Actually, this flakiness started after the trial period ran out and continues even after I purchased registered. I have the newest version.<br>Sometimes, Ping Plotter will draw all the hops on the graph but everything will be empty and all them will have 100% packet loss except for the destination. This may go on for a while before I finally get data. I haven't figured out what causes it or what solves it. When it does happen, I've noticed that a tracert still works fine. I am usiong Windows 2000 Server SP2.<br>How can I figure out what the problem is and how to fix it? Is it a bug?<br><br><br>

Most of the time when I've seen this, it has to do with a router not liking to have multiple outstanding ICMP echo requests. It sounds like maybe you have a router that doesn't like quite as many outstanding requests as you have there. The number of outstanding requests can change based on the latency of your hops - timeouts can cause more to be outstanding.<br><br>To test this theory, go to the faq and follow the directions to make Ping Plotter single threaded (it's currently the last entry in the FAQ). This will affect performance, but you can tweak things so it's not *too* bad.<br><br><A HREF="http://www.pingplotter.com/faq.html" target="_new">http://www.pingplotter.com/faq.html</A><br><br>Another possibility is that you have a router that drops timed out echo requests when under load. I have a Cisco 675 here that had this problem - under heavy load, all the intermediate hops would go to 100% packet loss, and even under moderate load, the latencies would jump up to something about 10 times the normal latency. To fix this, I updated to a new version of the bios (CBOS 2.4.x if you have a 675).<br><br>After you've tested this out, check back in here and let us know how it turns out.<br><br>Thanks!<br><br><br>

Hi,<br>This happened again today.<br>I followed the instructions at the end of the FAQ, and that didn't seem to make a difference.<br>However, the problem did seem to go away after bit by itself.<br>I don't think there is anything we can do about this. I have a 3COM 3C510 Home Ethernet Gateway connected to my DSL modem. The problem probably occurs at the next hop, the CLEC I guess. Maybe someone there is screwing around, or it has a high load -- who knows.<br>Here is something else I noticed: a router went down somewhere around hop 15 of about 23 hops to my destination. I noticed about 20 minutes later. Maybe since a router down route was down, ping plotter was swamping it with requests that were not coming back and a router in between got sick of that? Then I stopped it and fooled around with settings, and a little while later it was clear??<br>I don't much about it - do you think that's plausible?<br>Luther<br><br><br>

I don't think it's too likely that an outage at hop 15 will affect your connection at hop 2 (although this is speculation, sometimes things happen that are hard to determine or isolate). What the outage can do, however, is have more outstanding ICMP echo requests simultaneously, and some routers don't seem to work well with this. When you have an outage with a high timeout time (ie: 10 second timeout), this can mean that you have a pretty much continuous stream of oustanding packets - which a router might not work well with (this is how something at hop 15 can affect hop 1 or 2). Changing Ping Plotter to be single threaded should take care of this, though.<br><br>Something you can try is to see what the command line tool TRACERT does to the same destination. I like to use the -t flag on TRACERT as this turns off the reverse DNS lookups (which take *forever* it seems like). TRACERT always works single threaded - and just about all routers have been tested to work reliably with it.<br><br>