Common Issues

Last Updated: 2020-10-12

My Dates and Times are off by a few Hours

Ensure that your Grouparoo Application Servers and Grouparoo Database servers/instances are all in the same timezone. Furthermore, some Sources (particularly database sources like MySQL and Postgres) should also be in the same timezone.

Depending on your deployment environment, there are a number of ways you can set the system timezone. Common ways to do this include:

  • Setting the system-wide environment variable process.env.TZ which the Node.js Grouparoo process will read. Depending on your Operating System, this can be set a number of ways.
  • Depending on your Source database type, you may also be able to set the Timezone of the Database independently of the the host OS's timezone:

Slow Migrations

The Grouparoo application will automatically apply migrations to the Grouparoo database (defined by DATABASE_URL) when you start the application. Sometimes however, with large data volumes, these migrations may take a long time to complete. Depending on you deploy Grouparoo, this may cause problems.

If you need to run Grouparoo's migrations in a one-off process, you can with the following command from the root of your Grouparoo application:

./node_modules/@grouparoo/core/bin/migrator up

My Runs do not Complete

In general, Runs will wait for all of their Imports and Exports to complete before for finishing themselves. Encountering an error is OK when trying to import or export a Record, but not trying to do the import or export will cause a problem. If a Run should have imported 100 Records, but the system only processed 99, the Run will still be in the running state. What happened to that last Record Import?

  • Check for resque (background task) errors on your server by visiting /resque. Are there any errors? Does it look like a worker has been working on a Record for a long time (ie: over a minute)?

Grouparoo will try to retry "stuck" worker-type errors, but if your Grouparoo deployment is unstable, these types of errors will happen often. Here are some common causes:

  • Your grouparoo servers are getting rebooted often. Perhaps you have deployed Grouparoo on Kubernetes, and you are regularly re-balancing your cluster.
  • You are not giving the Grouparoo application enough time to shut down properly. When you signal a Grouparoo server to shutdown, it may be in the middle of processing a background task. Rather than immediately terminating the process, you should allow the application some time to gracefully exit
    • In practice, this looks like:
      • SIGTERM
      • (wait 30 seconds)
      • SIGKILL
    • Depending on your deployment orchestrator, there are different options to set. For example, in Kubernetes, the terminationGracePeriodSeconds directive should be used
    • Heroku will always give each Dyno 30 seconds to stop gracefully before terminating it.
    • Check that the Grouparoo process is receiving the shutdown signals. Do not use another runner (like npm start, pm2, or nodemon) to then run Grouparoo. They often will intercept system signals like SIGTERM and not pass them down to the child process.