:00:59.446364 >>> W.3813 WT.17:03:06.076458
R.7628 RT.11:00:59.446433 >>> W.3814 WT.17:03:06.105915
R.7630 RT.11:00:59.446503 >>> W.3815 WT.17:03:06.134876
R.7632 RT.11:00:59.446572 >>> W.3816 WT.17:05:06.228132
R.7634 RT.11:00:59.446642 >>> W.3817 WT.17:05:06.289309
R.7636 RT.11:00:59.446711 >>> W.3818 WT.17:05:06.324148
R.7638 RT.11:00:59.446780 >>> W.3819 WT.17:05:06.361119
R.7640 RT.11:00:59.446850 >>> W.3820 WT.17:05:06.402875
Resync开始时刻快照(SDSF作业输出)
Current Time:17:12:42.175123 ; Current line:13520
Current Time:17:12:42.269679 ; Current line:13521
Current Time:17:12:42.385770 ; Current line:13522
Current Time:17:12:42.475666 ; Current line:13523
IRX0562E Abnormal completion of the GET Data Management macro.
IRX0565E @3JRDDS6,STEP1 ,8916,D,INDD2 ,GET ,COMMAND REJECT ,000005BA00000 0,QSAM.
IRX0250E System abend code 001, reason code 00000003.
IRX0255E Abend in host command execio or address environment routine TSO.
IRX0670E EXECIO error while trying to GET or PUT a record.
4 *-* "execio 1 diskr indd2 "i" (finis"
+++ RC(20) +++
对比NetView Log, 发现程序出现System Abend的时刻正是Resync开始的时候
17:12:42 E IEA494I 981C,BPNV00,PPRC PAIR PENDING,SSID=9098,CCA=1C
17:12:42 E IEA494I 9A1E,DBLG03,PPRC PAIR PENDING,SSID=909A,CCA=1E
17:12:42 E IEA494I 981D,BPCIC1,PPRC PAIR PENDING,SSID=9098,CCA=1D
17:12:42 E IEA494I 9A1F,BPWK03,PPRC PAIR PENDING,SSID=909A,CCA=1F
17:12:42 E IEA494I 981E,DBLG01,PPRC PAIR PENDING,SSID=9098,CCA=1E
17:12:42 M ===> VPCEHSCS: FAILBACK SUCCESSFUL IN BP0M AT 17:12:42
17:12:42 E IEA494I 981F,BPWK01,PPRC PAIR PENDING,SSID=9098,CCA=1F
17:12:43 M ===> VPCEHSCS: DISABLE SUCCESSFUL IN BP0M A T 17:12:43
17:12:43 - CNM493I GEOMSGGP : 00799000 : VPCEMSGK 2 9C00
17:12:43 E IEA494I 9C00,BPSA09,PPRC PAIR FULL DUPLEX,SSID=909C,CCA=00
17:12:43 - CNM493I GEOMSGGP : 00799000 : VPCEMSGK 2 9E01
17:12:43 E IEA494I 9E01,BPPM05,PPRC PAIR FULL DUPLEX,SSID=909E,CCA=01
17:12:43 - CNM493I GEOMSGGP : 00799000 : VPCEMSGK 2 9A00
程序失败原因参考
(7) 执行DASD1_MAINT_BK脚本, 恢复测试前GDPS/PPRC关系
SDF LOG:
17:15:54 DASD='SWITCH HYPERSW AP SUSPEND' STARTED
17:15:57 HYPERSW AP RESUME COMPLETE
17:16:26 DASD='SWITCH HYPERSW AP SUSPEND' ENDED RC=0 17:16:27 SYSPLEX='CDS NORMAL' STARTED
17:16:58 SYSPLEX='CDS NORMAL' ENDED RC=0
17:16:58 IPLTYPE='BP0H MODE=SITE1' STARTED
17:16:59 IPLTYPE='BP0H MODE=SITE1' ENDED RC=0
17:16:59 IPLTYPE='BP0J MODE=SITE1' STARTED
17:16:59 IPLTYPE='BP0J MODE=SITE1' ENDED RC=0
17:16:59 IPLTYPE='BP0D MODE=SITE1' STARTED
17:16:59 IPLTYPE='BP0D MODE=SITE1' ENDED RC=0
17:16:59 DASD='START SECONDARY NOFLASH' STARTED
17:18:27 MIRRORING STATUS IS NOW OK
17:18:27 DASD='START SECONDARY NOFLASH' ENDED RC=0
17:18:27 DASD1_MAINT_BK PLANNED/STANDARD ACTION ENDED