![]() | |
![]() |
| | Thread Tools | Search this Thread | Display Modes |
#1
| |||
| |||
|
#2
| |||
| |||
|
|
Posts to several groups have yielded no answers yet. Perhaps this one has a more experienced readership. :,) Situation: (1) Application subscribes to events generated by a service, via Remoting and an event wrapper. Both are on the same machine. (2) LAN cable is disconnected from the machine. (3) App receives no more events! However, the remoting proxy is still active and the application can invoke methods in the service. The service reports an exception that "the connection has been forcibly closed" when it tries to raise the event to the app. Modified Situation: (1) LAN cable is disconnected from the machine. (2) Application starts and subscribes as before. (3) LAN cable is connected. (4) LAN cable is disconnected. (5) NO PROBLEMS! For obvious reasons, having the user disconnect the LAN while they start the application is not an acceptable procedure. What can be done? |
#3
| |||
| |||
|
|
Since you are on the local PC, you can use the IpcClient/Server channels OR you can install (i think available on xp sp 2 or later, but maybe some other ones) the "loopback" network driver from add/remove programs. |
#4
| |||
| |||
|
|
"John W." <JohnW (AT) discussions (DOT) microsoft.com> wrote in message news:FD23AB78-FF72-4F98-8546-ECDE6A434D68 (AT) microsoft (DOT) com... Situation: (1) Application subscribes to events generated by a service, via Remoting and an event wrapper. Both are on the same machine. (2) LAN cable is disconnected from the machine. (3) App receives no more events! However, the remoting proxy is still active and the application can invoke methods in the service. The service reports an exception that "the connection has been forcibly closed" when it tries to raise the event to the app. Modified Situation: (1) LAN cable is disconnected from the machine. (2) Application starts and subscribes as before. (3) LAN cable is connected. (4) LAN cable is disconnected. (5) NO PROBLEMS! For obvious reasons, having the user disconnect the LAN while they start the application is not an acceptable procedure. What can be done? Is this on XP? The TCP stack isn't loaded if you don't have a cable connected for some gawd awful reason even if you have a network card. Since you are on the local PC, you can use the IpcClient/Server channels OR you can install (i think available on xp sp 2 or later, but maybe some other ones) the "loopback" network driver from add/remove programs. I think there's a KB article out there about this issue. -- Doug Semler, MCPD a.a. #705, BAAWA. EAC Guardian of the Horn of the IPU (pbuhh). The answer is 42; DNRC o- Gur Hfrarg unf orpbzr fb shyy bs penc gurfr qnlf, abbar rira erpbtavmrf fvzcyr guvatf yvxr ebg13 nalzber. Fnq, vfa'g vg? |
#5
| |||
| |||
|
|
"Doug Semler" <dougsemler (AT) gmail (DOT) com> wrote in message news:eVJHJcfAIHA.5960 (AT) TK2MSFTNGP05 (DOT) phx.gbl... Since you are on the local PC, you can use the IpcClient/Server channels OR you can install (i think available on xp sp 2 or later, but maybe some other ones) the "loopback" network driver from add/remove programs. Add hardware that is.... Install Manually...Network Adapters...Microsoft...Microsoft Loopback Adapter -- Doug Semler, MCPD a.a. #705, BAAWA. EAC Guardian of the Horn of the IPU (pbuhh). The answer is 42; DNRC o- Gur Hfrarg unf orpbzr fb shyy bs penc gurfr qnlf, abbar rira erpbtavmrf fvzcyr guvatf yvxr ebg13 nalzber. Fnq, vfa'g vg? |
#6
| |||
| |||
|
|
Found a time slot to check it today (never would have found it without your updated directions!), and... Rats! That wasn't the solution. |
#7
| |||
| |||
|
|
Sorry...Google lookup DisableDHCPMediaSense...and follow the directions there. if that doesn't work, i'd suggest switching to IPCServer/Client communication when the remoting is on the local system (that way you don't require the TCP/IP stack....and it's much faster) -- Doug Semler, MCPD |
#8
| |||
| |||
|
|
Can't do the disable thing. This connection is one of several to identical services on other machines. I need to detect loss of connectivity to the other servers, and recovery of those connections occurs flawlessly right now. I just need the localhost connection to either break both links (events and remoting proxy) or leave both links intact instead of breaking just one and leaving the one I use to detect the break. IPC may be the solution path, but there's a bunch of learning to be done before I can go there, and then the localhost connection (and service?) becomes unique. Ugh. |
#9
| |||
| |||
|
|
Yeah I know. We actually redesigned our solution at one point to specficially avoid remoted events. Ingo Rammer's book has a good demonstration on how to do remoted events if you haven't already looked into that. -- Doug Semler, MCPD a.a. #705, BAAWA. EAC Guardian of the Horn of the IPU (pbuhh). The answer is 42; DNRC o- Gur Hfrarg unf orpbzr fb shyy bs penc gurfr qnlf, abbar rira erpbtavmrf fvzcyr guvatf yvxr ebg13 nalzber. Fnq, vfa'g vg? |
#10
| |||
| |||
|
|
I'm trying to understand IPC for this, but have run into two walls. (1) I built a mod of the IPC examples that resembles what we do and got it working in a variety of ways, but when I try to add events I get an error about something not being "allowed to be serialized at this level." Another variation dies with "the client doesn't have a sink channel...." (2) When I change the channel from TCP to IPC in my real apps, I get "Access denied" errors when the client app tries to connect with the server object. I suspect this is a higher-up cousin of the first of the above errors. I'm beginning to see why you abandoned the events. Wish I could. |
![]() |
| Thread Tools | Search this Thread |
| Display Modes | |
| |