about installing ntop

hi,
i have installed ntop .

apt-get instal ntop .

and I do not edit any file manually .

my network

192.168.10.x
||||||||
(switch) --- Ntop server 192.168.10.8:3000
|
(Router-192.168.10.1)

It seems that ntop is not shows data properly .
Some one gives download but it does not shows that it consumes
bandwidth !!!!
Do I need to adit any file to fit it my network ???

On 1 Jul 2006, in the Usenet newsgroup comp.os.linux.networking, in article
<(E-Mail Removed). com>, learnq wrote:

>my network
>
>192.168.10.x
>||||||||
>(switch) --- Ntop server 192.168.10.8:3000
> |
>(Router-192.168.10.1)

Problem - while ntop will put your Ethernet card into promiscuous mode, it
won't do anything to the switch.

>It seems that ntop is not shows data properly .
>Some one gives download but it does not shows that it consumes
>bandwidth !!!!

Unless you have connected the ntop server to a "monitor" port on the switch,
you will see no traffic passing between the router and the _rest_ of your
network. A switch only passes packets that are destined for the host on a
specific port. Thus, the only thing that the ntop server will see is the
traffic for 192.168.10.8 and any broadcast packets (generally ARP requests).

>Do I need to adit any file to fit it my network ???

No - but if you want to monitor the rest of the network, you need to read
the manual for your switch and find out how to set the port you are using
for the ntop server into 'promiscuous' or 'monitor' mode.

Old guy

In comp.os.linux.networking Moe Trin <(E-Mail Removed)>:
> On 1 Jul 2006, in the Usenet newsgroup comp.os.linux.networking, in article
> <(E-Mail Removed). com>, learnq wrote:

>>my network
>>
>>192.168.10.x
>>||||||||
>>(switch) --- Ntop server 192.168.10.8:3000
>> |
>>(Router-192.168.10.1)

> Problem - while ntop will put your Ethernet card into promiscuous mode, it
> won't do anything to the switch.

>>It seems that ntop is not shows data properly .
>>Some one gives download but it does not shows that it consumes
>>bandwidth !!!!

> Unless you have connected the ntop server to a "monitor" port on the switch,
> you will see no traffic passing between the router and the _rest_ of your
> network. A switch only passes packets that are destined for the host on a
> specific port. Thus, the only thing that the ntop server will see is the
> traffic for 192.168.10.8 and any broadcast packets (generally ARP requests).

Let's not forget about netbios and other crap chatty M$ clients
tend to happily broadcast...

>>Do I need to adit any file to fit it my network ???

> No - but if you want to monitor the rest of the network, you need to read
> the manual for your switch and find out how to set the port you are using
> for the ntop server into 'promiscuous' or 'monitor' mode.

Unlikely this will work with most cheapo home network equipment.

While looking for a GB switch for my (small) home lan it took
some time to find a switch that is configurable (RS232 login)
while remaining halfway affordable. The only one I found
available to me and bought was a 1000/100/10Mbit "longshine" 8
port switch.

http://www.longshine.de/longshine/p_...t.php?lang=eng

Gigabit 8-Port Switch
- 8 x 10/100/1000 MBit/s Port
- 1 x RS-232 Consoleport
- VLAN, Trunking # It supports QoS, monitoring and a few other
nifty features
- Desktop, optional 19" Bracket

Anything else in the lines from other manufacturers available to
me* was at least 3-5x the price.

Haven't used it much, unluckily it has a fan (pretty small 40mm
or so) which makes more noise then wanted. Have to exchange it
with a larger one. On testing (nfs) saw zero speed degradation if
using the switch or direct cables, so it seem to work quite fine.

Seems with another fan or for someone who has the switch in a
room where the noise doesn't matter it's a pretty good deal for
his money.

*YMMV

--
Michael Heiming (X-PGP-Sig > GPG-Key ID: EDD27B94)
mail: echo (E-Mail Removed) | perl -pe 'y/a-z/n-za-m/'
#bofh excuse 143: had to use hammer to free stuck disk drive
heads.

On Sat, 1 Jul 2006, in the Usenet newsgroup comp.os.linux.networking, in article
<ao7jn3-(E-Mail Removed)>, Michael Heiming wrote:

>Let's not forget about netbios and other crap chatty M$ clients
>tend to happily broadcast...

What's netbios? The only M$ products I have a a few mice, and they're
pretty quiet.

>Unlikely this will work with most cheapo home network equipment.

True

>While looking for a GB switch for my (small) home lan it took
>some time to find a switch that is configurable (RS232 login)
>while remaining halfway affordable. The only one I found
>available to me and bought was a 1000/100/10Mbit "longshine" 8
>port switch.

What's the bandwidth verses cost? I know in 100BaseT, the extra bux buys a
lot more bandwidth.

>Anything else in the lines from other manufacturers available to
>me* was at least 3-5x the price.

Gotta be there at the right time. I got a pair of 3Com Superstack II 12
port for nearly nothing when someone was upgrading to fiber. I saw no
reason to refuse the kind offer. Unfortunately, they kept the Gigabit
copper stuff, for use elsewhere.

>Seems with another fan or for someone who has the switch in a
>room where the noise doesn't matter it's a pretty good deal for
>his money.

As mentioned before, we have extra fans on just about everything because it
gets warm in where the computers are (it's +43C outside at the moment, and
thanks to the air conditioner, it's only +25C inside). But the stereo is
on, and that masks a lot of the other noise.

>*YMMV

As always, unless those are official government figures in which case YMWABL[1].

Old guy

[1] Your Mileage Will _ALWAYS_ Be Less

In comp.os.linux.networking Moe Trin <(E-Mail Removed)>:
> On Sat, 1 Jul 2006, in the Usenet newsgroup comp.os.linux.networking, in article
> <ao7jn3-(E-Mail Removed)>, Michael Heiming wrote:

[..]

>>While looking for a GB switch for my (small) home lan it took
>>some time to find a switch that is configurable (RS232 login)
>>while remaining halfway affordable. The only one I found
>>available to me and bought was a 1000/100/10Mbit "longshine" 8
>>port switch.

> What's the bandwidth verses cost? I know in 100BaseT, the extra bux buys a
> lot more bandwidth.

Iirc switch + 2 intel GB nics (cooper) were about 250,- Euro
(some years ago). Perhaps about 5x the price of 100 MBit/sec for
10x the bandwidth. Hard to compare if you go for 100MBit nics +
switch as cheap as possible it might be even more.

[..]

> As mentioned before, we have extra fans on just about everything because it
> gets warm in where the computers are (it's +43C outside at the moment, and
> thanks to the air conditioner, it's only +25C inside). But the stereo is
> on, and that masks a lot of the other noise.

Just turned on my mobile dual tube A/C, it's supposed to be hot
again today, it's just 7:30 (GMT+2). About 30-33C, doesn't sound
that much but high humidity kills you off even at those
temperatures in this country. ;(

LCD monitors help lots, CRT types produce magnitudes more heat.

[..]

--
Michael Heiming (X-PGP-Sig > GPG-Key ID: EDD27B94)
mail: echo (E-Mail Removed) | perl -pe 'y/a-z/n-za-m/'
#bofh excuse 45: virus attack, luser responsible

On Sun, 2 Jul 2006, in the Usenet newsgroup comp.os.linux.networking, in
article <jf2kn3-(E-Mail Removed)>, Michael Heiming wrote:

>In comp.os.linux.networking Moe Trin <(E-Mail Removed)>:

>> What's the bandwidth verses cost? I know in 100BaseT, the extra bux buys a
>> lot more bandwidth.
>
>Iirc switch + 2 intel GB nics (cooper) were about 250,- Euro
>(some years ago). Perhaps about 5x the price of 100 MBit/sec for
>10x the bandwidth. Hard to compare if you go for 100MBit nics +
>switch as cheap as possible it might be even more.

Misunderstood question. Ethernet switches, some dual speed hubs, and
routers are "Store and Forward" devices. The Ethernet switch has no idea
where the packet coming in on port 1 is going until it sees the first six
bytes of the packet. Even if it instantly set up to shuffle the bits to
the "right" output port, it's already 14 bytes behind (8 bytes of sync and
the first 6 bytes of the header), and that assumes that the output port
isn't busy at the time. (A router has an even harder task, because the
destination IP address is the 17th to 20th byte into the packet.) Thus,
the switch has to be able to store (at least some portion of) the packet
and that takes memory space. Don't you be thinkin' to long about this,
because the host on port 1 is being talkative, and here comes another
packet right now. Pull your fingers out!!!

The network interfaces almost certainly are going to convert the serial
bit stream of the Ethernet packet into something parallel - perhaps 32
or 64 bits wide - to gain speed capability. A Gigabit packet on a 64 bit
PCI bus (as an example) is a mere 15.7 mega(really-wide)words per second.
But what happens if a tenth of the hosts connected to your switch are
talking at the moment. Now you know why switches aren't 64 or 100 ports
wide.

A measure of performance is the _total_ bandwidth of the device - what
is the sum of the bits/second that the device itself (not one individual
interface) can handle. I've seen 100BaseT switches with 110 Mbps total
bandwidth - but you can also get total bandwidths up into 5 Gbps range for
a _VERY_ appreciable increase in cost.

>Just turned on my mobile dual tube A/C, it's supposed to be hot
>again today, it's just 7:30 (GMT+2). About 30-33C, doesn't sound
>that much but high humidity kills you off even at those
>temperatures in this country. ;(

My Atlas doesn't have a lot of relatively local weather statistics for
cities around the world, but that is 7-10C above average for the lower
elevations in Germany. Yes, I'm aware of humidity, having lived within
20 KM of the Gulf of Mexico (Houston TX sees above 90F and 90% for a lot
of the summer - and even Jean-David Beyer has mentioned similar numbers
a few tens of KM South of New York City) as well as parts of Southeast
Asia (Hong Kong can be extremely miserable for that). For much of the
year, we're at the opposite end - it's normally quite dry, and we average
18 cm of rain a year. It's been rather warm recently, and this creates a
"Thermal Low" pressure area, which tends to suck in moisture from the
oceans (the cause of all that fog in San Francisco is the thermal low in
the Central Valley some 80-120 KM to the East). It's quite sultry right
now, and the 10AM (GMT-7) reports from Phoenix are 94F (34.4C), with a
dew point of 57F which gives a Relative Humidity of 28%. The 10 day
forecast is calling for a 10% chance of scattered thunderstorms in the
afternoons - which is where most of that 18 cm per year comes from.
Before you laugh at that "high" humidity, know that the annual average is
more like 15%, which is how we can (sorta) tolerate those +45C highs that
are not all that uncommon here.

>LCD monitors help lots, CRT types produce magnitudes more heat.

You're telling me? Actually, I have four CRTs in this room, and only
one is turned on, along with two LCDs.

Old guy

In comp.os.linux.networking Moe Trin <(E-Mail Removed)>:
> On Sun, 2 Jul 2006, in the Usenet newsgroup comp.os.linux.networking, in
> article <jf2kn3-(E-Mail Removed)>, Michael Heiming wrote:

>>In comp.os.linux.networking Moe Trin <(E-Mail Removed)>:

>>> What's the bandwidth verses cost? I know in 100BaseT, the extra bux buys a
>>> lot more bandwidth.
>>
>>Iirc switch + 2 intel GB nics (cooper) were about 250,- Euro
>>(some years ago). Perhaps about 5x the price of 100 MBit/sec for
>>10x the bandwidth. Hard to compare if you go for 100MBit nics +
>>switch as cheap as possible it might be even more.

> Misunderstood question. Ethernet switches, some dual speed hubs, and

Badly phrased question?

[ how a switch works ]

> A measure of performance is the _total_ bandwidth of the device - what
> is the sum of the bits/second that the device itself (not one individual
> interface) can handle. I've seen 100BaseT switches with 110 Mbps total
> bandwidth - but you can also get total bandwidths up into 5 Gbps range for
> a _VERY_ appreciable increase in cost.

Not mentioned in the manual:

[ snip more features ]
- Supports auto MDI/MDI-X function on all ports.
- Supports Full/Half Duplex mode on 10/100Mbps speed and Full
Duplex mode on 1000Mbps speed.
- Wire-speed packet filtering and forwarding rate.
- Store-and-forward architecture filters fragment & CRC error
packets.
- Supports 7 groups port-base VLAN
- Supports 2 groups port-trunk
- Supports 4-level IP priority
- Supports 4-level Port priority
- Supports 8K MAC address learning table
- Supports 256KBytes buffer memory (SSRAM)
- Supports extensive LED indicators for network diagnostics

The 256KB SSRAM give some indication. In reality I did not care
much about total bandwidth, as there are seldom more then 2-4
systems connected at all. Perhaps you overlooked that this is my
home lan *not* some $$ production environment? ;-)

>>Just turned on my mobile dual tube A/C, it's supposed to be hot
>>again today, it's just 7:30 (GMT+2). About 30-33C, doesn't sound
>>that much but high humidity kills you off even at those
>>temperatures in this country. ;(

> My Atlas doesn't have a lot of relatively local weather statistics for
> cities around the world, but that is 7-10C above average for the lower
> elevations in Germany. Yes, I'm aware of humidity, having lived within
> 20 KM of the Gulf of Mexico (Houston TX sees above 90F and 90% for a lot
> of the summer - and even Jean-David Beyer has mentioned similar numbers
> a few tens of KM South of New York City) as well as parts of Southeast
> Asia (Hong Kong can be extremely miserable for that). For much of the
> year, we're at the opposite end - it's normally quite dry, and we average
> 18 cm of rain a year. It's been rather warm recently, and this creates a
> "Thermal Low" pressure area, which tends to suck in moisture from the
> oceans (the cause of all that fog in San Francisco is the thermal low in
> the Central Valley some 80-120 KM to the East). It's quite sultry right
> now, and the 10AM (GMT-7) reports from Phoenix are 94F (34.4C), with a
> dew point of 57F which gives a Relative Humidity of 28%. The 10 day

Yep, that was my point. Low humidity makes it far easier to bear
with high temps.

> forecast is calling for a 10% chance of scattered thunderstorms in the
> afternoons - which is where most of that 18 cm per year comes from.

18 cm rain/year? If it rains here 18 cm a single day, none would
be surprised here. It can take a week or longer once it starts
raining. There's always an umbrella in my daypack. Of course this
depends on the area, there are dryer places in Germany.

> Before you laugh at that "high" humidity, know that the annual average is
> more like 15%, which is how we can (sorta) tolerate those +45C highs that
> are not all that uncommon here.

>>LCD monitors help lots, CRT types produce magnitudes more heat.

> You're telling me? Actually, I have four CRTs in this room, and only
> one is turned on, along with two LCDs.

Have 2 LCD (2*19" in dual-head mode) running, reading mostly text
I never looked back to CRT. The heat output compared with a 21"
CRT, which has the same viewable size is astonishing. While LCD
just outputs very little warm air over a very small area, 21"
CRTs make a nice heat you could perhaps roast beans on...

I know you are aware about the fact, but I have to take
lurkers/etc into account. ;-)

--
Michael Heiming (X-PGP-Sig > GPG-Key ID: EDD27B94)
mail: echo (E-Mail Removed) | perl -pe 'y/a-z/n-za-m/'
#bofh excuse 88: Boss' kid fucked up the machine

On Sun, 2 Jul 2006, in the Usenet newsgroup comp.os.linux.networking, in article
<bsiln3-(E-Mail Removed)>, Michael Heiming wrote:
>In comp.os.linux.networking Moe Trin <(E-Mail Removed)>:

>> Misunderstood question. Ethernet switches, some dual speed hubs, and
>
>Badly phrased question?

Perhaps

>Not mentioned in the manual:

[snip marketing blather]

>The 256KB SSRAM give some indication.

It does, and it doesn't. I'd suspect because it isn't mentioned explicitly
that it's not a large number.

>In reality I did not care much about total bandwidth, as there are seldom
>more then 2-4 systems connected at all. Perhaps you overlooked that this
>is my home lan *not* some $$ production environment? ;-)

That's also why I haven't made the jump to Gigabit. Up until early last
year, I was running 10Base2 in the house, because that was more than
enough for what I was using. We updated to 100BaseT because it was supported
by the cable modem, and I got those two Superstacks for ten bucks each. I
already had the NIC capability except on the box that was running the POTS
modem and firewall (an old 386 laptop).

>Yep, that was my point. Low humidity makes it far easier to bear
>with high temps.

Except there is to much of a good thing. We have four humidifiers running
in the house. First week I was here, I walked into this room, reached to
the CPU of a workstation to unlock the keyboard (old IBM AT style case),
and drew an arc about a half inch long... which caused the computer to
reboot. I've also got scars on the arms from touching one of the cats, and
drawing a similar arc. Hmmm, maybe 6% humidity isn't such a good thing.
(The hygrometer is showing 47% now, while NWS is saying 21% downtown.

>18 cm rain/year? If it rains here 18 cm a single day, none would
>be surprised here. It can take a week or longer once it starts
>raining.

Obviously, it varies all over the world. I was running a field test in South
East Asia, and measured (drinking glass set on a table for 10 minutes) 17.8
cm/hour. We can get thunderstorms to drop as much as 3 cm in an hour or less,
but the storm cell is only a 4-8 KM across. This means 10 KM away may not
even see a single raindrop.

>There's always an umbrella in my daypack. Of course this depends on the
>area, there are dryer places in Germany.

whois doesn't tell me, but you're down near Bamberg, right? Or is that just
a colo. The maps I have suggest it's drier up near Erfurt to Mersburg,

>CRTs make a nice heat you could perhaps roast beans on...

The cat used to think that one monitor I had with a relatively flat top was
just a perfect place to sleep. Sure glad that was a rental system - but I
wonder what the tech thought when we returned it.

Old guy

In comp.os.linux.networking Moe Trin <(E-Mail Removed)>:
> On Sun, 2 Jul 2006, in the Usenet newsgroup comp.os.linux.networking, in article
> <bsiln3-(E-Mail Removed)>, Michael Heiming wrote:
>>In comp.os.linux.networking Moe Trin <(E-Mail Removed)>:

[stuff]

>>Yep, that was my point. Low humidity makes it far easier to bear
>>with high temps.

> Except there is to much of a good thing. We have four humidifiers running
> in the house. First week I was here, I walked into this room, reached to
> the CPU of a workstation to unlock the keyboard (old IBM AT style case),
> and drew an arc about a half inch long... which caused the computer to
> reboot. I've also got scars on the arms from touching one of the cats, and
> drawing a similar arc. Hmmm, maybe 6% humidity isn't such a good thing.

Not that much, thought about 30% in opposite to 90% which would
allow to bear better with high temps. Came home an hour ago and
it was about 28/82.4 C/F with approximately 95% humidity. A/C is
running and now it is 26/78.8 C/F but humidity perhaps 60% and
makes you feel already quite comfortable.

The lowered humidity really seems to make the difference, not the
little temperature reduction. Need to get my hands on some cheapo
humidity measurement equipment. Preferable with serial
connector, one could make some nice mrtg graphs. ;-)

> (The hygrometer is showing 47% now, while NWS is saying 21% downtown.

>>18 cm rain/year? If it rains here 18 cm a single day, none would
>>be surprised here. It can take a week or longer once it starts
>>raining.

> Obviously, it varies all over the world. I was running a field test in South
> East Asia, and measured (drinking glass set on a table for 10 minutes) 17.8
> cm/hour. We can get thunderstorms to drop as much as 3 cm in an hour or less,

You should be able to get this here although. ;-)

> but the storm cell is only a 4-8 KM across. This means 10 KM away may not
> even see a single raindrop.

>>There's always an umbrella in my daypack. Of course this depends on the
>>area, there are dryer places in Germany.

> whois doesn't tell me, but you're down near Bamberg, right? Or is that just

Not really, denic cut down access to whois due to multiple abuse.
The web server of my domain has moved to some location near some
time ago, iirc.
[..]

>>CRTs make a nice heat you could perhaps roast beans on...

> The cat used to think that one monitor I had with a relatively flat top was
> just a perfect place to sleep. Sure glad that was a rental system - but I
> wonder what the tech thought when we returned it.

;-)

--
Michael Heiming (X-PGP-Sig > GPG-Key ID: EDD27B94)
mail: echo (E-Mail Removed) | perl -pe 'y/a-z/n-za-m/'
#bofh excuse 143: had to use hammer to free stuck disk drive
heads.