Script-based publishing process

The publish flow, steps and logs described here apply to publishing to CHM, WebHelp, PDF or Collaborative Review.

After the user pushes the Publish button, in Publication Manager or in the Web Client, the process is automated.

The Phase I of the publish process is EXPORTFORPUBLICATION and is configured in XML Background Task Settings. The entries in EXPORTFORPUBLICATION are used to validate links, conrefs, keys, and so on. If the validation fails, the publication status is set to Draft until all references are resolved.

After the publication is published, you can download it to your system, upload it to the repository or have it distributed on a delivery platform.

When a publishing job is created, it is put on a centralized queue in the database. The message is picked up by a background task service. Depending on your configuration, this can be on the Content Manager main server or on a batch server.

The number of simultaneous publish actions can be configured via XML Background Task Settings. You can adapt the number of simultaneous publish actions per service or add extra services. However be aware that the more jobs you run simultaneously, the more resources it takes on the server (content is retrieved from the database which means more load on the database server and network, content is dumped onto the file system which means increased I/O, multiple java processes are running which means higher memory consumption) and all these jobs compete with each other for resources:
  • Content is retrieved from the database, it creates more load on the database server and on the network.
  • Content is dumped onto the file system, it increases the I/O activity.
  • Multiple java processes are running, it creates a higher memory consumption.

If you double the number of simultaneous processes, it does not necessarily mean that your overall publishing speed doubles.