analyze_replay_performance.py
After a database initialization
completes, run the analyze_replay_performance.py script on the processing
server that serviced the replay request to analyze transaction replay throughput statistics,
including the checkpoint creation time, any throttling, the number of transactions replayed,
and the duration to replay them. Look for database initialization requester
or start sending in-memory database replay to peer cluster log messages to
find the processing server on which to run the script. The script uses the
mtx_debug.log file to analyze the replay performance because the
server where it is run drove the replay, both locally and remotely.
Syntax
[-h] [-d] [-c] [-s] [-v]Options
The analyze_replay_performance.py script has the following command line options.
- -h, --help
- Help about the script options.
- -d, --debug
- Increments the debug flag count. The higher the debug flag count, the more debugging output is produced.
- -c checkpoint, --checkpoint checkpoint
- The name of the checkpoint file in the mtx_debug.log file to analyze. If you do not specify this option, the script uses the mtx_debug.log file.
- -s remote_server_start_time, --start-time remote_server_start_time
- Specifies the Transaction Server start time. For local replay, this time is detected by the system. For
remote replay, the time must be specified in the form
YYYY-MM-DD hh:mm:ss.msec - -v, --verbose
- Increases the verbosity of the script. Doing so causes more messages and statistics to be printed. By default, the verbosity is set to 0.
Local Transaction Replay Non-Verbose Mode
Note: Total replay times and transactions per second
(TPS) for individual databases are estimated, due to multi-threading. Overall replay
time and TPS are exact.
analyze_replay_performance.py mtx_debug.log
Replay target: 2:1:1:1:0:1
Replay checkpoint name: in-memory database
Service start time: 2014-07-23 21:50:25.404337
Replay request time: 2014-07-23 21:50:25.404337
Replay start: 2014-07-23 21:50:26.033924
Replay completion: 2014-07-23 21:50:40.304745
Replay duration (checkpoint): 0:00:14.270820
Replay duration (total): 0:00:14.270820
Total startup time: 0:00:14.900408
Replay threads: 1
Local replay throttle: 128 Messages Outstanding
Remote replay throttle: 1 Batches Outstanding (Max 128 Messages per Batch)
DATABASE File Count Txn Count Replay Time Txn/Sec
Overall 5 96841 0:00:14.270820 6785.94
subscriber_db 1 59854 0:00:09.155952* 6537.17*
balance_set_db 1 15053 0:00:02.304713* 6531.40*
activity_db 1 20000 0:00:03.069597* 6515.51*
sched_db 0 0 N/A * N/A*
event_db 1 1 0:00:00.062479* 16.01*
pricing_db 1 1933 0:00:00.798911* 2419.54*
transaction 0 0 N/A * N/A*Local Transaction Replay Verbose Mode
Note: Total replay times and TPS for individual
databases are estimated due to multi-threading. Overall replay time and TPS are
exact. The individual file replay time and TPS are exact.
analyze_replay_performance.py -v mtx_debug.log
Replay target: 1:1:1:1:0:1
Replay checkpoint name: mtx_ckpt_v4500.1.1393334498
Replay request time: 2014-02-25 05:23:44.110432
Replay start: 2014-02-25 05:23:44.112170
Replay completion: 2014-02-25 05:23:45.873697
Replay duration (checkpoint): 0.839398
Replay duration (total): 1.761527
Replay threads: 3
Local replay throttle: 128 Messages Outstanding
Remote replay throttle: 1 Batches Outstanding (Max 128 Messages per Batch)
DATABASE File Count Txn Count Replay Time Txn/Sec
Overall 19 3245 1.7615 1842.15
subscriber_db 4 400 0.039158* 10215.00*
subscriber_db_1.log. 1 136 0.030165 4508.54
subscriber_db_2.log.gz 1 158 0.033270 4749.01
subscriber_db_3.log.gz 1 1 0.004594 217.68
subscriber_db_4.log.gz 1 105 0.013228 7937.74
balance_set_db 4 100 0.062245* 1606.56*
balance_set_db_5.log.gz 1 44 0.008844 4975.18
balance_set_db_6.log.gz 1 11 0.009619 1143.57
balance_set_db_7.log.gz 1 35 0.014518 2410.80
balance_set_db_8.log.gz 1 10 0.002397 4171.78
activity_db 4 200 0.061704* 3241.27*
activity_db_9.log.gz 1 83 0.022846 3633.02
activity_db_10.log.gz 1 69 0.015164 4550.21
activity_db_11.log.gz 1 38 0.012051 3153.24
activity_db_12.log.gz 1 10 0.016506 605.84
sched_db 1 0 0.000139* 0.00*
sched_db_13.log.gz 1 0 0.000139 0.00
event_db 4 400 0.078254* 5111.56*
event_db_14.log.gz 1 187 0.045881 4075.76
event_db_15.log.gz 1 51 0.012566 4058.62
event_db_16.log.gz 1 117 0.020534 5697.86
event_db_17.log.gz 1 45 0.010229 4399.31
pricing_db 1 1845 0.245975* 7500.76*
pricing_db.log.gz 1 1845 0.245975 7500.76
transaction 1 300 0.021305* 14081.15*
transaction_1_1_2_334504_5.log.gz 1 300 0.021305 14081.15
Remote Transaction Replay
analyze_replay_performance.py mtx_debug.log
**** NOTE: Cannot determine server start time for remote server.
Assuming replay request start time.
Use the --start-time option to specify the server start time.
Source code version: 5220 revision 80766
Replay target: 2:1:1:1:0:1
Replay checkpoint name: in-memory database
Service start time: 2021-02-27 15:38:29.974376
Replay request time: 2021-02-27 15:38:29.974376
Replay start: 2021-02-27 15:38:32.994842
Replay completion: 2021-02-27 15:38:48.474500
Replay duration (checkpoint): 0:00:15.479657
Replay duration (total): 0:00:15.479657
Total startup time: 0:00:18.500123
Replay threads: 1
Local replay throttle: 128 Messages Outstanding
Remote replay throttle: 1 Batches Outstanding (Max 128 Messages per Batch)
DATABASE File Count Txn Count Replay Time Txn/Sec
Overall 6 37153 0:00:15.479657 2400.12
subscriber_db 1 12000 0:00:08.985681* 1335.46*
balance_set_db 1 2000 0:00:04.417732* 452.72*
activity_db 1 4020 0:00:05.653089* 711.12*
sched_db 1 35 0:00:00.062991* 555.63*
event_db 1 8522 0:00:04.701598* 1812.57*
pricing_db 1 10576 0:00:14.457673* 731.51*
alert_db 0 0 N/A * N/A*
transaction 0 0 N/A * N/A*
* Note: Aggregate replay times and TPS for individual
databases is approximate due to multi-threading.
Overall replay time and TPS are accurate.