Solution : https://service.sap.com/sap/support/notes/79341 (SAP Service marketplace login required)
Summary :
This SAP Note addresses issues where Oracle databases report "checkpoint not complete" errors in the alert log, typically indicated by the wait event "log file switch (checkpoint incomplete)". The underlying cause is the DBWR process's inability to write data fast enough due to high database activity, leading to unoverwritable redo logs. Solutions recommended include checking for excessive redo log generation, investigating potential I/O bottlenecks, and configuring database settings optimally. These measures can enhance DBWR's writing capability and avoid database work interruptions during high load scenarios. Strategies include increasing redo log space, adding more log groups, and scaling DBWR processes.
Key words :
dbf rm /oracle/<sid>/mirrloga/log_g11_m2, rm /oracle/<sid>/origloga/log_g11_m1, remove related operating system files, subsequent origlog redo logs reside, alter database add logfile group 11, alter database add logfile group 15, alter database drop logfile group 11, \oracle\<sid>\mirrloga/log_g11_m2, '/oracle/<sid>/mirrloga/log_g11_m2, \oracle\<sid>\origloga/log_g11_m1
Related Notes :
| 1259767 | Management of online redo log files using BRSPACE |
| 1068186 | "checkpoint not complete" message increasingly in alert file |
| 904490 | Automatic enhancement of control files |
| 830576 | Parameter recommendations for Oracle 10g |
| 745639 | FAQ: Oracle enqueues |
| 619188 | FAQ: Oracle wait events |
| 613872 | Oracle traces with ORADEBUG |
| 584548 | Unusually high number of redo logs |
| 521264 | Hang situations |