- Server process takes too long to find a free buffer (too long is determined by internal parameters).
- checkpoint queue (list of ditry buffers waiting to be written to the disk) becomes too long.
- Three second time out. (Every three second DBW will writer some dirty buffers to disk).
- alter system checkpoint is called. all dirty buffers are written to the disk.
checkpoints automatically occur when
- tablespace is dropped
- tablespace is made read only
- tablespace is put into hot backup mode
- tablespace is taken offline
LGWR writes when....
- User commits a transaction
- log buffer is one third full
- Just before DBWn writes
- Three second time out