BVI Interface & Dropped Packets

aquilla

Hi Guys,

I'm hoping can help me understand this. I have a fairly small LAN at home and at the centre is a Cisco 877W (running IOS 12.4(15)T3). I have both wired and wireless devices connecting so I have created a bridge-group.

Yesterday I added a PC (wired) to the network to do a Linux installation. Got the base OS on the PC and decided to do the rest of the config remotely from another PC. Tried to ping the IP of the new Linux box and noticed I was getting a lot of packet loss.

I have been testing this some more today and I installed Windows XP on the new box incase it was a Linux problem, but again I am still seeing a lot of packet loss (from Windows 7 PC -> New ITX Box):

Ping statistics for 192.168.1.99:
Packets: Sent = 300, Received = 105, Lost = 195 (65% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 2ms, Average = 0ms

As a test I replaced the network cables (currently CAT6) with some spares (CAT5e) but the problem continued. I then connected *just* these two hosts to a separate switch (Netgear GS10 and the packet loss vanished (the GS108 was NOT connected to the Cisco at any point during the test). So the cables were good and the PCs were good.

At this point I thought I might be something to do with the bridge-group configuration on the router so I created a new VLAN and moved these two hosts in the new VLAN and off the bridge-group. Surprise surprise the packet loss vanished:

Ping statistics for 172.16.30.99:
Packets: Sent = 1000, Received = 1000, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 7ms, Average = 0ms

Please can you help me try and understand why I am seeing packet loss over the BVI interface.

Network diagram here: http://www.zen7850.zen.co.uk/Network.png
Router configuration here: http://www.zen7850.zen.co.uk/Config.txt

Some notes:

Windows 7 PC (IP: 192.168.1.96) is my main PC
Mini ITX Box (IP: 192.168.1.99) is the new addition
There is a wireless laptop (IP: 192.168.1.100) and a Wifi printer (192.168.1.192) connected.
The Draytek Modem (IP: 192.168.1.253) is used to handle the DSL line as it's ADSL2+ and the Cisco doesn't cope very well.

Panzer919

did you happen to see what your interfaces were negotiating to?

You swapped the cables so that takes care of layer 1, Layer 2 would have been interface errors, speed and duplex mismatch, layer 3 is IP but since you could ping the devices that pretty well rules that and layer 4 out.

My guess without seeing it put together is a speed/duplex mismatch somewhere, when you put the PC's into the switch, it all negotiated correctly. Try to get into the habit of hard coding your speed and duplex setting on your interconnects if the device will allow it.

aquilla

Panzer919 wrote:

did you happen to see what your interfaces were negotiating to?

When connected to the Netgear GS108 switch both hosts negotiated to 1Gbps Full Duplex.

On the Cisco 877W they are currently:

Router# show inter fa2
FastEthernet2 is up, line protocol is up
Hardware is Fast Ethernet, address is 001c.0ed8.2b42 (bia 001c.0ed8.2b42)
Description: *** Link to Personal Workstation - NIC 1 ***
MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,
reliability 255/255, txload 2/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 100Mb/s

Router# show inter fa3
FastEthernet3 is up, line protocol is up
Hardware is Fast Ethernet, address is 001c.0ed8.2b43 (bia 001c.0ed8.2b43)
Description: *** Mini ITX ATOM Server ***
MTU 1500 bytes, BW 100000 Kbit, DLY 100 usec,
reliability 255/255, txload 1/255, rxload 1/255
Encapsulation ARPA, loopback not set
Keepalive set (10 sec)
Full-duplex, 100Mb/s

Speed and duplex are auto-negotiated on the Cisco 877W.

Panzer919

did you put your setup back the way it was or is it still "fixed"?

with the output you just showed are you getting packetloss?

aquilla

Panzer919 wrote: »

did you put your setup back the way it was or is it still "fixed"?

with the output you just showed are you getting packetloss?

My setup is back the way it was originally (except the replacement cables because I know they work). Yes I am still seeing packet loss. I set a ping running and it appears to start out OK and then produces very high packet loss (Screenshot - http://www.zen7850.zen.co.uk/Ping.png).

The final results of this quick ping test were:

Ping statistics for 192.168.1.99:
Packets: Sent = 100, Received = 50, Lost = 50 (50% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 0ms, Average = 0ms

Note: There is a link to the current router config in my first post.

Panzer919

so you get packet loss if you use the switch ports on the 877W, you do not get packet loss if you throw another switch in the mix.

I don't remember, do you have to specify the switchports as switchports? I've seen a few devices that had ports that were either switchports or router ports but you had to tell it to act as a switchport.

So
int fa0
switchport
switchport mode access
switchport access vlan 1 <- I know its redundant but its a habit.

try that and see if that works

aquilla

Panzer919 wrote: »

so you get packet loss if you use the switch ports on the 877W, you do not get packet loss if you throw another switch in the mix.

I don't remember, do you have to specify the switchports as switchports? I've seen a few devices that had ports that were either switchports or router ports but you had to tell it to act as a switchport.

So
int fa0
switchport
switchport mode access
switchport access vlan 1 <- I know its redundant but its a habit.

try that and see if that works

When the Cisco 877W is used and VLAN1 & WLAN interfaces are bridged I see packet loss across the LAN (varying amounts depending on how long the ping runs for).

When a new VLAN is created on the Cisco 877W and two hosts are moved to this VLAN (thus removing them from the bridge-group), no packet loss is seen between the hosts.

When two hosts are connected via a different switch (with no connection to the Cisco 877W) no packet loss is seen.

I have just tried setting all the ports to "switchport" and "switchport access vlan 1" but it doesn't make a difference.

aquilla

I'm beginning to wonder if this maybe a bug in the IOS since pinging across a VLAN works without any issue.

I know that IPv6 routing won't work on BVI interfaces (CSCta27529) in certain IOS's.

I'll see if I can upgrade the IOS over the weekend / next week and see what results that gives.

Panzer919

Good idea, in theory I don.t know of any reason why pings would be dropping. Granted I have never configured a BVI from scratch but from what I was researching on it, there should not be any issues.

aquilla

Panzer919 wrote: »

Good idea, in theory I don.t know of any reason why pings would be dropping. Granted I have never configured a BVI from scratch but from what I was researching on it, there should not be any issues.

Thanks for you help Panzer.

I'll let you know what happens next week re any IOS upgrade.

Panzer919

anytime

aquilla

Hi Panzer,

I managed to grab a new flash module and upgraded the IOS to 15.1 which seems to have cured the problem (and allows me to test IPv6).

Ping statistics for 192.168.1.99:
Packets: Sent = 6189, Received = 6189, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 25ms, Average = 0ms

So it looks like it may have been a possible bug in the IOS (wasn't able to find anything in bug tool kit though).

Regards,

Panzer919

good to know thanks