Didn't get DHCP IP unexpectedly in 1 area

[Sgj]

Hi...

Wanna know what is the Root Cause issue.

Case:

Suddenly, all devices in 1 area can't get DHCP IP.

Check the switch, it's still on and up.

Tried restarting Switch Access, but it's still not working properly.

When trying to restart the Distribution Switch, it works, and all devices in that 1 area get DHCP IP again.

I forgot to take logs before restarting the Distribution Switch or when the problem occurred. So, I don't know what happened when that problem occurred.

But I get notification error logs ("Flapping") from my customers like this:

This problem only occurs in 1 area, not in other areas.

The Distribution Switch model is C9200 with Version 16.12.01.

Need help with the Root Cause issue, why can it return to normal when restarting the Distribution Switch?

I mean, What is the root cause of the problem that causes flapping to occur?

Maybe someone has experienced a case like this.

Thanks in advanced

[MHM Cisco World]

broadcast storm is issue here, cpu of sw is full of this broadcast and dhcp packet drop.

if this happened again after reboot, check TN of STP.    

[Sgj]

Thank you for the reply...

I know the STP is Spanning Tree Protocol.

Sorry, but what is TN?

[MHM Cisco World]

sorry it typo, i mean TCN

Topolgy change notification,

Are you see any of this tcn before dhcp problem?

[Sgj]

I haven't done that.

Before replying to your message, I tried tracing the TCN of STP and it just appeared like this,

From the picture, to find out the TCN STP, is my method correct?

There's not showing anything.

[Sgj]

Hello...

Refer to your statement "broadcast storm is issue", there's no config or command with the broadcast storm about.

[paul driver]

Hello

When does that flapping occur, is it striaght after your reload the distribution switch or as/when its online?

[Sgj]

Hello...

It happens when it's online.

From the picture, the flapping starts at 5:42:15.

Before that time, the network in that area was normal.

But, when around time 5:42, all devices in that area can't get DHCP IP.

Until reboot Distribution Switch, the network in that area start normally and all device get DHCP IP again.

[paul driver]

Hello
So then you need to probably start looking for physical interfaces that are faulty or are part of a port-channel that would cause instability in the network and create such a large amount of flapping due to a possible stp loop in the network and the reason why it "fixes" when you reload the switch is you’ve temporally broken the loop with the reload.

show int gig1/0/19 < chec interface errors-resets etc..
show int gig1/0/22
show int gig1/0/23

show spanning-tree interface gig1/0/xx detail < check for bpdu's being received intermittently on a port as such that port could then may be deemed as being unreachable thus stp alternate/backup ports would then be opened and closed if the physical link is flapping

show spanning-tree vlan 3601 detail | in Num  < check for large amounts of transistions

Also suggest applying UDLD (aggressive mode), Loopguard for the L2 interconnects and disabling any Dynamic trucking (DTP) that you may have enabled.

Plus setting all access-ports to an administrative mode of access and applying Bpduguard and stp portfast.

[Sgj]

Hello,

I attach 3 files of show int gigx/x/xx, show spanning-tree interface gigx/x/xx detail, and show spanning-tree vlan 3601 detail | in Num.

Are you willing to guide me in your statement?

Also, suggest applying UDLD (aggressive mode), Loopguard for the L2 interconnects, and disabling any Dynamic trucking (DTP) that you may have enabled.

Plus setting all access-ports to an administrative mode of access and applying Bpduguard and STP portfast.

Thank you...

[MHM Cisco World]

Sorry for late reply, I have some business.
I see your attach, 
VLAN3602 <- 
this VLAN have no root STP port??
can I know the topology ? if this access SW is connect to Core "which is Root" and in same time connect to other SW with have priority better then it will elect as Root so are I am in right direction ? 

[Scott Leport]

What are connected to these switch ports? If it’s AP’s, this kind of message is quite common if you’ve got users roaming between locations. If that’s not the case then maybe you do have a broadcast storm as mentioned already. 

[Sgj]

Thanks for the replay

Refer to the picture, port 19, 21, 22, and 23, it was connected to Access Switch, not AP.