ā01-24-2024 07:23 AM
Hi All,
I am finally getting around to fixing the networking that has been hacked together over the years at my current work. There is issues with .253 and .251 switches where STP blocks on some ports causing traffic to grind to a halt.
Some switches are being replaced due to age and reliability issues. But this is the perfect time to also fix the mess of switches that have been added over the years.
Here is a diagram of what I have looked at is going on currently:
To me it doesn't make sense that .252 and .253 are not connected with a link. As well as .252 to .254 since they all sit on the same rack inches from each other.
Currently looking at the root bridge information, it appears .79 and .254 also think they are both root bridges. I realize that the way this network has been setup originally was also with no STP configuration, just leaving the switches to figure it out themselves via Mac and priority. All priority across the switches is set to 32768...
With that said. I'm looking to replace two switches at the moment, .253 definitely, as its STP blocking on its main GE50 port causing disruption to almost the whole network.
I'm looking for advice on how to redesign this with decent load balancing. These are all sg-300's of either 48 or 24 port design. Here is the root bridge and cost breakdowns for each switch:
.254 - last 4 characters of MAC: CB:6F, Root bridge ID: 0, root port: 0, root path cost: 0 priority 32768
.253 - last 4 characters of MAC: CF:AF, Root bridge ID: C0:A1, root port: GE52, root path cost: 40000, priority 32768
.252 - last 4 characters of MAC: C6:03, Root bridge ID: C0:A1, root port: GE52, root path cost: 40000, priority 32768
.251 - last 4 characters of MAC: E8:6F, Root bridge ID: C0:A1, root port: GE27, root path cost: 20000, priority 32768
.79 - last 4 characters of MAC: C0:A1, Root bridge ID: 0, root port: 0, root path cost: 0, priority 32768
I am getting two new CBS350-48FP-4X and would like to best utilize them until I can eventually get the rest of the network switches upgraded as well.
There are a couple of VLAN's being used here and I have realized looking through .79 and .251 there is a mismatch on trunk and access settings on the ports linking them. All advice is appreciated!
Thank you for your help!
ā01-24-2024 07:43 AM
I follow the path there is no loop
only one point
one port config as access and other side config as trunk this can cause issue, I think
so first step is show spanning tree
you share the root but you dont specify for which vlan this root
can you more elaborate
MHM
ā01-24-2024 08:03 AM
Hi MHM,
My apologies on missing those details.
It is GE52 set to Trunk VLAN 1UP on .79
GE27 set to Access VLAN 1UP on .251
The problem I am is GE50 on .253 is STP blocking as well as GE25 and GE26 are STP blocking on .251.
The switches that have workstations attached as endpoints are on .252 and .253. When these ports go blocking, I have issues with the almost the entire networking experiencing loss of connectivity for around a minute, which seems like the STP Topology looking for changes. I notice some of the switches noting "last topology change" at the time frame of when the network does this too.
ā01-24-2024 08:45 AM
since 252 and 253 is not direct connect then there is no issue
the issue in 251
ā01-24-2024 08:43 AM
Hello,
the design looks...well, messy. Do all these switches have the same functionality (that is, are these all access switches), and are they all of the same type ? Typically, the root bridge should be the most powerful switch, and it should be placed somewhere in the middle. Depending on what is physically possible, and on how many free and available ports you have, the best design would be to have (ideally redundant/port-channel) links between all switches, and to run STP.
Looking at what you have right now, the topology needs a major redesign.
ā01-24-2024 09:20 AM
Hi Georg,
Thanks for your reply. I am under the same belief I need a larger redesign. I appreciate MHM's notes on the trunk and access on .79 and .251 being a problem, which I will rectify.
Although I am thinking I want to make both of my new switches my root and secondary and place them where .252 and .253 currently reside. .254 is the gateway and I understand that is usually a core place to put a root bridge. Although .252 and .253 currently have most workstations and small endpoints attached to them directly.
So I am looking as a possible plan of replacing .252 and .253 with my new most powerful switches. Setting their priorities to 4096 and 8192 respectively and letting RSTP figure out the rest. I will upload a proposed new network map and am hoping for feedback on whether it will work better?
Thanks!
ā01-24-2024 10:32 AM
Hello,
sounds like a plan. By the way, you can also set the root and secondary root with the commands:
spanning-tree vlan vlan-id root primary
spanning-tree vlan vlan-id root secondary
ā01-24-2024 09:57 AM
I am thinking of something like this as a redesign. Thoughts on how this might work out?
Just noting .252 and .253 I would be setting as priority 4096, 8192 respectively and they have workstations on their ports that run to different ends of the building.
I think creating more direct routes to the gateway and between core switches would be better?
ā01-24-2024 10:36 AM
If you making both side trunk I think your issue will solved.
MHM
ā01-24-2024 12:37 PM
I agree MHM, I think the TRunk and Access mismatch from .79 to .251 was definitely part of the problem. I'm also looking at creating a more robust network that seems more logical. Do you think the updated planned map will help achieve that?
ā01-24-2024 01:10 PM
I WILL share some more point before we decide what is next best step
I will share you more info tonight
Thanks alot
MHM
ā01-25-2024 08:42 AM
Would anyone see issues with the latest map I've come up with? I have the new switches and I am starting to configure the ports to mirror that of the two switches I'll be removing. Once done that I am planning on making their priorities 4096 and 8192 respectively on .253 and .252 switches. They will be different IP's from .252 and .253 as I am planning on switching over with them in place. But I don't think that should matter.
ā01-25-2024 08:44 AM
Hi friend
Which one is CBS SW?
Thanks
MHM
ā01-25-2024 09:42 AM
The two CBS-350's will be where .252 and .253 are located. I'm going to be adding those extra links between them and .254
The new CBS switches are going to have different IP's at least to start, because I'm still actively running traffic on the old .252 and .253 sg-300's im replacing.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide