Troubleshooting¶
During the deployment process you may encounter problems. This is a list of known problems and how to address them.
If a problem is not mentioned explicitly it should be raised to team which is most likely responsible:
- All deployment steps prefixed with Post-Rollout: Application developers
- All other errors: Platform administrators
Platform-related errors¶
A Lagoon deployment can end up with a Failed status. If this is the case then the following errors may show up in the log output shown for the deployment in the Lagoon UI.
TLS handshake timeout¶
If the log output of a deployment contains the following during various stages of the deployment then redeploy the environment from the Lagoon UI:
Unable to connect to the server: net/http: TLS handshake timeout
This may be caused by a spike in load during deployment.
Context deadline exceeded¶
If the log output of a deployment contains the following then redeploy the environment from the Lagoon UI:
failed to call webhook: Post "[Ingress url]": context deadline exceeded
This may be caused by a spike in load during deployment.
Harbor authentication failed¶
If log output for the deployment contains the following during the Image Push to Registry stage of the deployment then redeploy the environment from the Lagoon UI:
level=fatal msg="copying system image from manifest list: trying to reuse blob
sha256:[SHA] at destination: checking whether a blob sha256:[SHA] exists in
[Harbor url]: authentication required"
This is caused by a bug in Harbor. It should be addressed in a future update of this part of the platform.
MySQL server has gone away¶
If log output for the deployment contains the following during the Post-Rollout drush deploy stage of the deployment then redeploy the environment from the Lagoon UI:
Drupal\Core\Database\DatabaseExceptionWrapper: SQLSTATE[HY000]: General error:
2006 MySQL server has gone away
This may be caused by intermittent database problems caused by spike in load during deployments.
Service deployment exceeded its progress deadline¶
If log output for the deployment contains the following during the Applying Deployments stage of the deployment then redeploy the environment from the Lagoon UI:
Waiting for deployment "[Service]" rollout to finish: 1 old replicas are pending
termination...
error: deployment "varnish" exceeded its progress deadline`
This may be caused by a spike in load during deployment.
Drush bootstrap failed¶
If log output for the deployment contains the following during the Post-Rollout drush deploy stage of the deployment then redeploy the environment from the Lagoon UI:
The command "/app/vendor/bin/drush updatedb --uri=https://[Domain]" failed.`
Exit Code: 1(General error)
[...]
Bootstrap failed. [...]
This may be caused by intermittent database problems caused by spike in load during deployments.