PowerConnect 6248 ssh unresponsive

Hello,

I'll help you to find what could be the source of this issue.

When you cannot open the SSH console, is the ping reachable?

Have you tried another SSH console software ?

Is the network bandwith usage or traffic of this switch higher than usual when issue occur.

Have you tried in same time to access to the swich with the serial port.

Is the switch firmware up to date?

I remain at your disposal.

Regards,

Hello there,

The serial console is quite responsive.

ICMP seems to be failing most of the time: 1/6 packets, 83% loss, min/avg/ewma/max = 121.846/121.846/121.846/121.846 ms

I'm not certain how to determine whether the network traffic is high.  I expect most of the traffic is ingressing from the cross connect on port-channel 6, which the 5424 is on, and egressing on interface 1/g33, which attaches to the DOCSIS modem.  See this config for details.

I don't expect more than maybe 100Mbit of traffic to be transiting the switch. But we can investigate this theory using the serial console. I am just not familiar enough with the system to know how to do this yet. I would be grateful for any advice you can offer.

There are also a number of Debian linux systems also attached to the switch via port-channels. These have either 4 or 6 GE ports in LACP bonds. At this point, I don't expect that any of these hosts are sending more than 1Mbit each.

I have not tried with a different ssh client version, but I have tried from different hosts, some directly attached to the switch, and some indirectly attached via a switch attached via a LAG to the 6248.

$ ssh -V
OpenSSH_7.9p1 Debian-10+deb10u2, OpenSSL 1.1.1d  10 Sep 2019

The switch firmware seems to be up to date. I had to update it with xmodem over night a few weeks ago.

switch01#show system
System Description: Dell Ethernet Switch
System Up Time: 2 days, 17h:16m:57s
System Contact:
System Name: switch01
System Location:
Burned In MAC Address: 0025.642A.C55F
System Object ID: 1.3.6.1.4.1.674.10895.3011
System Model ID: PCT6248
Machine Type: PowerConnect 6248
Temperature Sensors:

Unit     Temperature (Celsius)    Status
----     ---------------------    ------
 1        42                      OK

Fans:

Unit     Description    Status
----     -----------    ------
 1       Fan 1          OK
 1       Fan 2          OK
 1       Fan 3          OK


 1       Fan 4          OK

Power Supplies:

Unit     Description    Status         Source
----     -----------    -----------    ------
 1       Main           OK             AC
 1       Secondary      Not present    DC

switch01#show system id


 
  
Serial Number: CN0GP931282989AO0218A10
Asset Tag: none
Unit Service tag       Serial number           Asset tag
---- ------------      --------------          ------------
1    
  
    CN0GP931282989AO0218A10 none switch01#show switch Management Standby Preconfig Plugged-in Switch Code SW Status Status Model ID Model ID Status Version --- ---------- --------- ------------- ------------- ------------- ----------- 1 Mgmt Sw PCT6248 PCT6248 OK 3.3.18.1 
  
 

Are there any network loops? A loop could cause this behavior. Is the management on it its own VLAN?

Hi Josh,

I'd be happy to put management on its own VLAN.  I'm not certain how to change the management VLAN, though.  I looked it up in the manual but 

switch01#show ip interface management

IP Address..................................... 192.168.79.210
Subnet Mask.................................... 255.255.255.0
Default Gateway................................ 0.0.0.0
IPv6 Administrative Mode....................... Enabled
IPv6 Prefix is ................................ FE80::225:64FF:FE2A:C55F/64
IPv6 Gateway................................... FE80::E11:67FF:FE02:4822
Burned In MAC Address.......................... 0025.642A.C55F
Configured IPv4 Protocol....................... None
Configured IPv6 Protocol....................... DHCP
DHCPv6 Client DUID............................. 00:03:00:06:00:25:64:2a:c5:5f
IPv6 AutoConfig Mode........................... Disabled
Management VLAN ID............................. 1
switch01#conf

switch01(config)#ip address vlan 78

switch01(config)#end
switch01#show ip interface management

IP Address..................................... 192.168.79.210
Subnet Mask.................................... 255.255.255.0
Default Gateway................................ 0.0.0.0
IPv6 Administrative Mode....................... Enabled
IPv6 Prefix is ................................ FE80::225:64FF:FE2A:C55F/64
IPv6 Gateway................................... FE80::E11:67FF:FE02:4822
Burned In MAC Address.......................... 0025.642A.C55F
Configured IPv4 Protocol....................... None
Configured IPv6 Protocol....................... DHCP
DHCPv6 Client DUID............................. 00:03:00:06:00:25:64:2a:c5:5f
IPv6 AutoConfig Mode........................... Disabled
Management VLAN ID............................. 78

Oh, look. That worked. I can't ssh to it, though.

cjac@ogion:~$ ssh 192.168.79.210
ssh: connect to host 192.168.79.210 port 22: Connection refused
cjac@ogion:~$ nc -vz 192.168.79.210 22

nc: connect to 192.168.79.210 port 22 (tcp) failed: Connection refused

No, I don't think there are any bridge loops.  How can I diagnose this, though?  Can I see the throughput on a per-port basis, maybe?

actually, yes I can reach it on that VLAN.  I had disabled that VLAN in order to debug something earlier.  

OK.  Management is much more responsive on the management VLAN.  I'd still like to know why ssh and ping are so unresponsive on other VLANs.  If I turn off ip routing on all other interfaces, will that improve things?

Hi,

In my opinion, access via the management VLAN is faster because it avoids the need for inter-VLAN routing. What do you mean by disabling IP routing on all other interfaces?

Regards,

By disabling ip routing on all other interfaces, I meant

switch01#configure

switch01(config)#no ip routing

The ssh connection is now responsive.  I will see if I can figure out how to enable ssh pub key authentication.

Before I close this thread as resolved, can you recommend any tools to detect bridge loops or high network usage on any particular ports?

Thanks,

C.J. 

Thanks, I'll take a look at that.