致命错误 20 asm大量并行进程的遭遇

实例宕机，提示大量ora-20错误，这个我会，不就是连接数多吗？应用干的

看看告警日志，

2021-06-21T01:14:24.831701+08:00
NOTE: cleaning up empty system-created directory '+OCRDG/dbcluster/OCRBACKUP/backup01.ocr.265.1075742055'
NOTE: cleaning up empty system-created directory '+OCRDG/dbcluster/OCRBACKUP/backup00.ocr.259.1075756459'
NOTE: cleaning up empty system-created directory '+OCRDG/dbcluster/OCRBACKUP/16625263.267.1075770861'
2021-06-21T03:22:32.978113+08:00
ORA-00020: maximum number of processes (1024) exceeded
ORA-20 errors will not be written to the alert log for
the next minute. Please look at trace files to see all
the ORA-20 errors.
Process MZ00 submission failed with error = 20
2021-06-21T03:22:55.574149+08:00
Errors in file /u01/app/grid/diag/asm/+asm/+ASM1/trace/+ASM1_ora_229943.trc  (incident=57777):
ORA-00603: ORACLE server session terminated by fatal error
ORA-00020: maximum number of processes () exceeded
Incident details in: /u01/app/grid/diag/asm/+asm/+ASM1/incident/incdir_57777/+ASM1_ora_229943_i57777.trc
2021-06-21T03:22:56.354968+08:00
opiodr aborting process unknown ospid (229943) as a result of ORA-603
2021-06-21T03:22:56.396737+08:00
Errors in file /u01/app/grid/diag/asm/+asm/+ASM1/trace/+ASM1_ora_229941.trc  (incident=57769):
ORA-00603: ORACLE server session terminated by fatal error
ORA-00020: maximum number of processes () exceeded
Incident details in: /u01/app/grid/diag/asm/+asm/+ASM1/incident/incdir_57769/+ASM1_ora_229941_i57769.trc
2021-06-21T03:22:56.930931+08:00
Errors in file /u01/app/grid/diag/asm/+asm/+ASM1/trace/+ASM1_ora_229941.trc:
ORA-17503: ksfdopn:2 Failed to open file +OCRDG/orapwasm
ORA-00603: ORACLE server session terminated by fatal error
ORA-00020: maximum number of processes () exceeded
2021-06-21T03:22:57.039063+08:00

找到一个trc

Trace file /u01/app/grid/diag/asm/+asm/+ASM1/trace/+ASM1_ora_229943.trc
Oracle Database 19c Enterprise Edition Release 19.0.0.0.0 - Production
Version 19.9.0.0.0
Build label:    RDBMS_19.9.0.0.0DBRU_LINUX.X64_200930
ORACLE_HOME:    /u01/app/19.3.0/grid
System name: Linux
Node name: host
Release: 3.10.0-957.el7.x86_64
Version: #1 SMP Thu Oct 4 20:48:51 UTC 2018
Machine: x86_64
Instance name: +ASM1
Redo thread mounted by this instance: 0
Oracle process number: 1022
Unix process pid: 229943, image: oracle@host (TNS V1-V3)




*** 2021-06-21T03:22:55.574585+08:00
*** SESSION ID:(1037.14999) 2021-06-21T03:22:55.574603+08:00
*** CLIENT ID:() 2021-06-21T03:22:55.574607+08:00
*** SERVICE NAME:() 2021-06-21T03:22:55.574610+08:00
*** MODULE NAME:() 2021-06-21T03:22:55.574613+08:00
*** ACTION NAME:() 2021-06-21T03:22:55.574615+08:00
*** CLIENT DRIVER:() 2021-06-21T03:22:55.574618+08:00

2021-06-21T03:22:55.574435+08:00
Incident 57777 created, dump file: /u01/app/grid/diag/asm/+asm/+ASM1/incident/incdir_57777/+ASM1_ora_229943_i57777.trc
ORA-00603: ORACLE server session terminated by fatal error
ORA-00020: maximum number of processes () exceeded

kgefec: fatal error 20

的确是最大进程数，这里设置的是1024。
但是这是asm实例遇到的最大进程数，谁会连接到+ASM1上面呢？

此时实例2运行正常，+ASM2实例也有600多个asm_pp开头的进程，什么是pp进程，请参考

不讲武德的并行进程命名
http://blog.chinaunix.net/uid-20687159-id-5844801.html

为什么asm有大量并行进程？
bug？

三板斧定位法：
select osuser,username,count(0) from v$session group by osuser,username;
select sid,sql_id,event,status,state,event,last_call_et from v$session;
select pname,program,machine from v$process where addr=(select paddr from v$session where sid=&sid);

发现一个用户执行过gv$视图查询，但是会话一直没断开。
后来应用关闭后，这些asm_pp自动消失。
再观察。