Ping Localhost Erratic

Asus / Pp-dlw motherboard
May 27, 2009 at 10:00:01
Specs: Windows XP, Dual Intel Xeon / 4GB Ram
I'm seeing strange random and erratic ping results when I try to ping the local machine. It doesn't happened every time, I'd say one out of every four boots it responds as expected and is stable. The other times it looks something like this:

Reply from 10.0.0.175: bytes=32 time<1ms TTL=128
Reply from 10.0.0.175: bytes=32 time<1ms TTL=128
Reply from 10.0.0.175: bytes=32 time<543345ms TTL=128
Reply from 10.0.0.175: bytes=32 time<534546ms TTL=128
Reply from 10.0.0.175: bytes=32 time<523498ms TTL=128
Reply from 10.0.0.175: bytes=32 time<523478ms TTL=128
Reply from 10.0.0.175: bytes=32 time<589723ms TTL=128
Reply from 10.0.0.175: bytes=32 time<1ms TTL=128

It continues to cycle like this indefinatley, going from 1ms to in (or around) the 500000's. When it reports the large number it seems instentanious, like it will spit out 20 lines in the blink of an eye.

I have disabled the onboard NIC and installed a PCI one. I have done the WinSock XP fix and rebuilt TCPIP manually via Microsofts instructions. I have tried to disable everything in msconfig startup programs and I've tried to disable non microsoft services. Nothing changes this behavior.

I've never seen anything like this. Does anybody have any ideas why this might be happening? Any suggestions to try. The system is a production system so I cant wipe it out or have it down for extented periods of time.


See More: Ping Localhost Erratic

Report •


#1
May 27, 2009 at 10:08:53
do you get the same results if you ping loopback ip?

Report •

#2
May 27, 2009 at 13:06:41
Exactly what use could there be to ping the local IP address???

There is no use for this test is the answer.

Do you have any other real problems that you are working on and just got fouled up on this issue?

"Best Practices", Event viewer, host file, perfmon, antivirus, anti-spyware, Live CD's, backups, are in my top 10


Report •

#3
May 27, 2009 at 14:01:12
Backstory: I have a piece of software that, when launched, does a quick check on the ethernet card (TCPIP) to verify it is working properly. The software would usually fail to load and we found it was directly related to whether the ping was responing properly or not. meaning normal ping, software works, erratic ping software doesnt work.

The ping replies in my first post were a guess from memory, here is a copy and paste from pinging localhost 127.0.0.1


Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237796ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237797ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237797ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237798ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237799ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237800ms TTL=128
Reply from 127.0.0.1: bytes=32 time=-237800ms TTL=12
Reply from 127.0.0.1: bytes=32 time=237801ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237801ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237802ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237803ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237803ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237804ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237804ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237805ms TTL=128
Reply from 127.0.0.1: bytes=32 time=237806ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time<1ms TTL=128
Reply from 127.0.0.1: bytes=32 time=-244045ms TTL=12
Reply from 127.0.0.1: bytes=32 time=244046ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244047ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244048ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244048ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244049ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244050ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244050ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244051ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244051ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244052ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244053ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244053ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244054ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244055ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244055ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244056ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244057ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244057ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244058ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244059ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244059ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244060ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244060ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244061ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244062ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244062ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244063ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244064ms TTL=128
Reply from 127.0.0.1: bytes=32 time=244064ms TTL=128


Report •

Related Solutions

#4
May 27, 2009 at 14:20:29
OK so the problem is isolated to that nic.
We are not dealing with a dns or name resolution problem.
What you have done hardware wise should eliminate the nic hardware as the issue.

I would bring up task manager and run the ping test again. See if any tasks are consuming cpu cycles which can rob ping of the focus.

Make sure you are conflicter/virus/malware free,


Report •

#5
May 27, 2009 at 16:56:01
I ran task manager and no cpu usage appeared to be in sync with the problem. I proceeded to admin tool/services and stopped every stoppable service, there was no effect. I then went to the Task Manager / Processes tab and did an 'End Process' on every non critical process, no effect. Any suggestions?

Note:
I haven't been seen the erratic behavior in safe mode with networking.

Note:
The system used to harbor a Windows 2000 installation and it has removable bays so I removed XP and through in the Win2k drive. I could not reproduce the problem in Win2k. Pings report a constant <10ms.


Report •

#6
May 27, 2009 at 18:33:31
Then go to perfmon maybe, look for cpu and memory issues.

This is still an odd way to do stuff. On many OS's ping to localhost will work with no nic installed. Only suggested that tcp/ip stack is working. Not sure if MS changed that in xp.

Kind of got me contused here. You say two nics do the same thing. That kind or rules out either nic and drivers too.

Ping is a low priority so maybe other tasks are at work. Guess you could also view wireshark for more info on lan traffic.


Could ring the ethernet cable maybe crosstalk or other fails on a bend or connector.

"Best Practices", Event viewer, host file, perfmon, antivirus, anti-spyware, Live CD's, backups, are in my top 10


Report •

#7
May 27, 2009 at 21:08:13
All indications from your tests and the W2K swap is that it isn't hardware. But the symptoms point to hardware issues like cpu utilization, memory, overheating or the most likely; irq conflict.

Then it comes back to the OS.

I would try a system restore to an earlier date to see if the issue is resolved.

If not, a repair reinstall is in order.

I have had the procedure put me back at install defaults for OS [means you have to do all the updates again]. It kept all my data intact. I did have to reinstall a couple of photo apps that seem to have lost their way [would throw errors upon launch]. Everything works great.


Report •

#8
May 29, 2009 at 16:11:04
Well after another day of trying various things (including reinstalling windows on another dive, updating bios, etc.) I finally made some headway. I've found that the erratic ping (under XP only) is directly related to whether Hyper-threading is enabled in the BIOS. I've found that with Hyper-threading disabled, pinging localhost is stable and works as expected.

So my question is, why would this be? And maybe more importantly, what negative effect will running with hyper-threading disabled have?


Report •

#9
May 29, 2009 at 16:23:03
The MS tcp/ip stack is SMP [symetric multiprocessing] compliant. I have to guess they also wrote it to be hyperthreading. My understanding is hyperthreading is a subset of SMP.

Running hyperthreading disabled will only effect the performance of those OS pieces and programs written for hyperthreading.

You would not want to disable hyperthreading on a server running exchange or sql server for example since they are both written to use HT.

BTW EXCELLENT job on your troubleshooting. Superb!


Report •


Ask Question