Send netdisco-users mailing list submissions to
[email protected]
To subscribe or unsubscribe via the World Wide Web, visit
https://lists.sourceforge.net/lists/listinfo/netdisco-users
or, via email, send a message with subject or body 'help' to
[email protected]
You can reach the person managing the list at
[email protected]
When replying, please edit your Subject line so it is more specific
than "Re: Contents of netdisco-users digest..."
Today's Topics:
1. Netdisco 2 Port Information Update (Pavel Skovajsa)
2. Re: Netdisco 2 Port Information Update (Alan Buxey)
3. Re: Netdisco 2 Port Information Update (Pavel Skovajsa)
4. Re: Netdisco 2 Port Information Update (Oliver Gorwits)
5. Netdisco 1.3.2 Neighbor discovery (Simon De Matos)
6. Netdisco 2 - small bugs when running with a custom path
(Daniel Tuecks)
--- Begin Message ---
Hello there,
I use ND2 heavily in our enterprise for such things as 10Mbps Speed
reporting or Half Duplex Reporting. This is fairly simple, I use a
query below:
SELECT device.name,model,serial,device.ip,port,duplex,duplex_admin
FROM device_port INNER JOIN device ON device_port.ip=device.ip
WHERE up='up' AND speed='10 Mbps' AND port LIKE '%Ethernet%'
Going through these reports we have found many occasions of "false
positives" where a port which is reported as 10Mbps is actually
running 100Mbs in the moment. The obvious answer is that the
information in ND2 is not real-time and the port WAS running 10Mbps
sometime in the past.........which brings me to these questions:
1. I would expect ND2 updates the port information( (status,
speed,duplex, admin duplex)regularly. Can you explain the mechanism
used? This should be probably one of the 'housekeeping' processes,
that control how often is this information updated.
2. Is there any history information about the port status in the DB,
similarly to MAC 'first/last seen' info?
3. Why some (seemingly random) ports on the switches do not have any
duplex,admin_duplex information, and yet there is a 'connected device'
associated with these. If needed I can provide example offline. This
might be connected to Q4.
4. Why some (seemingly random) ports on the switches have a little
cross meaning "blocking" yet they have 'connected device' associated.
What does it mean 'blocking'? I have tons of these, they cannot be all
STP blocked....Might be connected to Q3.
5. Some ports on switches are reported with 'admin_duplex' as full,
yet they have 'duplex' as half, which is obviously nonsense. Is this a
snmp switch reporting issue or these fields get updated in different
time causing loss of sync?
That's it for now. Thanks a lot for looking into these and helping me
understand on how the mechanism works.
-pavel
--- End Message ---
--- Begin Message ---
There are jobs which go through all known switches. So whatever state the
ports etc were in when that happened is what is recorded. Don't forget that
when a machine is sleeping/suspended or off, then the port will usually drop to
10/half rather than be down/disconnected (WoL etc - you'll see the lights
blinking on back of PC or server ... well if it's got any lights. .. too many
NICs these days don't have visual status :(
alan
--- End Message ---
--- Begin Message ---
Exactly, I have not really explained my big plan here. I already
figured out that machines sleeping/suspended are causing majority of
these false positives.
My end goal here is to minimize these by figuring out WHICH ND process
updates the port information and WHEN the ND process runs for each
device. After that I was thinking about doing some of the following:
opt1. tweaking that process to run only during business hours
opt2. run the process every hour and somehow find 'alternating ports'
(ports which change speed/duplex often). That is why I was asking if
there is any historic info about port state.
hope it helps,
-pavel
On Fri, Nov 8, 2013 at 9:06 PM, Alan Buxey <[email protected]> wrote:
> There are jobs which go through all known switches. So whatever state the
> ports etc were in when that happened is what is recorded. Don't forget that
> when a machine is sleeping/suspended or off, then the port will usually drop
> to 10/half rather than be down/disconnected (WoL etc - you'll see the lights
> blinking on back of PC or server ... well if it's got any lights. .. too
> many NICs these days don't have visual status :(
>
> alan
--- End Message ---
--- Begin Message ---
Hi Pavel,
Here is a short explanation of what Netdisco is doing. If anything's
not clear, please ask.
On 2013-11-08 19:03, Pavel Skovajsa wrote:
1. I would expect ND2 updates the port information( (status,
speed,duplex, admin duplex)regularly. Can you explain the mechanism
used? This should be probably one of the 'housekeeping' processes,
that control how often is this information updated.
There are three main background jobs which Netdisco runs:
* discover - this is usually run once a day, and it updates information
in the database about switches and routers on your network, but _not_
information about end hosts such as workstations/servers/printers/etc.
This job discovers the ports on the device, and their configuration.
* macsuck - this is usually run every hour, and it finds out which MAC
addresses are connected to which switch ports.
* arpnip - this is usually run every hour, and it finds out the mapping
of IP addresses to MAC addresses on the network.
In the config file under "housekeeping" are the discoverall, macwalk,
and arpwalk jobs. These are triggers to enqueue a
discover/macsuck/arpnip for all known devices in the database (a
side-effect of discover is also to find new devices).
2. Is there any history information about the port status in the DB,
similarly to MAC 'first/last seen' info?
No, there isn't. However the lastchange field of the port should
contain a timestamp of when the port last changed state.
3. Why some (seemingly random) ports on the switches do not have any
duplex,admin_duplex information, and yet there is a 'connected
device'
associated with these. If needed I can provide example offline. This
might be connected to Q4.
Currently, Netdisco 2 will change the port state to "up" if the port
was previously down but an end host is now seen on the port. As a
consequence, the duplex and speed are cleared, because they're not known
until a discovery takes place (once a day).
However, you mention a connected device not a connected node (end
host). This makes me think there is a bug either in Netdisco or the SNMP
data returned by your device.
4. Why some (seemingly random) ports on the switches have a little
cross meaning "blocking" yet they have 'connected device' associated.
What does it mean 'blocking'? I have tons of these, they cannot be
all
STP blocked....Might be connected to Q3.
The spanning-tree blocking code probably isn't smart enough :-)
Certainly where there's a native (untagged) VLAN and some tagged VLANs,
I think it only considers the native VLAN. To be honest I have no idea
how this is worked out!
5. Some ports on switches are reported with 'admin_duplex' as full,
yet they have 'duplex' as half, which is obviously nonsense. Is this
a
snmp switch reporting issue or these fields get updated in different
time causing loss of sync?
It could be a bug in the SNMP reporting, in which case we'd probably
want to see the output of an snmpwalk of the interfaces tables.
regards,
oliver.
That's it for now. Thanks a lot for looking into these and helping me
understand on how the mechanism works.
-pavel
------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models.
Explore
techniques for threading, error checking, porting, and tuning. Get
the most
from the latest Intel processors and coprocessors. See abstracts and
register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Netdisco mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/netdisco-users
--- End Message ---
--- Begin Message ---
Hello,
I just installed Netdisco on a Debian Wheezy and I'm currently having
an error with the neighbor discovery.
We're using Cisco, HP and 3Com switches, so I followed jrbinks'
documentation to make Netdisco work properly with 3Com products
(patching Info.pm, getting the missing 3Com mibs) but when trying a
network discovery from one of the 3Com switches I get this:
[3Com-4800G-switch] Fetching Neighbor Information
SNMP::Info::_validate_autoload_method(hasLLDP) Unable to resolve method.
SNMP::Info::_validate_autoload_method(hasCDP) Unable to resolve method.
SNMP::Info::_validate_autoload_method(hasSONMP) Unable to resolve method.
SNMP::Info::_validate_autoload_method(hasFDP) Unable to resolve method.
SNMP::Info::_validate_autoload_method(hasEDP) Unable to resolve method.
SNMP::Info::_validate_autoload_method(hasAMAP) Unable to resolve method.
SNMP::Info::_validate_autoload_method(hasCDP) Unable to resolve method.
[3Com-4800g-switch] ! CDP not Enabled.
Network Discovery complete.
The discovery works fine with the other brands. I'm not familiar with
CPAN/SNMP::Info/Netdisco and my searches weren't really successful ...
Hoping someone here knows what could cause these errors (or what I'm
missing) because I'm currently running out of ideas.
Best regards,
Simon De Matos
----------------------------------
Service Infrastructures
Direction du Système d'Information
Pôle Ressources
----------------------------------
Université de Limoges
123, avenue Albert Thomas
87060 Limoges cedex
--- End Message ---
--- Begin Message ---
Hi all,
I've found two small bugs when deploying netdisco2 to a custom path (path:
/netdisco2).
The first one occurs, when you log into the new netdisco installation.
Right after the login process, one gets redirected to '/inventory' instead
of '/netdisco2/inventory'.
The second bug affects .csv export. This is basically the the same error as
the one above: The generated download link is '/ajax/....' (should be
'/netdisco2/ajax/...').
I've attached patches to fix these bugs. Hope this is of any help for you.
regards
Daniel
diff -crB netdisco-netdisco-ng-f1de543acf1d3f5fa4aeef6f704cae8a9010ec86/Netdisco/share/views/js/common.js netdisco-netdisco-ng-dtuecks/Netdisco/share/views/js/common.js
*** netdisco-netdisco-ng-f1de543acf1d3f5fa4aeef6f704cae8a9010ec86/Netdisco/share/views/js/common.js 2013-11-09 00:50:50.000000000 +0100
--- netdisco-netdisco-ng-dtuecks/Netdisco/share/views/js/common.js 2013-11-12 13:52:56.015474050 +0100
***************
*** 6,12 ****
if (show.length) {
$('#nd_csv-download')
! .attr('href', '/ajax/content/' + type + '/' + tab + '?' + query)
.attr('download', 'netdisco-' + type + '-' + tab + '.csv')
.show();
}
--- 6,12 ----
if (show.length) {
$('#nd_csv-download')
! .attr('href', uri_base + '/ajax/content/' + type + '/' + tab + '?' + query)
.attr('download', 'netdisco-' + type + '-' + tab + '.csv')
.show();
}
diff -crB netdisco-netdisco-ng-f1de543acf1d3f5fa4aeef6f704cae8a9010ec86/Netdisco/lib/App/Netdisco/Web/AuthN.pm netdisco-netdisco-ng-dtuecks/Netdisco/lib/App/Netdisco/Web/AuthN.pm
*** netdisco-netdisco-ng-f1de543acf1d3f5fa4aeef6f704cae8a9010ec86/Netdisco/lib/App/Netdisco/Web/AuthN.pm 2013-11-09 00:50:50.000000000 +0100
--- netdisco-netdisco-ng-dtuecks/Netdisco/lib/App/Netdisco/Web/AuthN.pm 2013-11-12 14:00:05.979236615 +0100
***************
*** 5,11 ****
use Dancer::Plugin::Auth::Extensible;
hook 'before' => sub {
! params->{return_url} ||= '/inventory';
if (! session('logged_in_user') && request->path ne uri_for('/login')->path) {
if (setting('trust_x_remote_user') and scalar request->header('X-REMOTE_USER')) {
--- 5,11 ----
use Dancer::Plugin::Auth::Extensible;
hook 'before' => sub {
! params->{return_url} ||= uri_for('/inventory');
if (! session('logged_in_user') && request->path ne uri_for('/login')->path) {
if (setting('trust_x_remote_user') and scalar request->header('X-REMOTE_USER')) {
--- End Message ---
------------------------------------------------------------------------------
November Webinars for C, C++, Fortran Developers
Accelerate application performance with scalable programming models. Explore
techniques for threading, error checking, porting, and tuning. Get the most
from the latest Intel processors and coprocessors. See abstracts and register
http://pubads.g.doubleclick.net/gampad/clk?id=60136231&iu=/4140/ostg.clktrk
_______________________________________________
Netdisco mailing list - Digest Mode
[email protected]
https://lists.sourceforge.net/lists/listinfo/netdisco-users
