Well, but via SNMP you have direct info about trunks, vlans and so on. Maybe if finding whole chain fails You can always display the vlan number on port, because it really is true.
It doesn't have to be though, many devices will just have the vlan at the default '1' value which is why we have the current solution.
I can provide snmp walk and oids for this, we are running mrtg and rrdtool
for
taking stats from linux machines, it is really no problem, mrtg does it automatically. If you make some interface for this I can try to put some linux stuff to Nav.
Do you have experience with Java programming? If so it should be easy to add yourself. If not please send any SNMP walks and we can try to add it for you.
However, we cannot display VLAN numbers for trunks without actually traversing the whole chain.
We can only access the VLANs allowed on the trunk (we already collect this info), but this might be very different from what is actually running (many sites simply allow all VLANs on the trunks). For this reason we try to determine which VLANs are actually active on the trunk.
Last question, bugfixes (ip prefix scope graph, cricket desc for routers
etc.)
- when? :)
Fixes will be part of the beta9 release, due out shortly (hopefully today or tomorrow). As for the cricket bug you can apply the fix ourself. Edit the file:
/usr/share/doc/nav/cricket/cricket-config/router-interfaces/.nav
On the line starting with descr = change the SQL statement to:
select nettype,netident from gwportprefix left join prefix using (prefixid) left join vlan using (vlan) where gwportid=$id and netident is not null
Then run makecricketconfig.
-- Kristian
Dnia czwartek, 20 stycznia 2005 11:10, Kristian Eide napisa?:
Well, but via SNMP you have direct info about trunks, vlans and so on. Maybe if finding whole chain fails You can always display the vlan number on port, because it really is true.
It doesn't have to be though, many devices will just have the vlan at the default '1' value which is why we have the current solution.
ok, that make sense, so maybe just take vlans which id is not "1" ;)
I can provide snmp walk and oids for this, we are running mrtg and rrdtool
for
taking stats from linux machines, it is really no problem, mrtg does it automatically. If you make some interface for this I can try to put some linux stuff to Nav.
Do you have experience with Java programming? If so it should be easy to add yourself. If not please send any SNMP walks and we can try to add it for you.
Well, Java is mystery for me... bash, awk, perl, sed - this is my language :) The best solution would be some interface for adding something new devices, and snmp-specific data for them. So, You want me to send snmpwalk, or just oids for collecting data from interfaces or cpu?
However, we cannot display VLAN numbers for trunks without actually traversing the whole chain.
We can only access the VLANs allowed on the trunk (we already collect this info), but this might be very different from what is actually running (many sites simply allow all VLANs on the trunks). For this reason we try to determine which VLANs are actually active on the trunk.
I haven't got any link that allows any vlan on trunk, but I have static vlans administration, in VTP the maybe different.
Last question, bugfixes (ip prefix scope graph, cricket desc for routers
etc.)
- when? :)
Fixes will be part of the beta9 release, due out shortly (hopefully today or tomorrow).
It is very nice to hear :-D
As for the cricket bug you can apply the fix ourself. Edit the file:
/usr/share/doc/nav/cricket/cricket-config/router-interfaces/.nav
On the line starting with descr = change the SQL statement to:
select nettype,netident from gwportprefix left join prefix using (prefixid) left join vlan using (vlan) where gwportid=$id and netident is not null
Then run makecricketconfig.
after running makecricketconfig:
=> Running makeinterfaceTargets with dir=router-interfaces <= Use of uninitialized value in concatenation (.) or string at /usr/lib/nav/makecricketconfig.pl line 831. [cut]
and now there are no interfaces for routers in cricket. But maybe I should wait...
thanks for this bugfix