Troubleshoot issues in parallel publishing to Brand Portal

Brand Portal is configured with Experience Manager Assets to have approved brand assets seamlessly ingested (or published) from Experience Manager Assets author instance. Once configured, Experience Manager Author uses a replication agent to replicate the selected asset(s) to Brand Portal cloud service for approved usage by Brand Portal users. Multiple replication agents are used Experience Manager 6.2 SP1-CFP5, Experience Manager CFP 6.3.0.2, and onwards to allow high-speed parallel publishing.

On configuring cloud service for Brand Portal under /etc/cloudservice, all necessary users and token are auto-generated and saved in the repository. Cloud service configuration is created, service users required for replication and replication agents to replicate content are also created. It creates four replication agents. So when you publish numerous assets from Experience Manager to Brand Portal, the assets are queued and distributed among the replication agents through Round Robin.

However, publishing can fail intermittently due to- large sling jobs, increased Network and Disk I/O on Experience Manager Author instance, or slowed performance of Experience Manager Author instance. It is, therefore, advised to test the connection with the replication agent(s) before you begin publishing.

Troubleshoot failures in first-time publishing: validating your publish configuration

To validate your publish configurations:

Tail logs while creating Cloud Service

Check tail logs. Check whether the replication agent is created or not. If the replication agent creation fails, edit the cloud service by making minor changes in cloud service. Validate and check again whether the replication agent is created or not. If not, re-edit the service.

If on repeatedly editing the cloud service it is not configured properly, report a daycare ticket.

Test connection with replication agents

View log, if errors are found in replication log:

Clean-up existing Brand Portal publish configurations

Most of the times when publishing is not working, the reason can be that the user who is publishing (for example: mac-<tenantid>-replication doesn’t have the latest private key, and hence publish fails with “401 unauthorized” error and no other error is reported in replication agent logs. You might want to avoid troubleshooting and create a configuration instead. For the new configuration to work properly, clean up the following from Experience Manager author setup:

Now the system is all cleaned up. Now you can attempt to create a cloud service config and still use the existing JWT application. There is no need to create an application, rather update the public key from the newly created cloud config.

Developer connection JWT application tenant visibility issue

If on https://legacy-oauth.cloud.adobe.io/, all the orgs (tenants) for which the current users hold system administrator are listed. If you don’t find the org name here or you can’t create an application for a required tenant here, please check if you have sufficient (system administrator) rights.

There is one known issue on this user interface that for any tenant only top ten applications are visible. When you create the application, stay on that page and bookmark the URL. You don’t need to go to the listing page of the application and find the application that you created. You can hit this bookmarked URL directly and update/delete the application whenever needed.

The JWT application might not be listed appropriately. It is, therefore, advised to note/bookmark the URL while creating JWT application.

Running Configuration stops working

If a replication agent (which was publishing to Brand Portal just fine) stops processing publish jobs, check replication logs. Experience Manager has auto-retry built-in, so if a particular asset publish fails, it is retried automatically. If there is some intermittent issue like network error, it might succeed during retry.

If there are continuous publish failures and queue is blocked, then you should check test connection and try to solve the errors that are being reported.

Based on the errors, you are advised to log a support ticket, so that Brand Portal engineering team can help you resolve issues.

Brand Portal IMS configuration token expired

If your Brand Portal environment stops abruptly, there is a possibility that the IMS configurations are not working properly. The system shows an unhealthy IMS configuration and reflects an error message (similar to the following) that your access token is expired.

com.adobe.granite.auth.oauth.AccessTokenProvider failed to get access token from authorization server status: 400 response: Unknown macro: {"error"}

To resolve this issue, it is recommended to manually save and close the IMS configuration and check health status again. If the configurations does not work, delete the existing configurations and create a new one.

Configure replication agents to avoid connection timeout error

Usually the publishing job fails with a timeout error if there are multiple pending requests in the replication queue. To resolve this issue, ensure that the replication agents are configured to avoid timeout.

To configure the replication agents: