Problem with switch SG500 of cisco

[tu.bn]

We are using switch of Cisco belong to series SG500. The current, we saw some of the problems after use the switch (detail below). Please support for us.

Here is a desciption of the problem
-When the switch is started, it runs very normally. The devices are connected to the switch, the switch will learn the MAC address. The number of MAC address learned on the switch about 2000 MAC.
-After switch operate about 2-3 weeks, we found that the number of MAC address learned on the switch reduced more than 80 percent, from 2000 MAC to around 300 MAC. The devices are being connected on the switch still operate normally. However, when we move the network cable of device is using from one port on a switch to another port, we were experiencing long delays in reconnecting to the network. The delays are about 10 minutes.

Detail test after switch operate about 2-3 weeks:
- First device configured with a static IP of 192.168.1.2 is plugged into port 2 of a cisco SG500 switch.
- Second device is configured with a static IP of 192.168.1.3 and is plugged into port 3 of the switch.
- On both device, We start a ping <IP of other PC> -t
- Then we moves the network cable from port 3 to port 4 of the switch. As soon as we moves the cable, pings on both device stop working, check the switch is not learned MAC of second device. After about 10 minutes, the pings start working again. But the switch is still not learned MAC of second device.

Here is detailed information of switch that we are using:
SG500-28 28-Port Gigabit Stackable Managed Switch
- PID VID: SG5000-28-k9 V02
- Firmware Version (Active Image): 1.3.0.62
- Stack unit mode: Standalone
- System Operational Mode: L2 Mode
- Jumbo Frames: Enabled
- Spanning tree protocol: MSTP
- SMNP : Enabled
- Port security : Disable
- Qos: Disable
- Access control : Disable
- DHCP : Disable

Looking forward hearing from you soon.

 

[imagoon]

I would open a case with TAC. 2000 MAC addresses means 2000 devices which means these are part of a very large or several layer 2 networks. Without knowing your environment and what the other switches are doing, no one here can really make a suggestion. For example I see you have jumbo frames enabled. Unless these are SAN switches that should be off but we don't know that because you don't state your use case. MSTP being on indicates the presence of vlans etc.

2000 MAC addresses could also indicated you have network loop some place also, and MSTP has started blocking ports / paths / vlans etc. Or spanning-tree is having trouble converging... or any number of issues.

 

[lif_andi]

The default aging timer is usually around 5 minutes. If this is about the delay you are getting then this is normal. You can change the default aging timer to something else, although this is not recommended. There might be a different aging timer either default or configured on the switch that increases this to 10 minutes. Or as it says in the documentation:

The aging time is a value between the user-configured value and twice that value minus 1. For example, if you entered 300 seconds, the aging
time is between 300 and 599 seconds.

The aging timer is automatically lowered if the switch detects a topology change, and if it sees a MAC address come in on a port where it was not expected. You can lower this timer yourself for testing, but if you want to leave it to its changed value be sure that you understand what it is you are changing and what the consequences are. For testing purposes you can clear the MAC address table by clicking "Clear Table". This should repopulate the table and you can see what traffic is going through, and also you can click this right after you switch the cables and see if this does not fix your problem.

Regarding the drop in MAC addresses, if the traffic from those devices is not going through this exact switch then this is normal. When you first connect the switch, and your topology is trying to converge, there would assumably be a lot of broadcast traffic, and unicast flooding, so the switch will learn a lot of addresses. After a while, when the dust has settled, the MAC addresses of the devices that are not using that switch to carry traffic, those MAC addresses should disappear from the table. If you are experiencing loss of connectivity though, I'd advise looking at the documentation for the switch, and if no solution is found, you should open a TAC case or post a question on the Cisco website.

On a side note, if everything is working normally, why are you moving the cable at all, and what exactly is the problem you are having ?

 

[tu.bn]

The default aging timer is usually around 5 minutes. If this is about the delay you are getting then this is normal. You can change the default aging timer to something else, although this is not recommended. There might be a different aging timer either default or configured on the switch that increases this to 10 minutes. Or as it says in the documentation:

The aging timer is automatically lowered if the switch detects a topology change, and if it sees a MAC address come in on a port where it was not expected. You can lower this timer yourself for testing, but if you want to leave it to its changed value be sure that you understand what it is you are changing and what the consequences are. For testing purposes you can clear the MAC address table by clicking "Clear Table". This should repopulate the table and you can see what traffic is going through, and also you can click this right after you switch the cables and see if this does not fix your problem.

Regarding the drop in MAC addresses, if the traffic from those devices is not going through this exact switch then this is normal. When you first connect the switch, and your topology is trying to converge, there would assumably be a lot of broadcast traffic, and unicast flooding, so the switch will learn a lot of addresses. After a while, when the dust has settled, the MAC addresses of the devices that are not using that switch to carry traffic, those MAC addresses should disappear from the table. If you are experiencing loss of connectivity though, I'd advise looking at the documentation for the switch, and if no solution is found, you should open a TAC case or post a question on the Cisco website.

On a side note, if everything is working normally, why are you moving the cable at all, and what exactly is the problem you are having ?

Our switch aging timer value is 300 seconds. Before we used switch sg300 is very good, after replace by sg500 happens problem. Switch sg500 is configured as switch sg300.

I don't think aging timer value is problem.