3

We recently moved offices, and the only changes I am aware of we have made to the network is to move the servers and desktops onto a new IP range 172.50.0.*/23. However, we are now experiencing intermitten and very odd communication between machines.

We have 2 Active Directory domains, and all the servers on one of the AD domains can be pinged OK, and they can all talk to each other fine.

However, on the second domain, we are experiencing intermittent and strange ping behaviours. What we are seeing is as follows:

  • Some servers simply don't respond to pings from other servers, but will respond to pings from other machines on the same subnet.

  • Some servers respond intermittently to pings from other servers, whilst at the same time responding reliably to other servers / workstations.

  • Some servers respond only once to a ping, and then never respond again such as:

    ping mailfilter.intranet.domain.com Pinging mailfilter.intranet.domain.com [172.50.0.240] with 32 bytes of data: Reply from 172.50.0.240: bytes=32 time<1ms TTL=128 Request timed out. Request timed out. Request timed out.

  • Sometimes I'm seeing slow response times, such as 800ms, which is certainly not normal considering both servers are connected to the same GB switch.

What we have tried is the following:

  • Reconfigure the switches back to Factory defaults

  • Replace all switches with different switches

  • Add devices onto the network one at a time until the issue returns, and there is no pattern. The servers respond to pings fine for a little while, and then after some time the issue returns

  • Moving servers from one domain to another, does not appear to resolve the issue for now, although it's only been a few hours...

I really don't know where to turn next. Is it possible there are some legacy IP based rules hidden somewhere which could be causing this?

3
  • 1
    How many switches are in use? Have you checked for a switch loop?
    – joeqwerty
    Apr 15, 2011 at 14:32
  • 1
    Just an observation from the ping command output in your question. I think you should be pinging using IP addresses at this time just to get DNS problems out of the equation and focus on IP addressing. Apr 15, 2011 at 16:42
  • 1
    I think Joe is on track, are you certain there isn't a loop anywhere on the network? if switching is sound, and subnet mask is the same, you are pinging port to port, it should work, there is nothing else to it. Jul 5, 2011 at 2:30

1 Answer 1

1

Duplicate IP's on the network?

Circular route somewhere?

Bad network card?

4
  • Well Windows is not registering a duplicate IP when it starts up, and I know it normally does if there is, so I would assume not. Apr 15, 2011 at 14:12
  • Sorry, on the circular route stuff, I don't think so, I unpatched every device and added them back one at a time. I got to around 7 devices before the problems returned and there was very little connected. Bad network card is not possible as we have changed the NIC in one of the servers and still have the same problem. Apr 15, 2011 at 14:13
  • Never "assume" windows does or doesn't do anything, and duplicate IP detection is tricky and can't always be done in a few seconds. Also, the bad network card doesn't have to be the server, it can be any card or port on the network. When you added devices back in, did you catch any correlation between a device and when the problems started again?
    – Caleb
    Apr 15, 2011 at 14:42
  • Nope, no correlation whatsoever. I even unplugged the offending device and the problems persisted... Really struggling to figure out where to look next. Apr 15, 2011 at 15:27

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .