posted
Here is a post I put in a tech forum. It is a revamped post from many months ago of the same problem. I can not find a hot fix or suggestion on what to do to fix the problem.
Any Ideas?:
I had a network that ran on NT 4.0. I had a separate cluster of computers that shared a work group that was off of our network that accessed a program (Read180) installed on a 2003 server.
My NT4.0 died and I upgraded to a 2003 SP1 server for our school network.
At that time, our READ180 program started getting real glitchy on the other 2003 server that was not on our network.
When I did a "ping -l 1500 -n 1000" on the Read180 server, it always started with 1 "request timed out." It would then ping successfully for about 90 seconds and then it would "request timed out" 4-6 times in a row. It would ping normally for another 40-90 seconds and time out again for 4-6 times. No matter how long the packet was or how many times I pinged, I would get the same cyclicle error message.
In my management console there are not errors on either server.
Well, it just so happened that we got a better version of the read180 program and we bought a super cool new server to host it.
However, I am getting the same request timed out error in the cyclical fashion. Because I have an updated program and a much better server, I don't think it is a hardware issue or that particular program.
I have changed locations of the server. I put it in two different hallways and eventually put it up in the ceiling directly connected to our backbone (all of our pulled direct cables are already in use).
Same error cycle.
I have updated drivers. I have even put the read180 server and it's client machines onto our network. I thought maybe it was a server 2003 AD thingy. I still get the same problem.
I am at my wits end. Any suggestions? About the only thing that has changed is that I am using the new server as a Print Server as well as Active Directory, DHCP, DNS, and File Share...and of course it is a 2003 server.
Posts: 2445 | Registered: Oct 2004
| IP: Logged |
quote: However, I am getting the same request timed out error in the cyclical fashion. Because I have an updated program and a much better server, I don't think it is a hardware issue or that particular program.
Perhaps I am missing what you are saying, but if that program is the only software on the server that is having the problem, how is it not that particular program?
Posts: 13123 | Registered: Feb 2002
| IP: Logged |
posted
SS: he's pinging the server the software is on. Pings are handled by the computer's TCP/IP stack, not a piece of 3rd party software.
Though that suggests the question, does Read180 install any networking components?
Are you always pinging from the same machine/a machine running the same software?
Do the number of pings in the time intervals vary greatly? If the server is internet-exposed, would you mind giving out the IP address so we can ping it?
Can you obtain the same results by pinging any other machine on your network?
Posts: 15770 | Registered: Dec 2001
| IP: Logged |
quote:but if that program is the only software on the server that is having the problem, how is it not that particular program?
The program is the only program on the server. That doesn't mean the problem is isolated to the program. For example, I am pinging the server not the program.
The program update is actually going from a basic program to an Enterprise Edition. I had to reinstall 5 different discs vs. the old Read180's single disc. The Read 180 server had no problems before my main server update.
The problem started when I updated my NT server. I think it is a server issue not a program issue.
Now as we are talking, I have pinged the read180 server from a different machine (that is on the same subnet as the read180 server) and it is not timing out.
Previously I tested on my computer that is on a different subnet as the Read 180 sever. We only have 2 subnets. When I pinged the 2 subnet gateways and 2 IP addresses on our main server (it has 2 nics with different IP addresses assigned to it from our two subnets), everything went through fine. When I pinged just the Read180 server that is on a different subnet (but the same subnet as one of our Main server's NIC, it did the cyclical time out.
So out of 5 ping requests there was only time outs on that particular Read180 server, but no time outs on our main server that has a NIC card that is on the same subnet as the Read180 sever. HOWEVER, on a different computer (where I am sitting now) that is on the same subnet as READ180, I am NOT getting "request timed out."
I wonder if it is a router issue. It is kind of confusing to explain.
Posts: 2445 | Registered: Oct 2004
| IP: Logged |
posted
It sounds like a router issue. A number of major ISPs currently block ICMP Echo traffic from crossing network boundaries, and I wouldn't be surprised if some routers had modifications intended to disrupt attempts by worms to use ping to detect potential targets (as has been part of some of the most devastating worm infestations in history).
Posts: 15770 | Registered: Dec 2001
| IP: Logged |
What is your infrastructure based on (switched/routed?)
Are there multiple routing hops between the pinger and pingee?
Could it be that there is a traffic issue, and you are losing packets through collisions?
If you are running over a switched network could you run a sniff-trace on a mirrored port to see what you are sending and receiving?
That's all that occurs to me at the mo - but keep us updated - and sorry if this is all stuff youhave already tried - great minds think alike Posts: 892 | Registered: Oct 2006
| IP: Logged |
quote:Yeah, knew that. Which is the general angle that I was getting at.
Sorry, I didn't make the mental connect.
quote:Though that suggests the question, does Read180 install any networking components?
It does use MySQL, JBoss, and Java. I believe it has a database that updates via the web.
quote: Are you always pinging from the same machine/a machine running the same software?
No, I have pinged from machines on both ip ranges. I have pinged from machines that use the software and machines that don't.
quote:Do the number of pings in the time intervals vary greatly? If the server is internet-exposed, would you mind giving out the IP address so we can ping it?
No, the interrupts are almost times perfect at around 90 seconds. I emailed you my servers IP.
quote: Can you obtain the same results by pinging any other machine on your network?
Oddly enough, I tried pinging 4 machines on the same ip range as our read 180 server. 2 of them had nothing but "request timed out" 2 of them had no interrupts. Pinging my server gets the cyclical errors. This is really weird.
I have never run an ip sniffer. I will check that out tonight.
Posts: 2445 | Registered: Oct 2004
| IP: Logged |
quote: Not to be insulting, but you did check here and/or talk to their tech support?
It is not insulting. There are many times when I focus on complex issues in my mind to only discover there was something super easy I overlooked.
This is not one of those times. I did call tech support. They assured me this was not a read180 issue. I looked through their knowledge base and only found out that if you can't ping the server you are not connected. I already knew that. I am trying to find out why it is not connecting.
Actually it is connecting; it just keeps loosing the connection. It will slow down really bad or crash the session.
Posts: 2445 | Registered: Oct 2004
| IP: Logged |
Printer-friendly view of this topic
![[Wink]](wink.gif)
