Kbase 17957: What is the significance of the 'checkpoint length' found in PROMON?
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
10/15/2008 |
|
Status: Verified
GOAL:
What is the significance of the 'checkpoint length' found in PROMON?
FACT(s) (Environment):
UNIX
Windows
Progress 8.x
Progress 9.x
FIX:
Within PROMON
-> R&D
-> 3. Other Displays
-> 4. Checkpoints
The following type of information is found:
Ckpt ---------- Database Writes ----------------
No. Time Len Dirty CPT Q Scan APW Q Flushes
518 17:29:12 0 1564 1539 40 10 0
517 17:16:18 774 1614 1598 567 44 0
516 17:07:18 540 2897 2873 305 0 0
515 17:02:08 310 2444 2419 150 0 0
514 16:57:20 288 2962 2944 53 0 0
513 16:48:55 505 2500 2477 582 0 0
512 16:41:43 432 2231 2210 294 0 0
511 16:34:25 438 2337 2311 304 0 0
The "len" or length column in the example above refers to elapsed seconds. It indicates how much time has passed from when a checkpoint began, until it ended. Notice the zeros in the "flushed" column. This is considered optimal.
The column labeled "dirty" indicates how many dirty database buffers were put on the checkpoint queue at the start of the checkpoint. At the end of the checkpoint these must all have been written to disk. If any of these database buffers were not written to disk at this time, they'll all be written at the end and eventually counted in the "flushed" column.
The snap shot example above presents a system that has been optimally tuned with respect to checkpoints. The checkpoints are about 5 to 10 minutes apart and all the dirty buffers are being written by the page writers (APW's) in the background as they should be.