Skip to content

Latest commit

 

History

History
35 lines (30 loc) · 1.65 KB

README.md

File metadata and controls

35 lines (30 loc) · 1.65 KB

QuestDB (WAL) Table Resumer

Overview

Command line utility written in python that is meant to:

  1. Connect to questdb instance
  2. Search for suspended WAL tables
  3. Search for WAL/segment ids of suspended tables from provided log file
  4. Search for corrupted transaction for each suspended tables
  5. Resume WAL by skipping corrupted transation

Essentially, it does what's described in Diagnosing corrupted WAL transactions but in automated way.

Running

  1. Install dependencies from requirements.txt, otherwise Pipfile can be utilised along with pipenv
  2. Fetch logs from your questdb instance, for questdb running in docker container: docker logs questdb > logs/qdb.log
  3. Make sure your questdb instance has PGWire enabled and that it accepts connections on 8812 port (or other if configured differently)
  4. Execute with --dry-run flag: 
    python main.py \
    --host <host> \
    --port <port> \
    --username <username> \
    --password <password> \
    --database questdb \
    --log-file logs/qdb.log \
    --dry-run
  5. Examine if there are no errors when connecting to questdb or parsin log file.
  6. Run again without --dry-run flag
  7. Enjoy resumed WAL table :)

Disclaimer

This utility was written to handle suspended WAL tables that have been suspended because of power failures on device hosting questdb. As pointed out in #4829 issue, skipping corrupted WAL transactions might not always work in case of serious storage corruption.