Solved

Cyclic timeout on the console port (mgmt / ethernet) - X690-48x-2q-4c

  • 30 April 2020
  • 7 replies
  • 440 views

X690-48x-2q-4c
ExtremeXOS version 22.6.1.4

Exactly every 15 minutes timeout appears on the mgmt (ethernet) port.
No ping response, hangs up the connection.
The switch to which the mgmt port is connected does not report breaks or problems with packets or transmission.

 

Please tell me what else I can check.


# show port mgmt no-refresh
Port Summary
Port     Display              VLAN Name           Port  Link  Speed  Duplex
#        String               (or # VLANs)        State State Actual Actual
========================================================================
mgmt                          Mgmt                E     A     1000   FULL
========================================================================

 

 

# show port mgmt collisions no-refresh
Port Collision Monitor
Port        Link           Collision Histogram
            State  1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16
================================================================================
mgmt        A      0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0

================================================================================

 

 

# show port mgmt information
Port      Flags               Link      ELSM Link Num  Num   Num Jumbo QOS     Load
                              State     /OAM  UPS STP VLAN Proto  Size profile Master
=====================================================================================
mgmt      Em---------fMB---x- active    - / -   0   0    1     1  9216 none    
=====================================================================================

 

# show port mgmt statistics no-refresh wide
Port      Link                Tx Pkt              Tx Byte               Rx Pkt              Rx Byte               Rx Pkt               Rx Pkt               Tx Pkt               Tx Pkt
          State                Count                Count                Count                Count                Bcast                Mcast                Bcast                Mcast
========= ===== ==================== ==================== ==================== ==================== ==================== ==================== ==================== ====================
mgmt      A                150163920          38342141361            155088566          13656681702                    0                   34                    0                    0
========= ===== ==================== ==================== ==================== ==================== ==================== ==================== ==================== ====================

 

 

icon

Best answer by wrepinski 9 May 2020, 09:05

Thanks for the suggestions. Everything checked.

I found a problem, it is a new VM on ESXi with many network interfaces. Unfortunately, I do not administer it. It has been turned off preventively until you check what it can cause. According to the information from the responsible person, the IP used by X690 is not there, but disabling this VM caused the problem to disappear.

View original

7 replies

Userlevel 3

That quite sounds like the upstream switch aging out the MAC address of the EXOS switch’s mgmt interface. To confirm, check the MAC address table timeout of the upstream switch.

To the same switch, the port next is connected to another switch with the mgmt x690 port. There is no problem there. The switch configuration for both ports is identical. I checked, MAC X690 is not lost from the arp / fdb table.
 

Userlevel 4

Hello wrepinski,

Check the following outputs 

* Site-B.4 # run script shell.py 


BusyBox v1.13.4 (2020-01-15 15:18:21 EST) built-in shell (ash)
Enter 'help' for a list of built-in commands.

/exos/bin $ ifconfig eth0
eth0 Link encap:Ethernet HWaddr 00:04:96:52:F6:EC
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1393691 errors:0 dropped:0 overruns:0 frame:0
TX packets:12577 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:10000
RX bytes:136671808 (130.3 MiB) TX bytes:534028 (521.5 KiB)
Interrupt:25 Base address:0xc000

/exos/bin $

Also check for duplicated IPs.

 

Best Regards,

Nikolay

Userlevel 3

A drawing would be nice as always.

From what I understand you essentially have a completely out-of-band mgmt switch (not EXOS) which is connected to several EXOS switches’ oob-mgmt ports. That makes sense.

One of the EXOS switch's mgmt interface stops responding regularly.

The problem is with your out-of-band mgmt network, not with the EXOS switch IMHO. As Nikolay pointed out, check for duplicates.


Switch X690 with IP * .26 always behaves properly.
The X690 switch with IP * .30 loses contact on the eth console (mgmt) port every 15 minutes. I did not find another device with this IP * .30 appeared, I was listening on the router and managed switch * .17.


The ifconfig eth0 command shows drop packages, but now they don't grow in any way, I waited for 4 timeout breaks (1h) and the drop counter did not increase.

 

Userlevel 3

Ok so when you are using Mgmt PC .35 and are working on .30 you are actually disconnected abruptly while working/typing/actually using the console?

If so, can you regularly check the FDB of managed switch .17 (while .30 is still reachable, and when it is no longer)

PS. Nice drawing

Thanks for the suggestions. Everything checked.

I found a problem, it is a new VM on ESXi with many network interfaces. Unfortunately, I do not administer it. It has been turned off preventively until you check what it can cause. According to the information from the responsible person, the IP used by X690 is not there, but disabling this VM caused the problem to disappear.

Reply