_MERGED.xml file is a bottleneck that prevents large data sets from publishing

[jeff_hooker]

We currently use Oxygen Webhelp as our primary online help format published out of our Tridion Docs CMS. While we are reasonably happy with it for most purposes, it has a bottleneck due to its reliance on the creation of a _MERGED.xml file.

While most of our documents are fewer than 4000 pages, the fact that some of our products involve extremely large register sets means that a certain percentage of our documents can be well over 100,000 topics, and even more if we need to support multiple versions of a product in a single documentation set.

The size of the _MERGED.xml file and the cost of keeping it stored in the Java memory space consistently grows past the point where our AWS publishing servers can sustain it, so publishing fails with a Java heap error.

When this happens we publish out to Eclipse help instead, but that comes at a cost of not being able to offer a single platform to our customers.

So, with the whining out of the way, my question is this; is there any means of optimizing the publishing of Oxygen Webhelp to enable unusually large data sets? If the cost of this is slower processing or reduced feature sets that's fine; we're just trying to get our content to publish.

[julien_lacour]

Hello Jeff,

The _MERGED.xml file is only generated in PDF scenarios (XSL-FO and HTML5 & CSS) not in WebHelp Responsive.
Are you sure you are getting this file from the WebHelp scenario?

Regards,
Julien

[jeff_hooker]

It is definitely a responsive web help publish job, and it is definitely currently generating (and using; I can see it in the ANT files) so what I can do to remove this dependency? Is this configurable?

[julien_lacour]

Hello Jeff,

Can you indicate which version of Oxygen WebHelp is used in your Tridion Docs CMS?
Also, do you have the possibility to share the console log for this output?

Regards,
Julien