Clarification PullPoint subscription and renewal

[jflevesque-genetec]

Hi everyone,

I feel like we have removed Base Notification eventing from the specification a while ago because it has always been optional on clients to support it and only in Profile S. Since then, the default has always been PullPoint.

We've recently hit an issue with a device where the Subscription termination time is not updated by the PullMessages request and the answer we received was that we had to call the Renew command on the subscription. Looking at the WSDL, the Renew command is not listed and in the Core Spec, section 9.1, there's some mention of Renew for Base Notification, and also one mention in section 9.1, 3rd step saying Additionally it provides the Renew and Unsubscribe operations of the Base Subscription Manager Interface.
But then the Diagram does not show the Renew operation in the lifecycle of the subscription.

9.1.1 says: By default the pull point keep alive is controlled via the PullMessages operation. In this case, after a PullMessages response is returned, the subscription should be active for at least the timeout specified in the PullMessages request.
So if I understand this right, if I create a Subscription with an initial termination time of 2 minutes, then the moment I start calling PullMessages, the TerminationTime of the subscription becomes TimeOfLastPullMessagesResponse+PullMessageTimeoutValue OR TimeOfLastPullMessages+2*PullMessageTimeoutValue. The moment a new PullMessages request comes in, this subscription termination time is automatically updated.

Do I follow the spec properly or am I missing something?

[bsriramprasad]

There is an existing DTT test "5.3.2 REALTIME PULLPOINT SUBSCRIPTION - RENEW" that is sending Renew command and expecting Renew response for a pull point subscription, If PullMessages is the ONLY mechanism to be used to renew the subscription and not via the Renew command this test needs to be removed, need to clone a ticket into DTT accordingly to do tool update and validate no regression in the existing profile tests.

[jflevesque-genetec]

Opened #515 to remove the mention of the Renew command.
@bsriramprasad should I open an issue in the dtt workspace to fix the Renew usage?

[bsriramprasad]

Opened #515 to remove the mention of the Renew command. @bsriramprasad should I open an issue in the dtt workspace to fix the Renew usage?

Yes, that would be good so that we don't miss to track the update.

[jflevesque-genetec]

Created #303 for the DTT

[kieran242]

@jflevesque-genetec I approved your PR and also Testing WG are reviewing and working on your DTT issue.

[kieran242]

@jflevesque-genetec @bsriramprasad @madhurao68 regarding https://github.com/onvif/dtt/issues/303 issue created please note that there has been significant feedback on this issue from GRSE and they would require feedback from the VE WG on this "will be in a conflict with the following profiles requirements: A, C, G, S" and they want wider clarification at F2F in IST.

I have came across two models from different manufacturers where it was necessary to use "remew" to maintain pull point subscription for eventing and both have Profile T conformance so I will be further interested in this conversation at the F2F and get better clarification as JF I was of the same opinion and understanding as yourself hence why I approved your PR.

[HansBusch]

For futher discussion see PR