Hi Michal,
i have a question, if a establish-sub is repeatedly used from same session , we are accepting the subscription and generating a new subscription in effect, its a duplicate subscription, at server should we not have a check for this or is it entirely on the client to manage its subscriptions?
in one of negative testing scenarios, an app repeatedly kept subscribing for same notification , which eventually hit fd limit and lead to netopeer2-server crash,
so just wanted to understand if we should check at server.
sample:
> establish-sub --stream NETCONF --filter-xpath /_ran-nr-nrm-cmAgent-oru-rpc:notification-result-execute-ru-rpc
DATA
<id xmlns="urn:ietf:params:xml:ns:yang:ietf-subscribed-notifications">16</id>
> establish-sub --stream NETCONF --filter-xpath /_ran-nr-nrm-cmAgent-oru-rpc:notification-result-execute-ru-rpc
DATA
<id xmlns="urn:ietf:params:xml:ns:yang:ietf-subscribed-notifications">17</id>
below is notification
notification notification-result-execute-ru-rpc {
description
"Relays the RU's rpc-reply for one dispatched rpc-request.
Correlated to the original request by ru-id + txn-id.
xml carries the raw rpc-reply content from the RU.";
leaf ru-id {
type string;
mandatory true;
description
"ru-id of the O-RU that replied.";
}
leaf txn-id {
type string;
mandatory true;
description
"Echoed from the originating rpc-request.";
}
anydata xml {
description
"Raw rpc-reply body from the RU, e.g.
<status>SUCCEEDED</status>
<wait-time>5</wait-time>
<measurement-id>1</measurement-id>";
}
}
in one negative case testing, there were about 300 fds consumed for same notification
113 /dev/shm/srsub__ran-nr-nrm-cmAgent-oru-rpc.notif {
Hi Michal,
i have a question, if a establish-sub is repeatedly used from same session , we are accepting the subscription and generating a new subscription in effect, its a duplicate subscription, at server should we not have a check for this or is it entirely on the client to manage its subscriptions?
in one of negative testing scenarios, an app repeatedly kept subscribing for same notification , which eventually hit fd limit and lead to netopeer2-server crash,
so just wanted to understand if we should check at server.
sample:
below is notification
in one negative case testing, there were about 300 fds consumed for same notification
113 /dev/shm/srsub__ran-nr-nrm-cmAgent-oru-rpc.notif {