Our first postcards :-)

[Morten Brekkevold]

For many years, we've asked NAV users to send us a postcard to tell us
they use NAV (see: http://metanav.uninett.no/postcard).

Up until now, we've received a total sum of 0 (zero) cards :-(

I just want to send a shout-out and a big thanks to Matěj Grégr, from
the University of Brno in the Czech Republic, who recently sent us no
less than _two_ postcards :-)

These are now proudly on display on the window facing the hallway
outside my office: http://yfrog.com/h8qbqgxj

[Joachim Tingvold]

Post by Morten Brekkevold
For many years, we've asked NAV users to send us a postcard to tell us
they use NAV (see: http://metanav.uninett.no/postcard).
Up until now, we've received a total sum of 0 (zero) cards :-(
I just want to send a shout-out and a big thanks to Matěj Grégr, from
the University of Brno in the Czech Republic, who recently sent us no
less than _two_ postcards :-)
These are now proudly on display on the window facing the hallway
outside my office: http://yfrog.com/h8qbqgxj

One more coming up next week. (-:

[Matej Gregr]

Hello,
I noticed, that NAV does not close arp entries after monitored node
become unavailable. For example: There is an active mac address
11:22:33:44:55:66 with end_time set to infinity. When the monitored node
goes down, the mac address remains in active state, so you can find it
using machinetracker as a still active address, which apparently is not
true.

I suggest, that there should be another rule in database, enssuring that
arp entries will be closed.

Something like:

CREATE OR REPLACE RULE netbox_status_close_arp AS ON UPDATE TO netbox
WHERE NEW.up='n'
DO UPDATE arp SET end_time=NOW()
WHERE netboxid=OLD.netboxid AND end_time='infinity';


Regards,

Matej

[Morten Brekkevold]

Post by Matej Gregr
Hello,
I noticed, that NAV does not close arp entries after monitored node
become unavailable. For example: There is an active mac address
11:22:33:44:55:66 with end_time set to infinity. When the monitored node
goes down, the mac address remains in active state, so you can find it
using machinetracker as a still active address, which apparently is not
true.
I suggest, that there should be another rule in database, enssuring that
arp entries will be closed.
CREATE OR REPLACE RULE netbox_status_close_arp AS ON UPDATE TO netbox
WHERE NEW.up='n'
DO UPDATE arp SET end_time=NOW()
WHERE netboxid=OLD.netboxid AND end_time='infinity';

A very astute observation :-) ipdevpoll will not update any data
related to a node while the node is down (though the ancient
Perl-implementation of the arplogger probably did).

Your rule appears to be just the thing we need here - I've nicked it
into NAV and attributed it to you :-)

https://bugs.launchpad.net/nav/+bug/790600