Home / server / Questions / 382824
In Process
zentenk
Asked: 2012-04-25 19:12:11 +0800 CST

Where can I start to diagnose why I occasionally cannot connect to a server

  • 772

I have a Dell Poweredge T310 with the default (free) NIC. I've got Ubuntu server 11.10 installed and I am running into a problem where at times I cannot connect to the server through any port. Whether it's an ssh or http request it would just timeout. (Connecting externally)

I used YouGetSignal to check whether the ports are open and they are closed.

However after 5 minutes or so of attempting I would be able to connect again.

I am quite new to networking and I am not sure where I should start looking at for diagnosing.

I don't believe it is a router problem since I can connect to the other devices connected to the network. I also do not have any power management software with the Ubuntu.

ubuntu nic dell-poweredge

3 Answers

  1. Perhaps you may want to try to update the firmware and driver for the nic card. Dell has a RedHat version of the firmware and driver, hopefully you can use it under Ubuntu.

    Do take note that, Ubuntu is not validated in Dell PowerEdge T310.

    Below are the firmware and driver for your Broadcom BCM5716:-

    Firmware http://www.dell.com/support/drivers/us/en/19/DriverDetails/DriverFileFormats?DriverId=R319248&FileId=2731119705&productCode=poweredge-t310&urlProductCode=False

    Driver http://www.dell.com/support/drivers/us/en/19/DriverDetails/DriverFileFormats?DriverId=YG02H&FileId=2731101505&productCode=poweredge-t310&urlProductCode=False

    • 1

  2. You need to look at the server's logs to determine what is going on. Start in /var/log/messages.

    Next, you need to sample traffic on the server to watch how it is seeing network communication at the wire level. Start by using Wireshark.

    After those investigative measures, you can begin the real fun of tracking down the specific trouble. However, before those simple things have been done, there's not much else to go on.

    Remember that the rules of troubleshooting, regardless of the discipline, are always:

    1. Gather Data
    2. Gather Data
    3. Gather Data

    Oh yeah, and something about making a hypothesis and then falsifying it. =)

    • 0

  3. traceroute (when it works and when it does not, and in both directions).

    Check logs for evidence of duplicate mac or ip addresses.

    Check http or ssh logs to see whether inbound or outbound communication is the problem. Run sshd in debug mode (with output redirected to a file, perhaps using tee) so you can see what, if anything, is recorded when you lose connectivity.

    Engage your network admin, if you have one, to check routers, switches and firewalls.

    • 0