compsol1993
asked on
Lost ability to connect to Client and Server PC's using Remote Web Workplace
I have a network that has been up and running for roughly a year. We have consistantly used RWW to access the server and client pcs to make changes at the site from our office, which is off-site.
The system is Small Business Server 2003 running Exchange 2003 for all email, internal and external. The server is also the DNS server. The network has between 27 and 32 pcs logged on at a time, the client pcs are all running XP Pro SP2. DHCP is supplied via a Linksys router, using port forwarding for RWW and Email.
There is an on-site tech that works with us on occasion, but takes care of most of the simple problems without having to call.
Three weeks ago, the DSL server was switched, and the tech at the new provider company instructed the on-site tech to change the ip address of the server. Without contacting us, she went ahead and made the change. Over the next couple of days there were major problems, so my co-worker went to look into the situation. It was found that the DHCP on the server had been turned on and was conflicting with the DHCP coming from the router. He also changed the ip address back to the original because it should never have been changed in the first place. At that time my co-worker also set up the port forwarding again for RWW and Email since the service provider had also had the on-site tech reset that. Not sure why???!!!
Everything worked smoothly for roughly a week, then out of the blue the RWW had problems connecting to the client PC's and the server. By trouble, I mean it wouldn't connect to them at all. I thought it was router problems, so again, I traveled to the site, and reset/reconfigured the router just incase something had blown on it. Nothing changed.
The problem I need an answer or help with is that I can login to RWW as any type of user, I can use OWA for mail, I can use the internal website. I CANNOT connect to any pc or server. I can see them, but when I choose to connect I get the always enjoyable message:
"Connectivity to the remote computer could not be established. Ensure that the remote computer is on and connected to the Windows Small Business Server Network."
So far this is what I have discovered:
1) I can login to RWW console, use OWA for mail, and use all other features of RWW including the internal web site.
2) I can choose to connect to Clients or Servers, and they will show in the "choose to connect" window, so I know it sees the online PCs, it won't go any farther
3) from an external pc, I cannot ping the "x.x.x.x:portNumber" for Remote Desktop Connection
4) from an external pc, I can ping the ip address of the router, but it is set to not accept requests. It is found, but will not return ping data.
5) from an external pc, I can use Remote Desktop Connection to access the server (because the port points there for RWW to work)
6) Internally, I can ping the name of a PC and get the resolved internal IP address
My conclusions thus far:
It seems as though all of the pieces of the puzzle are there. I can use Remote Desktop, which RWW uses to connect to network pcs, I can access RWW, I can ping the client pc names internally and get resolved IPs. I've come to the assumption that somewhere in RWW there is a problem accessing the current IP routing table for the network pcs. That is the place where I get the message.
I'm looking for ideas on the problem, but also, if available, ways to look at said routing table that the RWW is using. I assume it's using the same one that ping goes off of, but maybe it's just having trouble getting there.
This is very urgent, I have a remote user of the site that needs to be able to access their internal pc resources ASAP!
The system is Small Business Server 2003 running Exchange 2003 for all email, internal and external. The server is also the DNS server. The network has between 27 and 32 pcs logged on at a time, the client pcs are all running XP Pro SP2. DHCP is supplied via a Linksys router, using port forwarding for RWW and Email.
There is an on-site tech that works with us on occasion, but takes care of most of the simple problems without having to call.
Three weeks ago, the DSL server was switched, and the tech at the new provider company instructed the on-site tech to change the ip address of the server. Without contacting us, she went ahead and made the change. Over the next couple of days there were major problems, so my co-worker went to look into the situation. It was found that the DHCP on the server had been turned on and was conflicting with the DHCP coming from the router. He also changed the ip address back to the original because it should never have been changed in the first place. At that time my co-worker also set up the port forwarding again for RWW and Email since the service provider had also had the on-site tech reset that. Not sure why???!!!
Everything worked smoothly for roughly a week, then out of the blue the RWW had problems connecting to the client PC's and the server. By trouble, I mean it wouldn't connect to them at all. I thought it was router problems, so again, I traveled to the site, and reset/reconfigured the router just incase something had blown on it. Nothing changed.
The problem I need an answer or help with is that I can login to RWW as any type of user, I can use OWA for mail, I can use the internal website. I CANNOT connect to any pc or server. I can see them, but when I choose to connect I get the always enjoyable message:
"Connectivity to the remote computer could not be established. Ensure that the remote computer is on and connected to the Windows Small Business Server Network."
So far this is what I have discovered:
1) I can login to RWW console, use OWA for mail, and use all other features of RWW including the internal web site.
2) I can choose to connect to Clients or Servers, and they will show in the "choose to connect" window, so I know it sees the online PCs, it won't go any farther
3) from an external pc, I cannot ping the "x.x.x.x:portNumber" for Remote Desktop Connection
4) from an external pc, I can ping the ip address of the router, but it is set to not accept requests. It is found, but will not return ping data.
5) from an external pc, I can use Remote Desktop Connection to access the server (because the port points there for RWW to work)
6) Internally, I can ping the name of a PC and get the resolved internal IP address
My conclusions thus far:
It seems as though all of the pieces of the puzzle are there. I can use Remote Desktop, which RWW uses to connect to network pcs, I can access RWW, I can ping the client pc names internally and get resolved IPs. I've come to the assumption that somewhere in RWW there is a problem accessing the current IP routing table for the network pcs. That is the place where I get the message.
I'm looking for ideas on the problem, but also, if available, ways to look at said routing table that the RWW is using. I assume it's using the same one that ping goes off of, but maybe it's just having trouble getting there.
This is very urgent, I have a remote user of the site that needs to be able to access their internal pc resources ASAP!
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
These were all great things to try, and I will keep them in mind for the future. It ended up being something sortof simple that just took a while to finally find during research.
We found that mad.exe basically went haywire, and would not close the ports it was using for connections from the client pcs. When client was logging off, Outlook would close, but somewhere (I assume) in the application layer, there was a problem communicating between Microsoft Exchange Server Attendant and Rpc to release the port. I found this by running netstat -aon whereby I found that all of my server ports were being used up by a particular process and were stuck in the CLOSE_WAIT State. I tried to release them, but could not. I traced the PID back to the processes tab of the task manager, went to services and manually shut down the application. After a reboot, everything seems to be running fine.
I would attribute the problem with MAD.EXE to the technician on-site changing the IP address at the request of the ISP. My conclusion of this comes from other posts I've read on EE whereby this problem or similar problems have occured after a change in IP Address. Internal Ip that is.
Thanks for your response.
We found that mad.exe basically went haywire, and would not close the ports it was using for connections from the client pcs. When client was logging off, Outlook would close, but somewhere (I assume) in the application layer, there was a problem communicating between Microsoft Exchange Server Attendant and Rpc to release the port. I found this by running netstat -aon whereby I found that all of my server ports were being used up by a particular process and were stuck in the CLOSE_WAIT State. I tried to release them, but could not. I traced the PID back to the processes tab of the task manager, went to services and manually shut down the application. After a reboot, everything seems to be running fine.
I would attribute the problem with MAD.EXE to the technician on-site changing the IP address at the request of the ISP. My conclusion of this comes from other posts I've read on EE whereby this problem or similar problems have occured after a change in IP Address. Internal Ip that is.
Thanks for your response.
Actually, had you run the Change Server IP address wizard which I recommended above it SHOULD have corrected that problem.
Jeff
TechSoEasy
Jeff
TechSoEasy
ASKER
I have used the netstat command in a command prompt to discover that RWW is trying to use a number of ports other than 4125, and that each of them is stuck in a CLOSE_WAIT state. I have also found that Mad.exe is using my port 4125, and each of its connections are stuck in a CLOSE_WAIT state. Besides that, I have also found a number of connections using various high number ports such as 56124 for example are being made from "FDQN:ldap"....FDQN being the qualified name of my server internally (ex. servername.domain.local:ld
In the knowledge base article it explains how to modify the registry to make the port use problem go away, however, I know have multiple ports that are stuck in a CLOSE_WAIT state, and over the past coupld of days I have recieved resource warnings from the server, so I'm assuming that these sockets are never being released as they should.
I did try setting the TcpTimedWaitDelay key in the registry to 30 down from 60 as instructed by another post here on EE to try to get the CLOSE_WAIT to release sooner, but to no avail.
Questions:
1) What are the numerous connections I see from servername.domain.local:ld
2) And How do I clean out or remove these CLOSE_WAIT session from the command prompt?
I had a reboot performed two days ago, while the problem existed, and it did not work, so I did not try that yet today either.
Thanks for the help in advance.