I want to know where is the memory limitation come from. Recently, I got outofmemory error message in eazyBI log file and some import fails. The max is 2,863,661,056 according to the log however there is not such memory configuration in the instance. Jira Xmx is 16G and eazyBI Xmx is 4096M. I don’t know which memory configuration should be increased in this case. Jira or eazyBI?
Please guide me how to resolve our memory related problem.
Many thanks,
Terry
eazybi-child-node2.log:2021-07-07 17:45:30 +0900 ERROR: [:db2652a5] [Mondrian::OLAP::Error] org.olap4j.OlapException: mondrian gave exception while executing query / OutOfMemory used=2656991088, max=2863661056 for connection: Jdbc=jdbc:mysql://10.6.117.54:3308/eazybi_jira_dwh_131?useUnicode=yes&characterEncoding=UTF-8&sessionVariables=sort_buffer_size=8388608,sql_mode=NO_AUTO_VALUE_ON_ZERO&useSSL=false; JdbcDrivers=com.mysql.jdbc.Driver; UseContentChecksum=true; *** XML schema dump removed ***
eazyBI uses the Mondrian engine for report generation. Xmx (in your case Xmx is 4096M) represents Java heap memory used by Mondrian. However, OutOfMemmory (in your case 2863661056) error shows only one part of the dedicated heap memory, typically around 2/3 of the heap, and therefore shows less than actually used and assigned.
Typically two parameters are quite connected. If you increase the QueryTimeout parameter this will give more room for complex reports to run and eazyBI most probably will reach for more memory within a given timeframe. You can consider increasing the memory if you have enough physical space on your servers for this. Please keep QueryTimeout as minimal as possible, increase it to 2 or 3 min by max.
However, we suggest checking report performance and complexity at the first signs when reports throw Query Timeout reached errors or Out Of Memory errors. Here is a training video on tips how to detect problems in reports. We suggest using Usage Statistics to monitor slow and failing reports.
If the Query Timeout parameter is quite high, work on slow reports and decrease the query timeout gradually. This will gradually minimize memory usage as well.
Import fails should not be directly linked with OutOfMemmory errors. If you have any trouble detecting problems in imports based on the errors messages, please share full eazyBI logs files.
You can send them via email support@eazybi.com or upload them in our contact form.