bad event partition in SRM

Document ID : KB000122262
Last Modified Date : 04/12/2018
Show Technical Document Details
Issue:
Some of our clients are having issues running reports consistently. We recently ran a mysql check against our SRM database in PROD and seeing a bad partition. 

/ca/spectrum/mysql/bin 
./mysqlcheck --defaults-file=../my-spectrum.cnf -uroot -proot reporting 

reporting.event 
error : Partition EVENT_P98 returned error 
error : Unknown - internal error 147 during operation

Can this be resolved without initializing or rebuilding the SRM database?
Environment:
Linux OS
Spectrum 10.2.3
Cause:
It may be possible that low memory settings in the MySQL config file my-spectrum.cnf is preventing the mysqlcheck from reading larger table partitions. The table partition may not actually be corrupt. 
Resolution:
Try increasing innodb_buffer_pool_size memory setting and run mysqlcheck again: 

1. edit <SPECROOT>/mysql/my-spectrum.cnf file

Locate memory settings: 

innodb_file_per_table
innodb_buffer_pool_size = 256M
innodb_log_file_size = 256M
innodb_log_buffer_size = 4M

You can increase innodb_buffer_pool_size to 50% of available RAM. For example if there is 16GB of RAM available, the innodb_buffer_pool_size can be set to 8096M. For example: 

innodb_buffer_pool_size = 8096M

2. once the file has been edited, you must restart MySQL for the change to take effect: 

nav to <SPECROOT>/bin

to stop MySQL: 

./cmdC localhost 2 MYSQL MYSQL.OUT

to start MySQL back up: 

./launchinstdbapp localhost MYSQL y MYSQL.OUT