We are starting to ramp-up testing of our application, and we started getting these messages in our log file:
LOG: checkpoints are occurring too frequently (15 seconds apart)
HINT: Consider increasing the configuration parameter "checkpoint_segments".
LOG: recycled transaction log file "0000000000000045"
LOG: recycled transaction log file "0000000000000046"
LOG: recycled transaction log file "0000000000000044"
So I gather that we should bump up the number for checkpoint_segments, but I'm wondering:
1) what is reasonable rule-of-thumb for what this value should be? (we are at the default right now)
2) should checkpoint_timeout be changed at the same time?
3) should I increase wal_buffers in conjunction with checkpoint_segments?
Thanks!
- DAP
----------------------------------------------------------------------------------
David Parker Tazz Networks (401) 709-5130