Table of Contents outline true
...
- Export deleted MARC authority records tests finished successfully on qcon environment using 100K, 300K and 646K records.
- Comparing Test Duration and Response time between tests with 100K, 300K and 646K records
- Test Duration grew up according to grew records count.
- Response time for 100k and 300k records almost the same, but it grew up almost in two times compared 100k and 646k records.
- During testing, we noticed that mod-entities-links module used 53% CPU in spikes for 300k, but for 646k records used 39% CPU.
- No memory leaks are observed.
- After the first part of the tests, the database was cleaned up by removing deleted MARC authority records. This cleanup negatively affected the duration and response time for the 100k records in the repeated test.
Test Runs
| Test № | Scenario | Test Conditions | Results |
|---|---|---|---|
| 1 | Export deleted MARC authority records via API | 100k deleted records | Completed |
| 2 | 300k deleted records | Completed | |
| 3 | 646k deleted records | Completed |
...
Export deleted MARC authority records were started from CarrierIO (QCON_ECS_Export_Deleted_Records) on a qcon Quesnelia [ECS] environment..
...
- Test 1: Tested export 100k deleted MARC authority records, started on Central Office tenant(cs00000int) with one user.
- Test 2: Tested export 300k deleted MARC authority records, started on Central Office tenant(cs00000int) with one user.
- Test 3: Tested export 646k deleted MARC authority records, started on Central Office tenant(cs00000int) with one user.
- After tests all deleted MARC authority records were removed from Database and added again, then repeat tests with 100k and 300k MARC authority records on Central Office tenant(cs00000int) with one user.
At the time of the test run, Grafana was not available. As a result, response times for Export deleted MARC authority records were taken from a .jtl files. These results were visualized in JMeter using a Listener (Response Times Over Time).