Hi all,

I think I've identified a bug, it exists in version 1.0, I haven't tested version 1.01 yet, but didn't see any mention of issues with floating graphs, so I thought I'd submit this:

We were having a problem with a user's network connection, I started multiping, pinging at a rate of 10 per second toward 10 to 20 end station targets. The graphs are all floating arranged nicely on the screen. During the troubleshooting, the user rebooted and received a new DHCP address, so I removed the old target, added the new address target all without stopping and restarting the pinging. Everything was fine for a short while, then something odd started happening.

In the problem machine's new graph, you could see multiping plotting away the replies for about 5 to 10 seconds, then poof, the whole part it had just completed plotting turned red as if all the pings that you watched plot successfully had actually failed. Stopping the pings, then restarting didn't remove the problem either. I was able to reproduce the problem with 2 new windows created while pinging, both exhibiting the same behavior.

Saving the workspace, then closing the program and relaunching with the workspace did restore sanity.

Being an old programmer, I'd suspect some memory managment/pointer problem <img src="/forums/images/graemlins/tongue.gif" alt="" />

Hi, Scott.

Thanks for the report.

Is there any chance that the time changed on this computer during this event? It doesn't sound like that happened, but we've seen a similar issue at one case in the past where the time changed mid-collection.

Do you have a file called "MP_errors.log" in your MultiPing install directory? This is where "silent" errors go to die (in case there was an error that caused it). How about a "MultiPing.rip" file? This is where a bit more visible errors go to die, and this also gives us some additional troubleshooting information. If these files exist, have a look at the contents to see if there are any events near the time you saw these problems occure.

Now, for the next (inevitable) question: Is there any chance you can reproduce this situation in another instance of MultiPing?

As you describe the problem, I'm not sure if the new data that was being collected was all being shown as packet loss also, or if it was just the history before the problem that was shown as packet loss. Did you have a chance to change the graph scale or scroll back in time to see if all the history was shown as packet loss, or just a short period?

If you do see this problem again, I suspect that it won't be limited to just the floating graph - but would also be visible in a time graph docked to the main window (which is turned on by double-clicking a target).

Before I go elaborating any more, I'll wait to see if there are any hits on a possible time change in your computer's clock, with any of the error logs, or with any reproducibility.

- Pete

Pete,

The time didn't change on the machine, there were no .log or .rip files in the multiping directory. I've just tried to reproduce the problem using a slower ping rate (2 per second), deleting and adding new targets while pinging, and I was unable to reproduce the problem. I was able to produce the problem twice yesterday during our troubleshooting session, each time was after several minutes (more than 10, less than 30) of continuous pinging at the rate of 10 per second, then removing one or more targets, and adding new targets.

I believe, attempting to remember the scenarios, a few initial pings to the new target were successful (less than 5 seconds), then a period of several seconds (7-10) of loss, followed by good responses, then only that window would suddenly draw a red bar from where the period of loss ended to the right edge of that graph, obliterating the new "good response" part of the graph. And that pattern of behavior continued for only the new graph until I exited multiping and brought it back up.

I did not attempt to change the time period shown on the graph, and for those graphs there wasn't enough data collected to be able to scroll back.

I'll try to reproduce the problem again a few times today, then I'll probably upgrade to the new version.

Thanks,

-Scott

I was unable to reproduce the issue today. <img src="/forums/images/graemlins/confused.gif" alt="" />

I'll be upgrading now, so, unless you hear back from me or from someone else who's noticed the same problem, I'd just ignore this case for now...

We'll keep this as an open issue for now. Thanks for the efforts to reproduce. Let us know if you run in to any similar problems.