Monday, June 06, 2016
12:17 PM
This document illustrates how to set up Oracle DB monitoring using HP Oracle DBSPI.
HP Oracle DBSPI 12.x supports Oracle 12.x with latest hotfixes, some of the hotfixes are not available in Support WebSite , but we can get them from HP Support .
- Create DBSPI user account with all the privileges mentioned in dbspiocr.sql script (script available under instrumentation Dir)
- Deploy DBSPI Discovery Policy on Oracle nodes
- Run Database Configuration tool on node , input all required information
- Deploy Oracle DBSPI Policies
DBSPI Set up is :
· Two node RAC cluster
· Two ASM instances running
· ASM user is grid user
· Oracle user is oracleRun dbspicfg.sh and input required details on node A in below format.
AA and BB are the oracle SAID and AA1, BB1 are the two active instances.
Oracle ASM connect should be configured with "/"
LISTENERS should be configured with LISTENERS path.
SYNTAX_VERSION 4
ORACLE
HOME "/u01/app/oracle/product/12.1.0/db"
DATABASE "AA1" CONNECT "hp_dbspi/hp_dbspi@AA"
LOGFILE "mention log file path"
DATABASE "BB1" CONNECT "hp_dbspi/hp_dbspi@BB"
LOGFILE "mention log file path"
ORA_ASM
HOME "/u01/app/12.1.0/grid"
DATABASE "+ASM1" CONNECT "/"
ORACLE
HOME "/u01/app/12.1.0/grid/network/admin"
LISTENER "LISTENER_SCAN2" CONNECT ""
Follow the same procedure on Node B.
Known Errors :
DBSPI40-28: ASM Diskgroup not found for +DATA01 file in v$asm_diskgroup
update defults file on NodeA and NodeB with
[Node A ]# cat /var/opt/OV/dbspi/defaults
UNIX_ORACLE_USER grid
dbspicol ORA_RAC_MASTER +ASM1 ON
[Node B]# cat /var/opt/OV/dbspi/defaults
UNIX_ORACLE_USER grid
dbspicol ORA_RAC_MASTER +ASM2 ON
Fix
The solution to the contention issue is to do one of the following:
1) Ensure Oracle metrics 3,6,16,56,58,203, 206,216 are collected in the same schedule.
OR
2) Space out the scheduled collectors to ensure there is no overlap. For example, the 5 minute collector can run every 5 minutes but the 10 minute collector should start at 02,12,22,32,42,52, minutes past each hour.
04/29/03 09:08:11 ERROR dbspicao(9422) s555 [metric0023:sysstat.pc:333]: DBSPI9-5: Persistent store function 'get' on file '/var/opt/OV/dbspi/history/oracle/companyname-_s555_m0023' failed.
This is a known intermittent problem and the only known workaround is to remove the metric history files.
In this case the metric reporting the problem is 23 (the error message tells: companyname-_s555_m0023). Running the following will fix the problem:
rm /var/opt/OV/dbspi/history/oracle/*0023*
Assiciated KM Documents :
https://softwaresupport.hpe.com/group/softwaresupport/search-result/-/facetsearch/document/KM546343>
No comments:
Post a Comment