Resolve Analyzer Known Issues
Problem 1. Auto-Analyser doesn't work. Analyzer health check status failed: OpenSearch is not healthy
Problem Description
Analyzer log:
2021-09-09 11:34:47,927 - analyzerApp - ERROR - Analyzer health check status failed: OpenSearch is not healthy;
[pid: 10|app: 0|req: 1/3] 127.0.0.1 () {28 vars in 294 bytes} [Thu Sep 9 11:34:46 2021] GET / => generated 43 bytes in 1643 msecs (HTTP/1.1 503) 3 headers in 120 bytes (1 switches on core 0)
2021-09-09 11:35:48,737 - analyzerApp.utils - ERROR - Error with loading url: http://opensearch:9200/_cluster/health
2021-09-09 11:35:48,752 - analyzerApp.utils - ERROR - HTTPConnectionPool(host='opensearch', port=9200): Max retries exceeded with url: /_cluster/health (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7f5cb82d4290>: Failed to establish a new connection: [Errno 111] Connection refused'))
2021-09-09 11:35:48,753 - analyzerApp.esclient - ERROR - OpenSearch is not healthy
2021-09-09 11:35:48,753 - analyzerApp.esclient - ERROR - list indices must be integers or slices, not str
OpenSearch container restarting all the time:
STATUS NAMES
Up Less than a second (health: starting) reportportal_opensearch_1
Solution
Create a directory for OpenSearch and assign permissions with the following commands
mkdir -p data/opensearch
chmod 777 data/opensearch
chgrp 1000 data/opensearch
Recreate ReportPortal services.
Problem 2. Auto-Analyser doesn't work. KeyError: 'found_test_and_methods' not found
Problem Description
2021-09-09 11:35:48,737 - analyzerApp.utils - ERROR - KeyError: 'found_test_and_methods' not found
Solution
Regenerate index in the OpenSearch. Project settings -> Auto-Analysis -> Genetate Index
Problem 3. Amqp connection was not established
Problem Description
2021-09-09 11:32:00,579 - analyzerApp - INFO - Starting waiting for AMQP connection
2021-09-09 11:32:00,586 - analyzerApp.amqp - INFO - Try connect to amqp://rabbitmq:5672/analyzer?heartbeat=600
2021-09-09 11:32:00,595 - analyzerApp - ERROR - Amqp connection was not established
Solution
RabbitMQ container is not running. Wait for status running
or recreate the RabbitMQ container.
Problem 4. Performance
Problem Description
Slowing down analysis or waiting for a long time fore responce.
Analyzer logs:
DAMN ! worker 1 (pid: 9191) died, killed by signal 9 :( trying respawn ...
Respawned uWSGI worker 1 (new pid: 9490)
Solution
Increase VM stats. We recommend using the minimum memory:
- Analyzer - 1024 Mb
- Analyzer train - 512 Mb
Also you can reduce the number of Analyzer processes with processing environment variable UWSGI_WORKERS: 2
(default 4
), then:
- Analyzer - 768 Mb
However, UWSGI_WORKERS
will slow down the Analyzer.