5 Posts
0
1912
May 10th, 2021 03:00
S4048-ON timeouts on MGMT port
Hi all,
the management port on one of our switches is behaving weirdly - connections frequently timeout. Around 1/5 pings make it through and back with roundtrip times <1ms as expected, the rest is lost somewhere, and there seems to be no regularity. This is observable on other protocols as well, such as ssh or SNMP; you can't get a stable connection via ssh to work with.
On the serial console, everything works as expected. The switch logs a message on the console which might be relevant:
"Mgmt port receiving excessive traffic; will be rate controlled"
We do have a monitoring solution in place which polls all the switches in a 5m interval. However, all the other switches, including a very similar S4048T-ON, don't show this behavior, so I don't believe the message and the actual problem are directly connected.
Any ideas?
Best,
Steve
DELL-Chris H
Moderator
•
9.4K Posts
1
May 10th, 2021 08:00
5557720,
It sounds like a possible loop in the network. Would you verify what else is on that network and if the management port is on a separate management vlan from the other traffic?
Let me know.
5557720
5 Posts
0
May 10th, 2021 12:00
Hi Chris,
thanks for your quick reply. The S4048 is used as a core distribution switch and almost all access switches are connected to this one via fiber. The management port is patched onto one of these other access switches and has no separate VLAN assigned, so that might qualify as a loop. However - a similar S4048T in a similar setup doesn't exhibit the problem, and I wouldn't have expected the management interface to cause a loop in the first place.
Best,
Stefan
DELL-Chris H
Moderator
•
9.4K Posts
0
May 10th, 2021 13:00
Are both switches running the same firmware versions?
5557720
5 Posts
0
May 11th, 2021 01:00
Dell-DylanJ
4 Operator
•
2.9K Posts
1
May 11th, 2021 07:00
Hello,
My troubleshooting is a disjointed from Chris' thought process, apologies for the slight derail. If you haven't already, you might try replacing the network cable connected to the management port. If the other switch is on the same firmware version and running a similar config, then I'd be inclined to think the problem exists elsewhere. My recommendation would be to replace that cable and see if you're able to get a steady and reliable ping out from it. If not, then I'd try running a tracert from a couple of endpoints to the management IP, just to get a good sense of where the failure is occurring.
5557720
5 Posts
1
May 11th, 2021 22:00
I had a chance to look at the cabling yesterday and it seems the main difference was that one switch had it's management port connected to itself (not ideal, of course, but works), and the other one was connected via a separate switch which also had a separate fiber uplink (so indeed a loop, even though I would have thought the MGMT interface would be excempt from that).
So short term solution is to hook up the management port to the same switch, and as a follow up to create a seperate VLAN to isolate MGMT traffic from the rest.
Thanks for your input!
Steve