今天主要介绍Oracle数据库在并行操作过程中 slave 进程和 QC 进程经常遇到的等待事件及常用脚本。
一、PX Deq: Execution Msg,PX Deq: Execute Reply等待事件
1. PX Deq: Execution Msg
Occurs when a parallel slave is waiting to be told what to do. This is normally considered an idle event, but can cause excessive CPU in some cases. |
该事件是并行查询中的常见事件。当PQ slave进程在等待QC告诉它要做什么的时候就会出现此事件(eg: when waiting to be told parse / execute / fetch etc..)
v$session_wait 中该等待事件对应的参数:
- P1 = sleeptime/senderid
- P2 = passes
- P3 = not used
我们可以使用如下语句获取转换sleeptime/senderid的相关信息:
setSERVEROUTPUTon undefp1 declare instvarchar(20); sendervarchar(20); begin selectbitand(&&p1,16711680)-65535asSNDRINST, decode(bitand(&&p1,65535),65535,'QC','P'||to_char(bitand(&&p1,65535),'fm000'))asSNDR intoinst,sender fromdual wherebitand(&&p1,268435456)=268435456; dbms_output.put_line('Instance='||inst); dbms_output.put_line('Sender='||sender); end; /
如果P1的值为空,则意味slave 不需要等待任何进程
比如p1的值为268501004,则上面的sql会返回:
Instance=1 Sender=P012
passes 进程在得到信息之前循环轮转等待的次数
该等待事件是一个空闲等待事件,当此等待事件出现,进程会持续等待并逐渐增加等待次数直到获取信息!
解决方法:
作为 Coordinator 的 Process 在获取 Slave 进程的数据时,反应太慢了,导致某些 Slave进行因为 Queue 满而不得不等待,进而拖慢了整个并行执行的速度。
这常常是由于 CPU 数目不足或者 系统中运行的 进程太多导致。可考虑 减小并行度。
2. PX Deq: Execute Reply
Occurs when the query coordinator is waiting for a response from a parallel slave. This is normally considered an idle event, but can cause excessive CPU in some cases.
Waiting Process: QC |
协调器正在等待一个 从slaves 进程对控制信息的响应(确认通知)或者期望从slave进程集中获取数据。这个等待事件意味着QC等待slaves结束执行sql 并且将结果集发送给QC
v$session_wait 中该等待事件对应的参数:
- P1 = sleeptime/senderid
- P2 = passes
- P3 = not used
我们可以使用如下语句获取转换sleeptime/senderid的相关信息:
setSERVEROUTPUTon undefp1 declare instvarchar(20); sendervarchar(20); begin selectbitand(&&p1,16711680)-65535asSNDRINST, decode(bitand(&&p1,65535),65535,'QC','P'||to_char(bitand(&&p1,65535),'fm000'))asSNDR intoinst,sender fromdual wherebitand(&&p1,268435456)=268435456; dbms_output.put_line('Instance='||inst); dbms_output.put_line('Sender='||sender); end; /
如果P1的值为空,则意味slave 不需要等待任何进程
比如p1的值为268501004,则上面的sql会返回:
Instance=1 Sender=P012
等待时间:这是非空闲等待时间,QC 等待从slave 的响应或者查询的数据结果
解决办法:非优化的sql语句肯能是导致此等待事件的原因:slaves 需要花费很长时间来执行sql 语句而qc又在等待slave返回数据。
优化sql,查看slave 在执行的语句以及其执行计划,并做出尽量的优化,以便减少slave执行sql语句的时间!
二、相关脚本
1. gives an overview of all running parallel queries with all slaves.It shows the if a slave is waiting and for what event it waits.
selectdecode(px.qcinst_id, NULL, username, '-'|| lower(substr(pp.SERVER_NAME,length(pp.SERVER_NAME)-4,4)))"Username", decode(px.qcinst_id,NULL,'QC','(Slave)')"QC/Slave", to_char(px.server_set)"SlaveSet", to_char(s.sid)"SID", to_char(px.inst_id)"SlaveINST", decode(sw.state,'WAITING','WAIT','NOTWAIT')asSTATE, casesw.state WHEN'WAITING'THEN substr(sw.event,1,30) ELSE NULL endaswait_event, decode(px.qcinst_id,NULL,to_char(s.sid),px.qcsid)"QCSID", to_char(px.qcinst_id)"QCINST", px.req_degree"Req.DOP", px.degree"ActualDOP" fromgv$px_sessionpx,gv$sessions,gv$px_processpp,gv$session_waitsw wherepx.sid=s.sid(+) andpx.serial#=s.serial#(+) andpx.inst_id=s.inst_id(+) andpx.sid=pp.sid(+) andpx.serial#=pp.serial#(+) andssw.sid=s.sid andssw.inst_id=s.inst_id orderbydecode(px.QCINST_ID,NULL,px.INST_ID,px.QCINST_ID), px.QCSID, decode(px.SERVER_GROUP,NULL,0,px.SERVER_GROUP), px.SERVER_SET, px.INST_ID/
2. shows for the PX Deq events the processes that are exchange data.
selectsw.SIDasRCVSID, decode(pp.server_name,NULL,'AQC',pp.server_name)asRCVR, sw.inst_idasRCVRINST, casesw.state WHEN'WAITING'THEN substr(sw.event,1,30) ELSE NULL endaswait_event, decode(bitand(p1,65535), 65535, 'QC', 'P'||to_char(bitand(p1,65535),'fm000'))asSNDR, bitand(p1,16711680)-65535asSNDRINST, decode(bitand(p1,65535), 65535, ps.qcsid, (selectsid fromgv$px_process whereserver_name= 'P'||to_char(bitand(sw.p1,65535),'fm000') andinst_id=bitand(sw.p1,16711680)-65535))asSNDRSID, decode(sw.state,'WAITING','WAIT','NOTWAIT')asSTATE fromgv$session_waitsw,gv$px_processpp,gv$px_sessionps wheresw.sid=pp.sid(+) andsw.inst_id=pp.inst_id(+) andsw.sid=ps.sid(+) andsw.inst_id=ps.inst_id(+) andp1text='sleeptime/senderid' andbitand(p1,268435456)=268435456 orderbydecode(ps.QCINST_ID,NULL,ps.INST_ID,ps.QCINST_ID), ps.QCSID, decode(ps.SERVER_GROUP,NULL,0,ps.SERVER_GROUP), ps.SERVER_SET, ps.INST_ID
3. shows for long running processes what are the slaves do.
selectdecode(px.qcinst_id, NULL, username, '-'|| lower(substr(pp.SERVER_NAME,length(pp.SERVER_NAME)-4,4)))"Username", decode(px.qcinst_id,NULL,'QC','(Slave)')"QC/Slave", to_char(px.server_set)"SlaveSet", to_char(px.inst_id)"SlaveINST", substr(opname,1,30)operation_name, substr(target,1,30)target, sofar, totalwork, units, start_time, timestamp, decode(px.qcinst_id,NULL,to_char(s.sid),px.qcsid)"QCSID", to_char(px.qcinst_id)"QCINST" fromgv$px_sessionpx,gv$px_processpp,gv$session_longopss wherepx.sid=s.sid andpx.serial#=s.serial# andpx.inst_id=s.inst_id andpx.sid=pp.sid(+) andpx.serial#=pp.serial#(+) orderbydecode(px.QCINST_ID,NULL,px.INST_ID,px.QCINST_ID), px.QCSID, decode(px.SERVER_GROUP,NULL,0,px.SERVER_GROUP), px.SERVER_SET, px.INST_ID