info all檢查到RPEYAII延時較高，而且延時一直在增高
通過找到RPEYAII進程的入庫進程，去找到會話，從而找到會話信息（如sql，event等等），檢查sql是否有問題
如果沒有問題，只有從ogg參數或者分進程的方式去實現延時的降低。
[oracle@lzlhost:lzl:/oggfs]$ ps -ef|grep RPEYAII
oracle 84062 79100 0 09:26 pts/1 00:00:00 grep RPEYAII
oracle 91297 67772 7 Sep07 ? 03:17:21 /oggfs/replicat PARAMFILE /oggfs/dirprm/rpeyaii.prm REPORTFILE /oggfs/dirrpt/RPEYAII.rpt PROCESSID RPEYAII USESUBDIRS AFTERCSN 15890106219844
[oracle@lzlhost:lzl:/oggfs]$ ps -ef|grep 91297
oracle 85714 79100 0 09:29 pts/1 00:00:00 grep 91297
oracle 91297 67772 7 Sep07 ? 03:17:59 /oggfs/replicat PARAMFILE /oggfs/dirprm/rpeyaii.prm REPORTFILE /oggfs/dirrpt/RPEYAII.rpt PROCESSID RPEYAII USESUBDIRS AFTERCSN 15890106219844
oracle 91301 91297 12 Sep07 ? 05:10:30 oracleepmsc2 (DESCRIPTION=(LOCAL=YES)(ADDRESS=(PROTOCOL=beq)))
入庫進程找到了，91301
然後通過進程號去找會話
[oracle@lzlhost:lzl:/oggfs]$
set lin 200
col event for a30
col program for a30
select sid,username,sql_id,prev_sql_id,event,program,status from v$session where paddr=(select addr from v$process where spid=91301);
SID USERNAME SQL_ID PREV_SQL_ID EVENT PROGRAM STATUS
---------- ------------------------------ ------------- ------------- ------------------------------ ------------------------------ --------
2305 GOLDENGATE 6w2gx87g8jzyd 6w2gx87g8jzyd SQL*Net message from client replicat@lzlhost (TNS V1-V3) ACTIVE
用戶和program都可以確定是這個會話沒錯
然後通過sql去找執行計劃，看下執行計劃是否有異常
SQL_ID 6w2gx87g8jzyd, child number 0
-------------------------------------
INSERT /*+ RESTRICT_ALL_REF_CONS */ INTO "EPM_SC"."R_MID_TABLE"
("DATA_ID","MR_ID","APP_NO","MR_SECT_NO","MR_SN","AMT_YM","CONS_NO","CON
S_NAME","ELEC_ADDR","CONTACT_NAME","TEL","MOBILE","MP_ID","MADE_NO","MET
ER_ID","MR_DIGIT","READ_TYPE_CODE","LAST_MR_NUM","THIS_READ","T_FACTOR",
"THIS_READ_PQ","LAST_MR_PQ","MR_STATUS_CODE","EXCP_TYPE_CODE","THIS_YMD"
,"AVG_PQ","ALARM_FLUC_VALUE","CALC_FLAG","REMAIN_PQ1","ADJ_PQ","SETTLE_P
Q","T_PRC","THIS_RCVBLE_AMT","T_OWE","CARD_REMAIN_PQ","AMR_FLAG","POWERO
FF_CODE","METER_DATE","MODIFY_TIMES","ADJUST_MODE","CALC_ID","ORG_NO","B
IN_MT_POS","BAR_CODE","ADD_ELEC","LOCKED_READ","MR_SHARE","ELEC_TYPE_COD
E","PRC_CODE","ACT_MR_MODE") VALUES
(:a0,:a1,:a2,:a3,:a4,:a5,:a6,:a7,:a8,:a9,:a10,:a11,:a12,:a13,:a14,:a15,:
a16,:a17,:a18,:a19,:a20,:a21,:a22,:a23,:a24,:a25,:a26,:a27,:a28,:a29,:a3
0,:a31,:a32,:a33,:a34,:a35,:a36,:a37,:a38,:a39,:a40,:a41,:a42,:a43,:a44,
:a45,:a46,:a47,:a48,:a49)
-------------------------------------------------
| Id | Operation | Name | Cost |
-------------------------------------------------
| 0 | INSERT STATEMENT | | 1 |
| 1 | LOAD TABLE CONVENTIONAL | | |
-------------------------------------------------
PLAN GET DISK WRITE ROWS ROWS USER_IO(MS) ELA(MS) CPU(MS) CLUSTER(MS) PLSQL
END_TI I NAME HASH VALUE EXEC PRE EXEC PRE EXEC PER EXEC ROW_P PRE EXEC PRE FETCH PER EXEC PRE EXEC PRE EXEC PER EXEC PER EXEC
------ - --------------- ------------- ---------- ------------ -------- -------- ----- ----------- --------- ----------- -------- -------- ----------- --------
07 16 2 GOLDENGATE 0 43.W 16 0 0 43.W 1 0 19 103 47 0 0
07 17 2 GOLDENGATE 0 296W 16 0 0 296W 1 0 17 102 47 0 0
07 18 2 GOLDENGATE 0 71.W 16 0 0 71.W 1 0 3 84 46 0 0
07 19 2 GOLDENGATE 0 4.1W 16 0 0 4.1W 1 0 0 79 45 0 0
07 20 2 GOLDENGATE 0 4.1W 16 0 0 4.1W 1 0 0 78 44 0 0
07 21 2 GOLDENGATE 0 4.1W 16 0 0 4.1W 1 0 0 78 45 0 0
07 22 2 GOLDENGATE 0 4.1W 16 0 0 4.1W 1 0 0 79 45 0 0
07 23 2 GOLDENGATE 0 4.1W 16 0 0 4.1W 1 0 0 79 45 0 0
08 00 2 GOLDENGATE 0 4.1W 16 0 0 4.1W 1 0 0 79 45 0 0
08 01 2 GOLDENGATE 0 642W 16 0 0 642W 1 0 7 89 46 0 0
08 02 2 GOLDENGATE 0 613W 16 0 0 613W 1 0 2 87 48 0 0
執行計劃沒有優化空間，在數據庫層面無法優化了
進程拆分，將R_MID_TABLE單獨拿出來
用stats命令檢查RPEYAII中哪些表事務頻繁，將這些表單獨拿出來建立複製進程
GGSCI (lzlhost) 7> stats RPEYAII total
Sending STATS request to REPLICAT RPEYAII ...
Start of Statistics at 2019-09-09 14:50:20.
Replicating from EPM_SC.E_CONS_SNAP to EPM_SC.E_CONS_SNAP:
*** Total statistics since 2019-09-09 11:16:03 ***
Total inserts 21.00
Total updates 9345414.00
Total deletes 1475873.00
Total discards 0.00
Total operations 10821308.00
Replicating from EPM_SC.E_KWH_AMT to EPM_SC.E_KWH_AMT:
...
從total operations看出操作頻繁的表
--將下述操作頻繁的表排除
LZL.E_CONS_SNAP to LZL.E_CONS_SNAP
LZL.E_KWH_AMT to LZL.E_KWH_AMT
LZL.E_PL_AMT to LZL.E_PL_AMT
LZL.E_MP_PQ to LZL.E_MP_PQ
LZL.E_CONS_PRC_AMT to LZL.E_CONS_PRC_AMT
新增進程
add replicat RPEYAIN,exttrail ./dirdat/ai,checkpointtable goldengate.checkpoint
add replicat RPEYAIO,exttrail ./dirdat/ai,checkpointtable goldengate.checkpoint
alter RPEYAIN,extseqno 751,extrba 219048117 （seq和rba由info拆分的主進程查看）
alter RPEYAIO,extseqno 751,extrba 219048117
start進程檢查延時情況