TABLE OF CONTENTS

• Q: How do I provide my log files?
  • Nextflow Logs
  • Tower Logs
    • For Docker Compose deployments
    • For Kubernetes-based deployments
  • Troubleshooting
    • Work Directory Retrieval
    • Configure key-pair to AWS Batch Compute via Tower Forge

Wherever possible, the Seqera Labs Support Team requests that you share log files as part of your ticket submission by adding them as an attachment to your ticket submission.  The different types of logs that we may ask you to provide include the following:

1. Nextflow logs
2. Tower logs
   1. Backend logs
   2. Cron logs
   3. Frontend logs

Q: How do I provide my log files?

Nextflow Logs

Nextflow logs can be retrieved on your Tower instance through the following steps:

1. Log into your Tower instance and click 'Runs.'
2. Select the workflow that you want to share.
3. Move to the 'Execution log' tab.
4. Click the 'Download log files' dropdown -> Download the 'Nextflow log file'.
   1. You can also provide the 'Nextflow console output' for the full details of your standard output.

Tower Logs

Tower logs are found inside your deployment instance.

For Docker Compose deployments

1. Log in to your instance where the containers are running.
2. Run the following command to determine the target ID of the container:
   1. docker ps
   2. Get the container ID or the name (ie. ba54613516f or tower_backend_1)
      
3. Run the following command to get the logs of a specific container and dump it into a file.
   1. This command will send the full logs from the beginning:
      1. docker logs tower_backend_1 > /tmp/backend.log
   2. This command will send the logs of a specific date:
      1. docker logs tower_backend_1| grep Sep-07 > /tmp/backend.log
4. Attach the 'backend.log' file created to your ticket.
5. The process above also applies to frontend, backend, and cron logs; note that for each, you need to change the filename (i.e. backend.log), so it does not overwrite the existing file.

For Kubernetes-based deployments

1. Log in to your instance where the Kubernetes pods are running.
2. Run the following command to determine the target name of the pods for the log extraction:
   1. kubectl get pod -n [namespace]
   2. kubectl logs [pod name] -n [namespace] > /tmp/backend-pod.og
3. We may also need you to provide the pod deployment in some situations.  To do so run the command below and provide us the result.
   1. kubectl describe [pod name] -n [namespace] > /tmp/describe-backend-pod.log
4. Be mindful of the log file name as not to overwrite existing files. The process above is applicable for other pods depending on what is required for investigation.

Troubleshooting

Work Directory Retrieval

Occasionally, you may encounter an error downloading the Nextflow logs via Tower. These logs can also be retrieved from the work directory set on the Compute Environment. 

Azure:

AWS Batch

The example above applies to all compute environments.

Configure key-pair to AWS Batch Compute via Tower Forge

For Tower version 22.3 and above, we have added the implementation of key-pair to the instances created by Tower Forge for AWS Batch Compute Environment. This will allow the user to access the instance directly and retrieve the logs inside.