jlixfeld wrote:
> 
> Bill Fisher wrote:
> > 1. Sometimes a switch port is labeled with a Subnet list (usually ifIndex 
> > 1). The L2 engine does not automatically change network labels to the port 
> > number - you can do this manually.
> > 
> 
> 
> So what I hear you saying is that those ovals never disappear.  They show a 
> subnet it's a layer 3 connection and they (are normally supposed to) show a 
> port number of it's a layer 2 connection.
> 
> That said, if I manually change the label, it will be persistent regardless 
> as to whether it is a layer 3 connection or a layer 2 connection?
> 


There's no distinction between layer 3 and layer 2 connections internally. The 
ovals (aka network objects) may be hidden, but they remain persistent.  The 
only case where a network oval may be deleted is when it's not connected to 
anything *and* the label is empty.  Note that the lines connected to the 
network object may change.


jlixfeld wrote:
> 
> 
> Bill Fisher wrote:
> > 
> > 2. It is possible for devices to exist on your map that L2 cannot locate. 
> > When this happens, the connection logic defaults to layer 3.  In the Status 
> > window for an interface, look for the "Interface Connection" section. If 
> > this section exists, it will provide the L2 data used to locate the device.
> > 
> 
> 
> This section does not exist, so it seems that it is defaulting to layer 3.
> 
> Fast forwarding down the thread a bit, is it possible that if I'm hitting the 
> limited mode bug, that might account for why L2 might not be able to locate 
> some of the connections?
> 


Yes -- that's a symptom of the bug.


jlixfeld wrote:
> 
> Just out of curiosity, is there a release schedule for 5.4.5?
> 


 I expect we will put a 5.4.5b1 up on the beta download page soon.


jlixfeld wrote:
> 
> I do indeed have limited mode set because I don't use spanning-tree on my 
> network.  Our network is completely routed, but we use CDP everywhere so from 
> my perspective, limited mode is all we need.  Besides, I'm a little leary 
> about certain layer 2 detection mechanisms.  I already know that there is a 
> bug in the switch software I use that rears it's head if something polls 
> dotidTpFdbTable.  It will immediately crash any devices running the affected 
> software, so I try to keep layer 2 detection limited to CDP, although I'm 
> certainly amenable to enabling LLDP if the feeling is that it will help.
> 


CDP should be more than sufficient; LLDP will just supply redundant information 
in a slightly more verbose manner.


-------------------- m2f --------------------

Read this topic online here:
http://forums.dartware.com/viewtopic.php?p=3523#3523





____________________________________________________________________
List archives: 
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [email protected]

Reply via email to