-
Type:
Bug
-
Status: Closed
-
Priority:
Major
-
Resolution: Cannot Reproduce
-
Affects Version/s: 7.0.0
-
Fix Version/s: None
-
Labels:
-
Environment:IDM master runs on CentOS7, MySQL as repo. Java 11, Latest DJ master.
-
Target Version/s:
-
Story Points:5
Noticed about 10% throughput drop with recon create from IDM to DJ. With commit 55394c2, the throughput is about 520 users/second, with commit 0a27773(from OPENIDM-13657), the throughput is about 459 users/second.
Attached are the metrics from tests using the two commits.
To reproduce using Pyforge:
1. Set Pyforge config as below:
[Stress]
num_users = 50
duration = 2000
concurrency = 10
max_throughput = -1
[OpenIDM]
host_name = ${Default:host_name}
java_home = /usr/java/jdk11
java_args = ${Default:java_args}
version = 7.0.0-SNAPSHOT
previous_version = 6.5.0
protocol = http
repo_type = mysql
use_docker_repo = False
enable_metrics = True
jvm_debug_options = ${Default:jvm_debug_options}
cloud = ${Default:cloud}
enable_policy_enforcement = False
2. Run test:
run-pybot.py -c stress -s idm.regular_recon.ReconManUserToLDAP OpenIDM
3. Check pyforge/results/debug.txt for recon throughput.
Note that no throughput drop on recon update was seen.