20. How to

20.1. Run security checks for dependencies

You can run the crontabber job that checks for security vulnerabilities locally:

make dependencycheck

20.2. Connect to PostgreSQL Database

The local development environment’s PostgreSQL database exposes itself on a non-standard port when run with docker-compose. You can connect to it with the client of your choice using the following connection settings:

  • Username: postgres
  • Password: aPassword
  • Port: 8574

20.3. Reprocess crashes

20.3.1. Reprocessing individual crashes

If you have appropriate permissions, you can reprocess an individual crash by viewing the crash report on the Crash Stats site, clicking on the “Reprocess” tab, and clicking on the “Reprocess this crash” button.

20.3.2. Reprocessing lots of crashes if you’re not an admin

If you need to reprocess a lot of crashes, please write up a bug. In the bug description, include a Super Search url with the crashes you want reprocessed.

20.3.3. Reprocessing crashes if you’re an admin

If you’re an admin, you can create an API token with the “Reprocess Crashes” permission. You can use this token in conjunction with the scripts/reprocess.py script to set crashes up for reprocessing.

For example, this reprocesses a single crash:

$ docker-compose run processor bash
app@processor:app$ ./scripts/reprocess.py c2815fd1-e87b-45e9-9630-765060180110

This reprocesses crashes all crashes with a specified signature:

$ docker-compose run processor bash
app@processor:app$ ./scripts/fetch_crashids.py --signature="some | signature" | ./scripts/reprocess.py


If you’re reprocessing more than 10,000 crashes, make sure to add a sleep argument of 10 seconds (--sleep 10). This will slow down adding items to the reprocessing queue such that the rate of crashes being added is roughly the rate of crashes being processed. Otherwise, you’ll exceed our alert triggers for queue sizes and it’ll page people.


August 17th, 2017: Everything below this point is outdated.

20.4. Populate PostgreSQL Database

Load Socorro schema plus test products:

socorro setupdb --database_name=breakpad --createdb

20.5. Create partitioned tables

Normally this is handled automatically by the cronjob scheduler Service: Crontabber but can be run as a one-off:

python socorro/cron/crontabber_app.py --job=weekly-reports-partitions --force

20.6. Populate Elasticsearch database


See the chapter about Crash storage: Elasticsearch for more information.

Once you have populated your PostgreSQL database with “fake data”, you can migrate that data into Elasticsearch:

python socorro/external/postgresql/crash_migration_app.py

20.7. Sync Django database

Django needs to write its ORM tables:

export SECRET_KEY="..."
cd webapp-django
./manage.py migrate auth
./manage.py migrate