Connectivity problems with SGE2000 in ring topology

[nic_risc90]

Hi,

I´ve got a strange problem with the connectivity within a ring topology.

Please view the attached screenshot for a short overview of the scenario.

There are five switches from the SGE2000 series and one SG200-50.

The SG200-50 is the root bridge and RSTP is configured.

From the root bridge switch (Switch 5) I can ping Switch 0.

But I cannot ping Switch 0 from PC0.

On the other hand I am able to ping all other Switches (1-5) from PC0.

In addition to that I can establish a RDP connection from PC0 to Server0.

This tells me that there is no general problem with the connection

from PC0 to the Switch0.

Right now when I want to configure Switch0 the only way is to connect to Server0 first.

But we have to move Server0 and this is why I want to manage Switch0 from PC0 like

any other Switch on that network.

Hope someone has a clue because I really don´t know what could cause this.

Regards

Richard

[Tom Watts]

Hi Nic,

The SG200-50 is the root bridge and RSTP is configured.

By default the SGE switches are CST. Did you update the spanning tree properties for RSTP and increment the bridge priorities for a desirable spanning-tree topology?

From the root bridge switch (Switch 5) I can ping Switch 0.

But I cannot ping Switch 0 from PC0.

From any other switch, are you able to manage switch 0?

On the other hand I am able to ping all other Switches (1-5) from PC0.

In addition to that I can establish a RDP connection from PC0 to Server0.

This is telling us you are on the same vlan and same subnet for ip communication for other hosts (not necessarily switch management)

Right now when I want to configure Switch0 the only way is to connect to Server0 first.

What is special about Server0? Does it have dual NIC? Does it have console access? Does it have multiple IP assigned to the NIC? Can try make a PC1, PC2, PC3, etc? Can you move PC0 to the same port as Server0 then test connectivity? Are you able to manage the switch at all or just looking for ICMP replies?

-Tom
Please rate helpful posts

[nic_risc90]

Hi Tom,

thanks for the quick answer.

Thomas Watts schrieb:
By default the SGE switches are CST. Did you update the spanning tree properties for RSTP and increment the bridge priorities for a desirable spanning-tree topology?
RSTP is configured with a priority of 28672 for the root bnidge and a priority of 32768 for the other switches. Switch0 is aware that the Switch with priority 28672 (Switch5) is the Root Bridge.
RSTP is configured on all switches
From any other switch, are you able to manage switch 0? 
I am using http to manage the switches, therefore I can´t tell you.
This is telling us you are on the same vlan and same subnet for ip communication for other hosts (not necessarily switch management)
All switches are using Vlan1 for the management. I use http to manage them. I can manage all switches except from Switch0 from computer PC0
What is special about Server0? Does it have dual NIC? Does it have console access? Does it have multiple IP assigned to the NIC? Can try make a PC1, PC2, PC3, etc? Can you move PC0 to the same port as Server0 then test connectivity? Are you able to manage the switch at all or just looking for ICMP replies?
Server0 has only one NIC and is directly connected to the Swtich0. He is in the same subnet like PC0 and all the switches.
 From PC0 I can ping Server0 and all Switches except Switch0. Server0 is the only physical server that is connected to Switch0 that I can get access to. I can´t connect another device to Switch0.
I am able to manage the switch when I first connect to that server. This is the actual problem that I can´t get http access to that one switch from PC0 while on the other hand I can access http on all other devices.
 It is not about pinging. I want to manage the switch with http from PC0.
-Tom 
Please rate helpful posts

Regards Richard

[Tom Watts]

Hi Nic, that's a bit strange. Are you connecting anything to ports 12 or 24? These are stacking switches, by default they are in stack mode and those ports are invalid to connectivity.

If your port connections are good, is it possible to just reset the switch to default setting and isolate it from the network to verify working order?

-Tom
Please rate helpful posts

[nic_risc90]

Hi Tom,

it is in standalone mode and some ports between 12 and 24 are connected and working.

Isolating it is not that easy because it plays a main role in this ring topology.

Any other idea what we should check before we reset it?

Regards Richard