Catalyst 4500 listening on TCP 6154 on all interfaces

Mon May 7 20:02:12 UTC 2018

Cisco contact me off-line and ask me to share my datas. They will open a
bug id and investigate.
Nothing to say, very pro active.

The bug id is CSCvj35885

Cisco also confirmed that this tcp port is for internal communication
(internal to the device) only and should not be exposed.

Next time I will follow your recommendation about opening a tac case for
information request, and not bother the community.

Thank to all for your tips and ideas.

Best regards,
Fred

On 07.05.2018 21:22, Spaans, Joel H wrote:
> This has not been my experience. TAC specifically has an option when opening a case to "Ask a question". It's purpose is for non-outage queries such as these. I've asked them things such as "How many ARP entries does an ASA 5585X support?" Sometimes I find conflicting information so I need to ask TAC or I'm just too busy to find the answer. 
>
> I've learned not to be hesitant to engage them, we pay for the support after all. 
>
> Yes, sometimes you will get an engineer who is not helpful. Let them close the case and open another case or insist that the case be moved to another engineer. 
>
> -----Original Message-----
> From: NANOG <nanog-bounces at nanog.org> On Behalf Of frederic.jutzet at sig-telecom.net
> Sent: Monday, May 7, 2018 10:45 AM
> To: Jay Farrell <jayfar at jayfar.com>; nanog at nanog.org
> Subject: Re: Catalyst 4500 listening on TCP 6154 on all interfaces
>
> I've been told that the TAC center will not take the time to answer as it's not a 'real' problem, service affecting issue.
> And the Cisco community forum on that topic was useless (nobody answer to a person which already open a topic about this issue 10 months ago).
> But you are the 4rd person to tell me to open a TAC, I could have tried first.
> In the meantime Cisco contact me off-list, so I will try with them.
>
>
>
>
> On 07.05.2018 16:59, Jay Farrell via NANOG wrote:
>> Just a wild thought – why not open a TAC case with Cisco and ask them?
>>
>> On Mon, May 7, 2018 at 3:06 AM, frederic.jutzet at sig-telecom.net < 
>> frederic.jutzet at sig-telecom.net> wrote:
>>
>>>> - a nsa backdoor :-)
>>> it would be a very bad backdoor as it's really easy to see the port 
>>> listening...
>>>
>>>
>>>> - a default active service
>>> Maybe, but a service which is not officially registered:
>>> https://www.iana.org/assignments/service-names-port-numbers/service-n
>>> ames-
>>> port-numbers.xhtml?search=6154
>>>
>>> in contrary to the SMI (zero touch feature on tcp 4786) which is 
>>> registered since almost 10y:
>>> https://www.iana.org/assignments/service-names-port-numbers/service-n
>>> ames-
>>> port-numbers.xhtml?search=4786
>>>
>>>
>>>
>>> Could it be possible that this kind of tcp port is not registered by 
>>> Iana because it meant to be used for internal communication only 
>>> (internal to the device), or should you register any port usage (even
>>> 'private') ?
>>>
>>>
>>> And yes I've tried to reset to default the config, shutdown all 
>>> interface, remove all L3 ip/feature (no ip blabla), and I still see 
>>> by default 2 TCP ports on listening state:
>>>
>>> Cat4500-SUP7L-E#sh ip prot
>>> *** IP Routing is NSF aware ***
>>>
>>> Cat4500-SUP7L-E#
>>> Cat4500-SUP7L-E#sh run | in ip
>>>  address-family ipv4
>>>  address-family ipv6
>>> no ip routing
>>> ip vrf Liin-vrf
>>> no ip mfib
>>> no ip bootp server
>>> no ip dhcp-client broadcast-flag
>>> no ip igmp snooping
>>> no ipv6 traffic interface-statistics
>>>  no ip address
>>>  no ip route-cache
>>>  no ip address
>>>  no ip route-cache
>>> no ip forward-protocol nd
>>> no ip http server
>>> no ip http secure-server
>>> Cat4500-SUP7L-E#
>>> Cat4500-SUP7L-E#
>>> Cat4500-SUP7L-E#show tcp br all
>>> TCB       Local Address               Foreign Address             (state)
>>> 5B40BB30  0.0.0.0.4786               *.*                         LISTEN
>>> 5CD5D2D8  0.0.0.0.6154               *.*                         LISTEN
>>> Cat4500-SUP7L-E#
>>>
>>>
>>>
>>> I will now try to negate all potential active service from the 'show 
>>> run all' config but it's not optimal as for example 'vstack' (port 
>>> 4786) does not appear in the default config so it would not be 
>>> disable by this trivial method.
>>>
>>>
>>> Fred
>>>
>>>
>>> On 05.05.2018 13:22, marcel.duregards at yahoo.fr wrote:
>>>> As the zero touch feature is on TCP 4786 (SMI), I vote for either:
>>>>
>>>> - a nsa backdoor :-)
>>>> - a default active service
>>>>
>>>> Have you tried to zeroize the config and restart then check if TCP 
>>>> 6154 is still on LISTEN state ?
>>>>
>>>>
>>>> -
>>>> Marcel
>>>>
>>>>
>>>>
>>>> On 03.05.2018 06:51, frederic.jutzet at sig-telecom.net wrote:
>>>>> Hi,
>>>>>
>>>>> We have Cat 4500 series on SUP7L-E with IOS/XE 03.06.02.E/152(2).E2 
>>>>> which have TCP port 6154 listening on all interfaces.
>>>>>
>>>>> Any idea what it could be ?
>>>>>
>>>>> #show tcp brief all
>>>>> TCB       Local Address               Foreign Address
>>>  (state)
>>>>> ...
>>>>> 5A529430  0.0.0.0.6154        <<<<<<<<<<<<<<<<
>>>>>
>>>>>
>>>>> #show tcp tcb 5A529430
>>>>> Connection state is LISTEN, I/O status: 1, unread input bytes: 0 
>>>>> Connection is ECN Disabled, Mininum incoming TTL 0, Outgoing TTL 
>>>>> 255 Local host: 0.0.0.0, Local port: 6154 Foreign host: UNKNOWN, 
>>>>> Foreign port: 0 Connection tableid (VRF): 1 Maximum output segment 
>>>>> queue size: 50
>>>>>
>>>>> Enqueued packets for retransmit: 0, input: 0  mis-ordered: 0 (0 
>>>>> bytes)
>>>>>
>>>>> Event Timers (current time is 0xF58354):
>>>>> Timer          Starts    Wakeups            Next
>>>>> Retrans             0          0             0x0
>>>>> TimeWait            0          0             0x0
>>>>> AckHold             0          0             0x0
>>>>> SendWnd             0          0             0x0
>>>>> KeepAlive           0          0             0x0
>>>>> GiveUp              0          0             0x0
>>>>> PmtuAger            0          0             0x0
>>>>> DeadWait            0          0             0x0
>>>>> Linger              0          0             0x0
>>>>> ProcessQ            0          0             0x0
>>>>>
>>>>> iss:          0  snduna:          0  sndnxt:          0
>>>>> irs:          0  rcvnxt:          0
>>>>>
>>>>> sndwnd:      0  scale:      0  maxrcvwnd:   4128
>>>>> rcvwnd:   4128  scale:      0  delrcvwnd:      0
>>>>>
>>>>> SRTT: 0 ms, RTTO: 2000 ms, RTV: 2000 ms, KRTT: 0 ms
>>>>> minRTT: 60000 ms, maxRTT: 0 ms, ACK hold: 200 ms
>>>>> uptime: 0 ms, Sent idletime: 0 ms, Receive idletime: 0 ms Status 
>>>>> Flags: gen tcbs Option Flags: VRF id set, keepalive running, nagle, 
>>>>> Reuse local address
>>>>>   Retrans timeout
>>>>> IP Precedence value : 0
>>>>>
>>>>> Datagrams (max data segment is 516 bytes):
>>>>> Rcvd: 0 (out of order: 0), with data: 0, total data bytes: 0
>>>>> Sent: 0 (retransmit: 0, fastretransmit: 0, partialack: 0, Second
>>>>> Congestion: 0), with data: 0, total data bytes: 0
>>>>>
>>>>>  Packets received in fast path: 0, fast processed: 0, slow path: 0  
>>>>> fast lock acquisition failures: 0, slow path: 0
>>>>> TCP Semaphore      0x5BEB9B10  FREE
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> (The command "show control-plane host open-ports" is not available 
>>>>> on this platform/code)
>>>>>
>>>>>
>>>>>
>>>>> I also think that if it would be a local socket for internal 
>>>>> process communication, it would be 127.0.0.1:6154 instead of 0.0.0.0:6154.
>>>>> So this is listening on all interfaces, virtuals and physicals and 
>>>>> seam not to be for internal internal process communication.
>>>>>
>>>>>
>>>>> Fred
>>>>>
>
>