3

I have a server that I configured using zoneedit. Upon running "traceroute" on it I notice that it takes more than 100 hops.

  • How do I lower the number of hops?
  • Is this really that bad or am I too paranoid?
  • Would the nameserver play a role in the number of hops.

2 Answers2

3

Over 100 hops indicates that there is a problem somewhere, or that you don't have access to a segment of the path.

Do you get * * * or repeated IP addresses?
Try with the -m option to limit the number of hops to a number < 100

  traceroute -m 90 mydomain.tld

thus you should get at least some indications about what is going wrong.

Déjà vu
  • 5,546
  • 9
  • 36
  • 55
  • It is all '* * *'. I had tried with default of 30 and later increased it to 100. Would the nameserver of the server make a difference? We currently use zoneedit and if I wanted to make our server easily findable to lower the latency what would I need to do? –  Sep 23 '10 at 17:13
  • 2
    It seems the *hops* you see are actually blanks `* * *` because there is a firewall or router that prevents the *trace* to go further. Not a real problem. I don't think the DNS is related to that - DNS is used at first to get the IP of the server / domain you want. You shouldn't worry much. – Déjà vu Sep 23 '10 at 23:33
  • 1
    ring0: Correct, DNS is not the issue here. There is a firewall or a router blocking icmp requests along the way. – pauska Sep 24 '10 at 04:57
2

Try using tcptraceroute (or similar) with a destination port of 80 (or any other port you might think be open on the remote server) and see how that goes.

Like ring0 said above, this is not a DNS issue - it's a router or a firewall "out there" who blocks standard traceroute packets (icmp).

pauska
  • 19,620
  • 5
  • 57
  • 75