License Server cookies problems

[tema_blr]

We caught strange behavior in our own deployed license server.
From the documentation:

For the licensing mechanism to run properly, your browser needs to accept cookies. Otherwise, Oxygen XML Web Author will not be loaded.

I researched oXygen Web Author communication with License Server and found that it automatically generates "_uid" cookie which will be used by the license server for managing client(browser) licenses.
But we have non-standard behavior in different browsers(always with allowed cookies), we have a new "_uid" in each request to the license server.
So, after a few minutes of 1 user work - there are 10 used licenses and about 100 rejected requests...
In the documentation https://www.oxygenxml.com/doc/versions/ ... nsing.html says:

While no personal information is sent to the server, a cookie that identifies the user is auto-generated.

What does it mean:

personal information

? Maybe we can avoid using cookies or there is another solution to this problem?

[cristi_talau]

Hello,

This behavior might happen if you customized the application HTML to make additional requests to the server. We can give you more details if you send us a HAR log [1] with the requests made by the browser on our support email address: support@oxygenxml.com .

Regarding the personal information mention in our documentation, it means that Web Author does send any personal information about the user to the license server, only an auto-generated ID that is solely used for enforcing the licensing.

Best,
Cristian

[1] https://support.zendesk.com/hc/en-us/ar ... leshooting

[tema_blr]

Hi there!
This problem is not related to customizing the application HTML.
We found origins of this bug - our protection system blocks some of License Server requests and for each request(from Web Author to the LS) new license will be created. License Server checks out all the licenses to one user if that user is running TrendOffice MicroScan.
Could you help with solving this problem?

[cristi_talau]

Hello,

Web Author needs all the requests to arrive to their destination to function properly. From what I read, TrendMicro OfficeScan uses machine learning to identify requests that they consider problematic. So, the best way to proceed would be to disable TrendMicro for Web Author or report the problem to TrendMicro to adjust their machine learning algoritm for these cases.

What we can try to do on our end is to identify which requests are blocked and maybe make them less susceptible to machine learning misclasification.

In a Web Author deployment, there are only two kinds of requests:
- From the browser to Web Author.
- From Web Author to the license server.

I would expect that the blocked requests are in the first category. In this case, it would help to have a "HAR" log [1] of the requests to see which requests get blocked.

Best,
Cristian

[1] https://support.zendesk.com/hc/en-us/ar ... leshooting