cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
425
Views
0
Helpful
6
Replies

Please Help.2950s,2900XLs

vitaliyr
Level 1
Level 1

Good Day

Our company has a problem. Would you be so kind to help us.

Here is a short description of our problem

Our clients have a network with the topology that described below. Two commutators 2900XL are connected by FEC and it is used 2/1-2/4(100FX). Commutators have IP-10.1.1.2 and IP-10.1.1.3 correspondingly. To each switchs (2900XL) is connected Cat2950s using by trunk connection (1/1-1/4100FX).The switches 2950 (10.1.1.4, 10.1.1.5, 10.1.1.6) are connected to the first switch with ip 10.1.1.2(2900XL) and the switches 2950(10.1.1.7,10.1.1.8,10.1.1.9) are connected to the second switch with ip 10.1.1.3(2900XL). . All commutators 2950 have 2950 IOS- 12.1(19)EA1a, and all commutators 2900XL have IOS-12.0(5)WC9. The switch 2900XL(10.1.1.2) are connected to the 2621 to achieve routing between VLANs.

There are four problems:

1). Ping from any 2950(10.1.1.4(5,6)) was successful and passed to the 2900XL(10.1.1.2). Ping command from 2950(10.1.1.4(5,6)) did not passed to the other 2950(10.1.1.4(5,6)) and 2900XL(10.1.1.3) which are connected to the 2900XL(10.1.1.2), till then 2900XL has not made Ping by itself to the switch. In other words commutators can make Ping command through 2900XL. After 3 or 5 minutes situation comes back.

2). All uplinks 2950XL have ignored frames. Broadcasting frames about 70-80 %.

3). At one of the 2950(10.1.1.7) big,very big , large and huge buffers have failures

4). On the both 2900XL at the loading procces have appeared mistake “LINK-4-ERROR Fast1/1 is experiencing errors”.. When we unplug this interface and make a reload we have the same problem on the next interface.

There are no redundant links in configuration. Speed and duplex on the trunk connection at the both side are tunned up in a similar manner. FEC is tunned up correctly. The type of topology is a star. The protocol CDP is switched off.

Please give your advise what is the problem? If this is the mistake of the equipment then what switch is wrong?

Whait to hear from you soon.

Thank you in advance.

6 Replies 6

carenas123
Level 5
Level 5

Answers according to your questions,

1. For this problem, I think 2900xl(10.1.1.2) is not forwarding first ARP packets of 2950 switches before ping packets to all ports i.e to connected 2900xl. Whenever, 2900xl pings itself to the other 2900xl and it's connected switches, then it's getting ARP reply and storing it in it's MAC address table. So that, as you said, after 3-5 min, that entry is getting aged out.

2. For this problem, a lot of choices, bad cable and security issues. Another option to check with "no keepalive" command on 2950 uplinks.

3. It might be due to broadcast storms.

4. Due to IOS image, try to upgrade to another IOS or hardware fault

sorry,topology ---

2950(10.1.1.5) 2950(10.1.1.8)

\ |

10.1.1.1- 2900XL(10.1.1.2)-2900XL(10.1.1.11)

||||

2900xl(10.1.1.4)-10.1.1.6

/ | \

10.1.1.3 10.1.1.7 10.1.1.9

||

10.1.1.10

10.1.1.1-router

10.1.1.2(4,11)-2900XL.

Other -2950

2). All uplinks 2950 have ignored frames. Broadcasting frames about 70-80 %. 10.1.1.11-has not ignoreds

3). At one of the 2950(10.1.1.3) big,very big , large and huge buffers have failures

4). On the both 2900XL at the loading procces have appeared mistake “LINK-4-ERROR Fast1/1 is experiencing errors”.. When we unplug this interface and make a reload we have the same problem on the next interface.

STP is turn off.

sorry,topology ---

....2950(10.1.1.5)...2950(10.1.1.8)

...................\...............|

....................\..............|

10.1.1.1-----2900XL(10.1.1.2)----2900XL(10.1.1.11)

.......................||||

................2900xl(10.1.1.4)-10.1.1.6

................../........|.......\

.10.1.1.3....10.1.1.7...10.1.1.9

.............................||

..........................10.1.1.10

10.1.1.1-router

10.1.1.2(4,11)-2900XL.

Other -2950

2). All uplinks 2950 have ignored frames and broadcasting frames about 70-80 %. 10.1.1.11-has not ignoreds

3). At one of the 2950(10.1.1.3) big,very big , large and huge buffers have failures

4). On the both 2900XL at the loading procces have appeared mistake “LINK-4-ERROR Fast1/1 is experiencing errors”.. When we unplug this interface and make a reload we have the same problem on the next interface.

STP is turn off.

I have observed the problem and payed attention to the following thing. Such situation with ping happened only with VLAN 1 is not regular. Sometimes everything works ok. Maybe it's not connected with ARP tables?

In "show process cpu" on 2900XL (10.1.1.2,10.1.1.4) - BROADCAST STORM - 4%. How can it be (FEC has adjusted correct, STP is turn off(topology star)?)

Bad cables on all uplinks?

Device fault manager (CW2000) dose not show presence of any physical problems with switches.

The IOS version is the latest.

Hi,

I'd recommend to enable STP.

There might be a bug concerning FEC with STP off.

Generally it's recommended not to stop STP unless absolutely necessary.

Regards,

Milan

Hi,Milan. When STP was turned on the situation was the same. That's why I don't see any difference.

It's interested that on uplink of 10.1.1.11(2900XL) have not ignored(and broadcast~1%) frames in difference 2950s