Error after creating Notifications

YANG
1

I have successfully implemented notifications and subscribed to these notifications and everything works fine i.e. whenever a particular variable changes in the database all the subscribers get a notification. I have tested tested this feature using a third party Netconf Browser (MGsoft). I followed the example to implement the notifications.

After this when i do a “get” from the netconf CLI, even though i get everything there are errors in the log.

    root@orion:/# /opt/confd/bin/netconf-console --get
. /*got all data successfully*/
.
    root@orion:/# tail /var/log/messages     
    Jan  1 00:34:44 orion daemon.err confd[550]: devel-c no registration found for callpoint overheat/get_log_times of type=external
    Jan  1 00:34:44 orion daemon.crit confd[550]: - no registration found for callpoint overheat/get_log_times of type=external
    Jan  1 00:34:44 orion daemon.err confd[550]: devel-c no registration found for callpoint overheat/get_log_times of type=external
    Jan  1 00:34:44 orion daemon.crit confd[550]: - no registration found for callpoint overheat/get_log_times of type=external
    Jan  1 00:34:44 orion daemon.err confd[550]: devel-c no registration found for callpoint overheat/get_log_times of type=external
    Jan  1 00:34:44 orion daemon.crit confd[550]: - no registration found for callpoint overheat/get_log_times of type=external
    Jan  1 00:34:44 orion daemon.err confd[550]: devel-c no registration found for callpoint overheat/get_log_times of type=external
    Jan  1 00:34:46 orion daemon.info confd[550]: netconf id=15 close-session attrs: message-id="0"
    Jan  1 00:34:46 orion daemon.info confd[550]: netconf id=15 sending rpc-reply, attrs: message-id="0"
    Jan  1 00:34:46 orion daemon.info confd[550]: audit user: admin/0 Logged out ssh <local> user
    root@orion:/#

overheat in the name of the notification stream i created. i did not add anything called “get_log_times”.
I do not see this error in the log when i add a xpath filter and then get each module individually.
for ex:
root@orion:~# /opt/confd/bin/netconf-console --get -x /system
root@orion:~# /opt/confd/bin/netconf-console --get -x /bridge

The error log is only seen when i do a get of the whole data model at once, and the behaviour is same even when i do the get operation from the netconf browser.

I would like to get some insight as to why this is happening.

2

What does your YANG model look like under /overheat?
If you do a

$ confd --check-callbacks

What does that say about your callpoint under /overheat ?

3

Overheat is not part of the yang model its only a name of a stream.

memset(&ncb, 0, sizeof(ncb));
ncb.fd = notifsock;
strcpy(ncb.streamname, "overheat");
ncb.cb_opaque = NULL;
   
if (confd_register_notification_stream(dctx, &ncb, &live_ctx) != CONFD_OK)
{
    confd_fatal("Couldn't register stream %s\n", ncb.streamname);
}

This notification is tied to the overheat stream and find below the elements related to the notification in the YANG model.

   notification temp-alarm {
    		leaf temperature-alarm {
    			type leafref {
    			path "/.../.../.../temperature-alarm";
        		}
        		 mandatory true;
    		}
    	}

    leaf temperature-alarm {
                    description 
                    "Alarm for temperature overheat";
                    type temperature-alarm;
                }
   
  typedef temperature-alarm {
        type enumeration {
            enum "normal";
            enum "too-high";
        }
        default normal;
    }
4

When registering your notification stream provider, did you set get_log_times = NULL? Assuming that you are not implementing your own external replay support.

From $CONFD_DIR/examples.confd/netconf_notifications/notifier_builtin_replay_store.c

    ncb.fd = workersock;
    ncb.get_log_times = NULL;
    ncb.replay = NULL;
    strcpy(ncb.streamname, "interface");
    ncb.cb_opaque = NULL;
    if (confd_register_notification_stream(dctx, &ncb, &live_ctx) != CONFD_OK) {
        confd_fatal("Couldn't register stream %s\n", ncb.streamname);
    }
5

I added ncb.get_log_times = NULL; now but i still get the same error in the log.

6

Your issue is likely that in confd.conf (the confd configuration file) you enabled replay support but did not enable the built in replay store. From the ConfD NETCONF notification example:

<notifications>
    <eventStreams>
      <stream>
        <name>interface</name>
        <description>Example notifications</description>
        <replaySupport>true</replaySupport>
        <builtinReplayStore>
Here you need to add: <enabled>true</enabled>
or you may want to disable replay support above
          <dir>./</dir>
          <maxSize>S1M</maxSize>
          <maxFiles>5</maxFiles>
        </builtinReplayStore>
      </stream>
    </eventStreams>
  </notifications>

Since you did not enable the built in replay store, but you enabled replay, ConfD expects that you registered the callback to handle the replay, e.g, the get_log_times callback.

1 Like