0

I'm tracing a path to my server using tcptraceroute. The server is showing intermittent network availability. It's as if some packets get through fine, others take a long time and some never get there.

Here are two tcptraceroute's to the server. Any idea what would cause such varying output?

 1  VLAN host (<VLAN IP>)  0.786 ms 1.250 ms  1.637 ms  
 2  * * *  
 3  server host (<serverIP>)  0.632 ms 0.646 ms  0.647 ms

and

1  VLAN host (VLAN IP)  0.848 ms  1.259 ms  1.682 ms
 2  * * *
 3  * * *
 4  * * *
 5  * * *
 6  * * *
 7  * * *
 8  * * *
 9  * * *
10  * * *
11  * * *
12  * * server host (serverIP)  0.615 ms

2 Answers 2

0

Do you know anything about the network between those two IP addresses? I can think of several reasons you might see something like that, but it's only guessing.

If your ISP routes directly to another vendor in the same data center as your IP is hosted, it's a quick route. If that link is down and your ISP routes to the same other vendor via a twisted route across it's own network and those servers do not respond to the trace pings, you get a longer route with no addresses shown. It's possible for a 'brittle' configuration in your ISP to do this same thing if their main DNS server is off line and the backup DNS server is not updated correctly. There are other ways to get this, but knowing nothing about the network between the two addresses means it is all just guessing.

1
  • The problem was actually KeepAlive was set too high on apache under a production server. Apache's sockets were filled causing what looked like network outages.
    – kobrien
    Jun 28, 2011 at 10:57
0

Apache keepalive was set too high. Apache's sockets were filled and was unable to serve more connections making it look like there was network outages.

You must log in to answer this question.

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