adrci: A survival guide for the DBA
Starting with 11gR1, we have a new way to deal with Oracle Errors & Tracefiles: There is now a special command line utility dedicated for that purpose called adrci (Automatic Diagnostic Repository Command Interpreter). This posting is intended to show you the (in my view) essential commands, a DBA ought to know in order to use it. We will look at
- Viewing the alert.log
- The relation between incident & problem
- Creation of Packages & ZIP files to send to Oracle Support
- Managing, especially purging tracefiles
I will at first create a problem. Don’t do that with your Production Database! Especially: Never do DML on dictionary tables!
[oracle@uhesse ~]$ sqlplus / as sysdba
SQL*Plus: Release 11.2.0.2.0 Production on Wed Jun 1 10:25:06 2011 Copyright (c) 1982, 2010, Oracle. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production With the Partitioning, OLAP, Data Mining and Real Application Testing options
SQL> select * from v$version;
BANNER --------------------------------------------------------------------------------
Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production PL/SQL Release 11.2.0.2.0 - Production CORE 11.2.0.2.0
Production TNS for Linux: Version 11.2.0.2.0 - Production NLSRTL Version 11.2.0.2.0 - Production
SQL> show parameter diagnostic
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
diagnostic_dest string /u01/app/oracle
SQL> grant dba to adam identified by adam;
Grant succeeded. SQL> connect adam/adam
Connected.
SQL> create table t (n number);
Table created.
SQL> select object_id from user_objects;
OBJECT_ID ---------- 75719
SQL> connect / as sysdba
Connected.
SQL> update tab$ set cols=2 where obj#=75719;
1 row updated.
SQL> commit;
Commit complete.
SQL> alter system flush shared_pool;
System altered.
SQL> connect adam/adam Connected.
SQL> select * from t;
select * from t *
ERROR at line 1:
ORA-03113: end-of-file on communication channel
Process ID: 2236 Session ID: 29 Serial number: 9
I flushed the Shared Pool to get the Data Dictionary Cache empty. Else the select may not crash the session as it did. Imagine the user calls me now on the phone. Our first idea as an experienced DBA: We look at the alert.log! Right so. Please notice that we now have two different kinds of the alert.log.
One is present in the conventional text format, per OFA in $ORACLE_BASE/diag/rdbms/name of the db/name of the instance/trace This location is determined by the new initialization parameter DIAGNOSTIC_DEST, while BACKGROUND_DUMP_DEST is deprecated in 11g.
1. Viewing the alert.log
The other one is in XML format placed in $ORACLE_BASE/diag/rdbms/name of the db/name of the instance/alert This version of the alert.log is accessed by adrci:
[oracle@uhesse ~]$ adrci ADRCI: Release 11.2.0.2.0 - Production on Wed Jun 1 10:20:08 2011
Copyright (c) 1982, 2009, Oracle and/or its affiliates. All rights reserved.
ADR base = "/u01/app/oracle"
adrci> show home
ADR Homes: diag/tnslsnr/uhesse/listener diag/rdbms/orcl/orcl
Please notice that we have different ADR Homes. In my case only two, because I am not using Grid Infrastructure on this Demo System, else there would be another one. I specify my Database Home first. Then I look at the alert.log. Good news if you are on Windows: Finally, you can tail -f your alert.log also 
adrci> set home diag/rdbms/orcl/orcl
adrci> show alert -tail -f
2011-06-01 10:16:35.337000 +02:00 db_recovery_file_dest_size of 4032 MB is 0.00% used.
This is a user-specified limit on the amount of space that will be used by this database for recovery-related files, and does not reflect the amount of space available in the underlying filesystem or ASM diskgroup. Starting background process CJQ0 CJQ0 started with pid=21, OS id=2204 2011-06-01 10:18:42.668000 +02:00 Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x0] [PC:0x90D891A, qcstda()+702] [flags: 0x0, count: 1] Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_2236.trc (incident=6153): ORA-07445: exception encountered: core dump [qcstda()+702] [SIGSEGV] [ADDR:0x0] [PC:0x90D891A] [Address not mapped to object] [] Incident details in: /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_6153/orcl_ora_2236_i6153.trc Use ADRCI or Support Workbench to package the incident. See Note 411.1 at My Oracle Support for error and packaging details. 2011-06-01 10:18:47.518000 +02:00 Dumping diagnostic data in directory=[cdmp_20110601101847], requested by (instance=1, osid=2236), summary=[incident=6153]. 2011-06-01 10:18:48.727000 +02:00 Sweep [inc][6153]: completed Sweep [inc2][6153]: completed
2. The relation between Incident & Problem
You see the incident was recorded in the alert.log. And it tells you “Use ADRCI or Support Workbench to package the incident.” We will soon see how to do that. First I’d like to explain the relation between incident and problem: An incident is the concrete occurrence of a problem. In other words: The same problem may have multiple incidents. To show that, I will open another terminal and do again a select against the table t, while still tailing the alert log from the first session.
Second terminal:
[oracle@uhesse ~]$ sqlplus adam/adam
SQL*Plus: Release 11.2.0.2.0 Production on Wed Jun 1 10:21:52 2011 Copyright (c) 1982, 2010, Oracle. All rights reserved.
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production With the Partitioning, OLAP, Data Mining and Real Application Testing options
SQL> select * from t where n=42;
select * from t where n=42
* ERROR at line 1: ORA-03113: end-of-file on communication channel Process ID: 2299 Session ID: 36 Serial number: 11
First terminal:
2011-06-01 10:21:31.367000 +02:00 Starting background process SMCO SMCO started with pid=19, OS id=2268 2011-06-01 10:22:08.781000 +02:00 Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x0] [PC:0x90D891A, qcstda()+702] [flags: 0x0, count: 1] Errors in file /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_2299.trc (incident=6201): ORA-07445: exception encountered: core dump [qcstda()+702] [SIGSEGV] [ADDR:0x0] [PC:0x90D891A] [Address not mapped to object] [] Incident details in: /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_6201/orcl_ora_2299_i6201.trc Use ADRCI or Support Workbench to package the incident. See Note 411.1 at My Oracle Support for error and packaging details. 2011-06-01 10:22:11.135000 +02:00 Dumping diagnostic data in directory=[cdmp_20110601102211], requested by (instance=1, osid=2299), summary=[incident=6201]. 2011-06-01 10:22:13.370000 +02:00 Sweep [inc][6201]: completed Sweep [inc2][6201]: completedI have seen the second incident recorded. I exit out of the tail -f with CTRL+C and continue:
adrci> show problem
ADR Home = /u01/app/oracle/diag/rdbms/orcl/orcl:
*************************************************************************
PROBLEM_ID PROBLEM_KEY LAST_INCIDENT LASTINC_TIME -------------------- ----------------------------------------------------------- -------------------- ---------------------------------------- 1 ORA 7445 [qcstda()+702] 6201 2011-06-01 10:22:08.924000 +02:00 1 rows fetched
So I have one problem with the ID 1 and the last incident occurred at 10:22. Are there more?
adrci> show incident
ADR Home = /u01/app/oracle/diag/rdbms/orcl/orcl:
*************************************************************************
INCIDENT_ID PROBLEM_KEY CREATE_TIME
-------------------- ----------------------------------------------------------- ----------------------------------------
6153 ORA 7445 [qcstda()+702] 2011-06-01 10:18:42.995000 +02:00 6201 ORA 7445 [qcstda()+702] 2011-06-01 10:22:08.924000 +02:00 2 rows fetched
I want to see some more detail about the incidents:
adrci> show incident -mode detail -p "incident_id=6201"
ADR Home = /u01/app/oracle/diag/rdbms/orcl/orcl:
************************************************************************* ********************************************************** INCIDENT INFO RECORD 1 ********************************************************** INCIDENT_ID 6201 STATUS ready CREATE_TIME 2011-06-01 10:22:08.924000 +02:00 PROBLEM_ID 1 CLOSE_TIME FLOOD_CONTROLLED none ERROR_FACILITY ORA ERROR_NUMBER 7445 ERROR_ARG1 qcstda()+702 ERROR_ARG2 SIGSEGV ERROR_ARG3 ADDR:0x0 ERROR_ARG4 PC:0x90D891A ERROR_ARG5 Address not mapped to object ERROR_ARG6 ERROR_ARG7 ERROR_ARG8 ERROR_ARG9 ERROR_ARG10 ERROR_ARG11 ERROR_ARG12 SIGNALLING_COMPONENT SQL_Parser SIGNALLING_SUBCOMPONENT SUSPECT_COMPONENT SUSPECT_SUBCOMPONENT ECID IMPACTS 0 PROBLEM_KEY ORA 7445 [qcstda()+702] FIRST_INCIDENT 6153 FIRSTINC_TIME 2011-06-01 10:18:42.995000 +02:00 LAST_INCIDENT 6201 LASTINC_TIME 2011-06-01 10:22:08.924000 +02:00 IMPACT1 0 IMPACT2 0 IMPACT3 0 IMPACT4 0 KEY_NAME ProcId KEY_VALUE 25.3 KEY_NAME Client ProcId KEY_VALUE oracle@uhesse (TNS V1-V3).2299_140262306875136 KEY_NAME PQ KEY_VALUE (0, 1306916528) KEY_NAME SID KEY_VALUE 36.11 OWNER_ID 1 INCIDENT_FILE /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_2299.trc OWNER_ID 1 INCIDENT_FILE /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_6201/orcl_ora_2299_i6201.trc 1 rows fetched
I want to look at the incident tracefile mentioned above:
adrci> show trace /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_6201/orcl_ora_2299_i6201.trc /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_6201/orcl_ora_2299_i6201.trc ---------------------------------------------------------- LEVEL PAYLOAD ----- ------------------------------------------------------------------------------------------------------------------------------------------------ Dump file /u01/app/oracle/diag/rdbms/orcl/orcl/incident/incdir_6201/orcl_ora_2299_i6201.trc Oracle Database 11g Enterprise Edition Release 11.2.0.2.0 - 64bit Production With the Partitioning, OLAP, Data Mining and Real Application Testing options ORACLE_HOME = /u01/app/oracle/product/11.2.0/db_1 System name: Linux Node name: uhesse Release: 2.6.32-100.28.5.el6.x86_64 Version: #1 SMP Wed Feb 2 18:40:23 EST 2011 Machine: x86_64 Instance name: orcl Redo thread mounted by this instance: 1 Oracle process number: 25 Unix process pid: 2299, image: oracle@uhesse (TNS V1-V3) *** 2011-06-01 10:22:08.929 *** SESSION ID:(36.11) 2011-06-01 10:22:08.929 *** CLIENT ID:() 2011-06-01 10:22:08.929 *** SERVICE NAME:(SYS$USERS) 2011-06-01 10:22:08.929 *** MODULE NAME:(SQL*Plus) 2011-06-01 10:22:08.929 *** ACTION NAME:() 2011-06-01 10:22:08.929 Dump continued from file: /u01/app/oracle/diag/rdbms/orcl/orcl/trace/orcl_ora_2299.trc 1> ***** Error Stack ***** ORA-07445: exception encountered: core dump [qcstda()+702] [SIGSEGV] [ADDR:0x0] [PC:0x90D891A] [Address not mapped to object] [] 1< ***** Error Stack ***** 1> ***** Dump for incident 6201 (ORA 7445 [qcstda()+702]) ***** 2> ***** Beginning of Customized Incident Dump(s) ***** 2> ***** Beginning of Customized Incident Dump(s) ***** Exception [type: SIGSEGV, Address not mapped to object] [ADDR:0x0] [PC:0x90D891A, qcstda()+702] [flags: 0x0, count: 1] Registers: %rax: 0x0000000000000000 %rbx: 0x00007f915c77f0e0 %rcx: 0x0000000000000007 %rdx: 0x0000000000000000 %rdi: 0x00007f915c77be98 %rsi: 0x0000000000000000 %rsp: 0x00007fffc65178e0 %rbp: 0x00007fffc6517960 %r8: 0x0000000000000028 %r9: 0x0000000000002000 %r10: 0x00000000093849c0 %r11: 0x0000000000000168 %r12: 0x00007f915c77ade8 %r13: 0x000000008edbb178 %r14: 0x00007f915c777da0 %r15: 0x00007f915c77ae28 %rip: 0x00000000090d891a %efl: 0x0000000000010246 qcstda()+686 (0x90d890a) mov -0x40(%rbp),%rdi qcstda()+690 (0x90d890e) mov %rdx,0x18(%rbx) qcstda()+694 (0x90d8912) mov 0x60(%r15),%rsi qcstda()+698 (0x90d8916) mov %ecx,0x8(%r15) > qcstda()+702 (0x90d891a) mov %ecx,(%rsi) qcstda()+704 (0x90d891c) mov 0x78(%rdi),%rdx qcstda()+708 (0x90d8920) test %rdx,%rdx qcstda()+711 (0x90d8923) jnz 0x90d8d03 qcstda()+717 (0x90d8929) mov -0x70(%rbp),%rdi *** 2011-06-01 10:22:08.963 dbkedDefDump(): Starting a non-incident diagnostic dump (flags=0x3, level=3, mask=0x0) 3> ***** Current SQL Statement for this session (sql_id=8r222qucmawdt) ***** select * from t where n=42 3< ***** current_sql_statement ***** 3
3. Creation of Packages & ZIP files to send to Oracle Support
I may not be able to solve the problem myself. Oracle Support will help me with that one. I gather all the required information with a method called “Incident Packaging Service” (IPS):
adrci> ips create package problem 1 correlate all
Created package 2 based on problem id 1, correlation level all
This did not yet create a ZIP file and is therefore referred to as “Logical Package”. The ZIP file is generated from the Logical Package that was created:
adrci> ips generate package 2 in "/home/oracle"
Generated package 2 in file /home/oracle/ORA7445qc_20110601112533_COM_1.zip,
mode complete
4. Managing, especially purging of tracefiles
Now to the management of tracefiles. You may notice that 11g creates lots of tracefiles that need to be purged from time to time. In fact, this is done automatically, but you may want to change the default purge policy:
adrci> show tracefile -rt
01-JUN-11 10:31:48 diag/rdbms/orcl/orcl/trace/orcl_mmon_2106.trc
01-JUN-11 09:43:43 diag/rdbms/orcl/orcl/trace/orcl_ckpt_2100.trc
01-JUN-11 09:22:13 diag/rdbms/orcl/orcl/trace/alert_orcl.log
01-JUN-11 09:22:11 diag/rdbms/orcl/orcl/trace/orcl_diag_2088.trc
01-JUN-11 09:22:10 diag/rdbms/orcl/orcl/trace/orcl_ora_2299.trc
01-JUN-11 09:22:10 diag/rdbms/orcl/orcl/incident/incdir_6201/orcl_ora_2299_i6201.trc
01-JUN-11 09:18:47 diag/rdbms/orcl/orcl/trace/orcl_ora_2236.trc
01-JUN-11 09:18:47 diag/rdbms/orcl/orcl/incident/incdir_6153/orcl_ora_2236_i6153.trc
I have already got some tracefiles. How long are they going to stay?
adrci> show control
ADR Home = /u01/app/oracle/diag/rdbms/orcl/orcl:
*************************************************************************
ADRID SHORTP_POLICY LONGP_POLICY LAST_MOD_TIME LAST_AUTOPRG_TIME LAST_MANUPRG_TIME ADRDIR_VERSION ADRSCHM_VERSION ADRSCHMV_SUMMARY ADRALERT_VERSION CREATE_TIME -------------------- -------------------- -------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- -------------------- -------------------- -------------------- -------------------- ---------------------------------------- 1335663986 720 8760 2011-05-27 10:16:46.997118 +02:00 1 2 80 1 2011-05-27 10:16:46.997118 +02:00 1 rows fetched
The ordinary tracefiles will stay for 30 days (720 hours), while files like incident files stay one year (8760 hours) by default. We can change that policy with for example:
adrci> set control (SHORTP_POLICY = 360)
adrci> set control (LONGP_POLICY = 2190)
adrci> show control
ADR Home = /u01/app/oracle/diag/rdbms/orcl/orcl:
*************************************************************************
ADRID SHORTP_POLICY LONGP_POLICY LAST_MOD_TIME LAST_AUTOPRG_TIME LAST_MANUPRG_TIME ADRDIR_VERSION ADRSCHM_VERSION ADRSCHMV_SUMMARY ADRALERT_VERSION CREATE_TIME -------------------- -------------------- -------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- -------------------- -------------------- -------------------- -------------------- ---------------------------------------- 1335663986 360 2190 2011-06-01 11:42:17.208064 +02:00 1 2 80 1 2011-05-27 10:16:46.997118 +02:00 1 rows fetched
Also, we may want to purge tracefiles manually. Following command will manually purge all tracefiles older than 2 days (2880 minutes):
adrci> purge -age 2880 -type trace adrci> show tracefile -rt
01-JUN-11 10:46:54 diag/rdbms/orcl/orcl/trace/orcl_mmon_2106.trc
01-JUN-11 09:43:43 diag/rdbms/orcl/orcl/trace/orcl_ckpt_2100.trc 01-JUN-11 09:22:13 diag/rdbms/orcl/orcl/trace/alert_orcl.log 01-JUN-11 09:22:11 diag/rdbms/orcl/orcl/trace/orcl_diag_2088.trc 01-JUN-11 09:22:10 diag/rdbms/orcl/orcl/incident/incdir_6201/orcl_ora_2299_i6201.trc 01-JUN-11 09:22:10 diag/rdbms/orcl/orcl/trace/orcl_ora_2299.trc 01-JUN-11 09:18:47 diag/rdbms/orcl/orcl/incident/incdir_6153/orcl_ora_2236_i6153.trc 01-JUN-11 09:18:47 diag/rdbms/orcl/orcl/trace/orcl_ora_2236.trc 01-JUN-11 09:17:19 diag/rdbms/orcl/orcl/trace/orcl_dbrm_2090.trc 01-JUN-11 09:16:44 diag/rdbms/orcl/orcl/trace/orcl_j002_2210.trc 01-JUN-11 09:16:30 diag/rdbms/orcl/orcl/trace/orcl_ora_2187.trc 01-JUN-11 09:16:19 diag/rdbms/orcl/orcl/trace/orcl_mman_2094.trc 01-JUN-11 09:16:16 diag/rdbms/orcl/orcl/trace/orcl_vktm_2082.trc 01-JUN-11 09:16:14 diag/rdbms/orcl/orcl/trace/orcl_ora_2016.trc 30-MAY-11 14:07:02 diag/rdbms/orcl/orcl/trace/orcl_mmon_2093.trc 30-MAY-11 11:15:30 diag/rdbms/orcl/orcl/trace/orcl_ora_3414.trc 30-MAY-11 11:00:01 diag/rdbms/orcl/orcl/trace/orcl_j000_2245.trc 30-MAY-11 10:56:58 diag/rdbms/orcl/orcl/trace/orcl_dbrm_2077.trc 30-MAY-11 10:56:20 diag/rdbms/orcl/orcl/trace/orcl_j002_2201.trc 30-MAY-11 10:56:06 diag/rdbms/orcl/orcl/trace/orcl_ora_2178.trc 30-MAY-11 10:55:58 diag/rdbms/orcl/orcl/trace/orcl_mman_2081.trc 30-MAY-11 10:55:55 diag/rdbms/orcl/orcl/trace/orcl_vktm_2069.trc 30-MAY-11 10:55:53 diag/rdbms/orcl/orcl/trace/orcl_ora_2006.trc
Conclusion: With adrci, we have a new and efficient utility to deal with Oracle Errors – especially for collecting information to send them to Oracle Support. This functionality is called Incident Packaging Service. 11g is generating lots of tracefiles. We can control the purging policy of them with adrci. Finally, we can now tail -f our alert.log from any OS.
No comments:
Post a Comment