Cisco Switching/Routing :: 3750 Stack To Upstream ISP Separate Switches - MAC Flapping
Mar 11, 2012
I have an issue with a Cisco 3750 switch stack which is connected to two seperate upstream Cisco 3750's which are administered by an ISP. The ISP is experiencing MAC address flapping from one of my VLAN SVI's i am using to route traffic upstream
As you can see I utilise a VLAN SVI to route traffic to the upstream 126.96.36.199 (illustration only) IP. As per the diagram of the topology attached, the ISP is receiving a MAC address flapping error confirming the SVI MAC address from my switch stack is being learnt on the trunk port connecting switch 1 and switch 2, and also the port (Gi1/0/48) directly connected to my switch stack. As these are all Layer 2 links essentially being passed upstream and then connected between the two ISP switches, we have a 3 way triangular loop formed.
If I was to remove the port channel configuration from the two ports associated with the VLAN SVI, am i right in suggesting this would still form a layer 2 loop? The two ports would still be a member of the SVI VLAN, and it is the VLAN MAC address which is being learnt by the two ISP switches on different interfaces.
May 1 16:00:30: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/12, changed state to down May 1 16:00:39: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/12, changed state to up May 1 16:17:28: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/12, changed state to down May 1 16:17:36: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/12, changed state to up May 1 16:25:28: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/11, changed state to down May 1 16:25:37: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/11, changed state to up May 1 17:19:25: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet0/11, changed state to down
Configuration as below:
macro name vmware switchport trunk encapsulation dot1q switchport mode trunk no switchport access vlan
Note:Both active blade switch and standby one are connecting to same aggregate switch 3750. Yet the active one, with same config as standby one, has not occurred this issue.
Looking for feedback from other organizations that have large 3750 stacks. I've got one stack of (8) 3750's composed of (6) 3750G's and (2) plain 3750's. This particular stack is usually unresponsive to SNMP queries and often fails to write config when we make changes. After a couple tries it will finally go. Part of my probably here is likely the plain 3750's that always boot faster and come up as the master. I should manually set the master to one of the G's. What I'm wondering is who else has 7-9 3750's stacked and are they performing well for SNMP, telnet, etc? I've got another newer stack of 7 3750E's that I need to add one more switch to. Need to decide if I want extended downtime to break the stack up or just add the 3750X to make member 8 and hope it performs well. I have 50+ 3750 stacks working great on our campus.
I am trying to do ios upgrade on 5 stacked 3750 switches. All the switches have different model number, so i am wondering which image file i should download. As far as i understand all the stacked switches should have the same IOS, i may be wrong. The switches have the following model numbers and SW images;
We have several stacks of 3750 switches where this problem is occurring. All of the switches are running IOS version 12.2(50)SE3. (Yes, I know it's old.)
We're seeing inconsistency in how the stack reports members when issuing the following commands:
sow switch show version show inventory
Not all members are showing up in the output of the "show version" and "show inventory" commands. For example:
HUNTI-WV-WDAC03#sh switch H/W Current Switch# Role Mac Address Priority Version State ---------------------------------------------------------- 1 Member 108c.cff2.ea00 15 0 Ready 2 Member 8cb6.4f0a.a600 1 0 Ready *3 Master 8cb6.4f0a.9c00 1 0 Ready (all three switches in stack are displayed)(code)
I am wanting to etherchannel from a 3750 stack to core Layer 3 switches (also 3750) with a cable going to each core switch, I have put both core switches and the stack under a 28bit subnet mask, but I dont seem to be load balancing across both links.
I have 3 3750 staking switch with the following configuration:
H/W Current Switch# Role Mac Address Priority Version State ---------------------------------------------------------- 1 Member 442b.0350.9400 1 1 Ready 2 Member 442b.0357.8780 2 1 Ready 3 Master 442b.036f.a800 3 1 Ready
The new switch is installed to be wired to the master switch. the swtich 3 (* 3 Master Ready 442b.036f.a800)my question is: if you disconnect the wiring from the master stack to connect to the new switch, it is possible that the new switch change the current configuration of stacking.
-Can I just add the new switch with clean configuration without changing the current configuration ? -How is assigned the stack number? i need the new switch be with the GI4 / x?. I can I manually by assigning ports? -If I want to the new 3750 member witch lower prioridad, can I change priority before adding to stack?
Lets say i have 2 3750 switches stacked via backend stack cables. Now if a packet needs to go from 1 switch in the stack to second switch in the stack, will it travel via stack cable or do we need to connect both switches via uplink ports (ethernet or sfp). I tried reading datasheet but it no wheres mention the actual frame path between switches in stack.
I'm thinking about using a 3750 stacking around 4 or 5 switches intead of using one 4500 series switch.
1-My first argument was concerning the budget (one 4500 instead of four or five 3750 in stack mode) ---- Here i think 4 or 5 stacked 3750 is cheap than one 4500 2- My second argument was about harware redundancing that i don't have with one 4500 and that i have with stack mode ----Here i have more hardware redundancing with 3750 3- The bug that we notice today in 2960 and 3750 and not in 4500 4- Simple to manage a 4500 compare to 3750 stack design 5- I was looking also for the one with best backplane ----Here i didn't have any infos concerning the best backplane between 4 or 5 stack 3750 and a 4500. 6- I'm also looking for power redundancing.
We have two switches WS-C2960S-24TS-L running in stack with running IOS c2960s-universalk9-mz.122-55.SE3.We have connected two checkpoint devices on Gi1/0/24 and Gi2/0/24 with a simple configuration:
interface range GigabitEthernet1/0/24, gi2/0/24 switchport access vlan x switchport mode access spanning-tree portfast
On the devices log we get the following message:
%SW_MATM-4-MACFLAP_NOTIF: Host 0000.0000.fe21 in vlan x is flapping between port Gi2/0/24 and port Gi1/0/24
The mac-address 0000.0000.fe21 is not found on any checkpoint interface. "Show mac-address int gi1/0/24 or gi2/0/24" does not show the correct mac-address of the checkpoint interface connected to the port, but instead shows 0000.0000.fe21.Is there a bug in the IOS, something wrong with the switch config or does it have something to do with the checkpoint config?
I configure 3750 stack switch as core and 2960 stack switches as access layer switches.I connected my laptop to one of my core stack in VLAN 10 and I am pinging to one of my server in VLAN 1. What will be the minimum latency at the time of inter VALN routing
I've been fighting what seems to be an increased number of outqueue drops on our core stack and edge switches for the last 3 or 4 weeks.(The core consists of a stack of 5 3750s in 32-gig stack mode. The wkgrp switches are 3560s. all are at 12.2.52) The wkgrp switches are directly connected to users. We use Nortel IP phones with the phone inline with the user PC. auto-neg to 100/full. [code] However I have tried turning off QOS on a couple of workgroup switches (no mls qos, but left individual port configurations the same) but am still seeing drops.Since I have disabled qos on the switches in question (no mls qos) (not the core tho) I am presuming these commands have no affect on the switch operation and therefore cannot be related to the problem. With QOS turned off one would presume that it is general congestion - especially at the user edge where busy PC issues might contribute. So I wanted to see if I could see any instances of packets in the output queues building up.
I wrote some scripts and macros that essentially did a snapshot of 'show int' every 20 seconds or so, and looked for instances of 'Queue: x/' where x was greater than zero.What I found after several days of watching the core stack, and a few of the workgroup switches that are most often displaying the behavior, was that I NEVER saw ANY packets in output queues. I often saw packets in Input queues for VLAN1, once in a great while I would see packets on input queues for fa or Gi interfaces, but NEVER on output queues. [ code] Additionally, when I look (via snmp) at interface utilization on interfaces showing queue drops (both core and wkgroup), they are occurring at ridiculously low utilization levels (as low as 4 to 8%). I've tried to look for microbursts between the core and a wkgroup switch where the core interface was experiencing drops, but haven't seen any (using observer suite). [code] While the queue-drop counts aren't critically high at this point, they are happening more frequently than in the past and I would like to understand what is going on... In most cases, no error counters are incrementing for these interfaces. Is there some mechanism besides congestion that could cause output queue drops?
I recently upgraded my network to have two 3750x core, one interface on the Cisco is connecting to a Net gear switch via a fiber converter. I am keep getting the vlan flapping error message in my log as below.
003396: Sep 17 01:46:16.328: %SW_MATM-4-MACFLAP_NOTIF: Host 5c0e.8ba7.0a5c in vlan 2 is flapping between port Gi2/0/15 and port Gi2/0/13 003397: Sep 17 01:46:19.843: %SW_MATM-4-MACFLAP_NOTIF: Host 5c0e.8ba7.0a5c in vlan 2 is flapping between port Gi2/0/15 and port Gi2/0/13 003400: Sep 17 01:49:58.769: %LINEPROTO-5-UPDOWN: Line protocol on Interface GigabitEthernet2/0/17, changed state to down [Code] .....
After my research i think this is a looping issue but I'm unsure how to address it.
Having a design where an IBM bladecenter has two Nexus 4001i switches and each switch is connected to a Nexus 5k vPC pair, is it possible to configure active-active nic teaming on the hosts (blades) considering that the Nexus 4k switches are unaware of each other and blade has one connection to each 4k over either emulex or qlogic 10G CNA?
As I understand Cisco Catalyst 3750-x supports stackwise and stackpower technology.Do I need to purchase a seperate module to enable stackwise and stackpower? Or are stackwise and stackpower modules included be default on the switch already?
I have two 3750s stacked and one 3750 standalone all in one data closet. The both have the same versions. I want to combine them to make one stack. How do I force the standalone to be the master? sw 1 priority XX(Higher then the existing stack)? Both the stack and the standalone have management addresses. Would it be ok to delete the stack ip address? That way, the standalone(master's address will be the only management address)?How should I connect the standalone 3750 to the stack? At present, the 3750 stack is connected via a cascade cable going from the stack 1 port of the top switch to the stack 2 port of the bottom switch.
The stack cable connected to Switch 1 Port 1 and Switch 4 Port 2 will not come back online. The logs show that there was a Stack line change. I have replaced the 1 meter Stack cable from Switch 1 to Switch 4 three times and it still does not come back online. This is the part that is interesting.. I have disconnected Port 1 Switch 4 and connected it to Port 2 Switch 4 and then Switch 4 came back online. This made me think Port 2 on Switch 4 was working correctly. Then I disconnected Port 2 Switch 1 and connected it to Port 1 Switch 1 and then Switch 1 came back online.
i have two cisco 3750 in stack firts, master switch have all swports busy second have only 5 how can i see cpu utilizations of the second switch, and its reasonable to move somme devices from first switch to second for deacreases of cpu and memory utilisatio ns of the master? the stack tehnologies use both cpu as one? or its only for simple management ?
i want to remove an access layer 3750 48 POE master switch and replace with a new 3750 switch in the stack. i want to copy the same configuration to new switch since the old switch is having POE issue and I got replacement of the malfunctioned switch. we have VTP domain is configured in the network...
I have been given 2 3750 switches and need to put a config onto it wuld is for 2 3750's in a stack, how should I put this on the 2 switches? Is it bested to install it onto 1 switch first then physically attach the 2nd after?
I'm setting up a web cache using the wccp protocol on a Catalyst 3750 stack.
Probably missing something real simple here but when I from the global configuration mode are trying to enter the ip wccp command it just says "invalid input" from wccp. There is no such command.. should be supported on my device from IOS 12.2(37)
We have two 6k5 configured as vss and it is working fine. For each floor we have a 3750 stack. In each 3750 stack we have 4 switches. From the first 3750 we have a fiber link to the first 6k5 and from the fourth switch we have another fiber link to the other 6k5 switch, for each floor/3750 stack, both configured as etherchannel. We installled each 3750 stack/floor in different maintenance windows; in all these installations we have problems configuring the etherchannel. After we configure the etherchannel on 3750 stack and in the vss, the etherchannel does not go up/up. We see the etherchannel up/down; after a lot off work, try and error, configuring shut/ no shut, adding and excluding links from the etherchannel; in any aleatory moment the etherchannel goes up/up and works fine. Is there any bug related with these IOS versions, or is there any right procedure to configure etherchannel from vss to 3750 stack?
I currently have 5 3750s connected via Stackwise. This stack is connected back to my core (6509s) via fiber (1Gig to Core1 & 1Gig to Core2). I want to move the 5th 3750 to another closet due to an addition of a new patch panel. Due to max length of 10 ft. for stackwise, my plan is to connect this 5th switch back to the stack using trunked fiber ports, with them being an etherchannel (2Gig). Also plan to use single fiber port form 2 different switches in the stack to provide some redundancy for the 5th switch. Should I be alert of any issues with this connection design? Do I need to properly remove the 5th switch from the stack using, no sw 5 provison sw-model or can I just disconnect stackwise and connect back with fiber?
I had 2 3750 stacked and one of them decided to die (power supply) cisco replaced it and now i want to re-add that switch to stack. I have already upgraded my replacement switch to the same IOS version.I was reading that i can do it without powering down operational 3750 so the action plan so far is:
-power down NEW member -connect stacking cable -power up
as you can see my switch 2 is now a master and it still holds config of all Gi ports for switch 1. If i connect my NEW switch without any config (vlan.dat and startup-config) will my existing switch 2 copy all ports config to new switch ?
We had an "Event" on our process network at the mill yesterday. I connected a new WS-C3560V2-48TS-S to our network and we lost communications with all of the other switches.
The core is a stack of 3 WS-C3750 switches, one is a -24TS-1U and the other two are -12S-S. These connect to 10 WS-C3560G-24TS and 5 Rockwell Stratix 8000 (IE3000) over fiber. I am planning on replaceing two 24 port switches with the new 48 port. I had the switch configured and running at my work bench. It was connected to the network with one sfp module and ran all weekend with no issues. Yesterday afternoon I took it to the network cabinet and installed it. I powered it up and connected two sfp modules to the fiber patch panel and made the connection at the core stack.
Everything looked OK. I had communication link lights working on everything. Within minutes, we lost communication with every switch connected to the core stack. I shut down the new 48 port switch and the network slowly came back up.
The new 48 port is configured with Flex Links for the fiber redundancy protocol and was connected to each of the 3750G-12S-S stack members.
I though it might have been a power issue but the stack is UPS protected and shows it has been up for over 10 weeks. I'm not even sure "Crash" is the best decription for what happened. The new switch has a high enough IP address that it would not take over as a IGMP Querier.
I've a situation where I need to add a new 3750x to a existing stack of 3750. [code] When I tried to stack them together, I get a version mismatch error. Is this because of the difference in SW Image? What are my options next ? My ultimate goal is to make the new switch stack correctly with the exisitng switches.